US 9,916,703 B2Grant

Calibration for autonomous vehicle operation

Issue Date:2018-03-13

•18 Claims

•47 Drawing Sheets

Abstract

Various embodiments relate generally to autonomous vehicles and associated mechanical, electrical and electronic hardware, computer software and systems, and wired and wireless network communications to provide an autonomous vehicle fleet as a service. In particular, a method may include receiving data associated with a sensor measurement of a perceived object, determining a label associated with the perceived object based on an initial calibration, retrieving log file data associated with the label, determining a calibration parameter associated with the sensor measurement based on the retrieved log file data, and storing the calibration parameter in association with a sensor associated with the sensor measurement. Sensors may be calibrated on the fly while the autonomous vehicle is in operation using one or more other sensors and/or fused data from multiple types of sensors.

Metadata

Assignee

Zoox, Inc.

Inventors

Jesse Sol Levinson
Bertrand Robert Douillard
Gabriel Thurston Sibley

Application Information

Application Number:US 14/756,996

Filing Date:2015-11-04

Priority Date:2015-11-04

Art Unit:3662

Classifications

IPC:

G01C25/00G07C5/08B60Q1/26G05D1/00G05D1/02G08G1/00G08G1/005G08G1/056G08G1/0968

Field of Search:

G07C 5/0808G01C 25/00G05D 1/024G05D 1/0257G05D 1/027G05D 1/0038G05D 1/0274G05D 1/0291G05D 1/0272B60Q 1/26G08G 1/202G01S 7/497G01S 7/4972

Patent Drawings (47 sheets)

Description

Cross-Reference to Related Applications

[0001] This application is related to U.S. patent application Ser. No. 14/932,959, filed Nov. 4, 2015, (now U.S. Pat. No. 9,606,539) entitled “AUTONOMOUS VEHICLE FLEET SERVICE AND SYSTEM,” U.S. patent application Ser. No. 14/932,963, filed Nov. 4, 2015, (now U.S. Pat. No. 9,612,123) entitled “ADAPTIVE MAPPING TO NAVIGATE AUTONOMOUS VEHICLES RESPONSIVE TO PHYSICAL ENVIRONMENT CHANGES,” U.S. patent application Ser. No. 14/932,940, filed Nov. 4, 2015, (now U.S. Pat. No. 9,734,455) entitled “AUTOMATED EXTRACTION OF SEMANTIC INFORMATION TO ENHANCE INCREMENTAL MAPPING MODIFICATIONS FOR ROBOTIC VEHICLES,” U.S. patent application Ser. No. 14/933,602, filed Nov. 4, 2015, (now U.S. Pat. No. 9,632,502) entitled “SYSTEMS AND METHOD TO OPTIMIZE NAVIGATION OF AUTONOMOUS VEHICLE TELEOPERATION AND PLANNING LOGIC,” U.S. patent application Ser. No. 14/757,016, filed Nov. 5, 2015, entitled “SIMULATION SYSTEM AND METHODS FOR AUTONOMOUS VEHICLES,” U.S. patent application Ser. No. 14/756,992, filed Nov. 4, 2015, entitled “ADAPTIVE AUTONOMOUS VEHICLE PLANNER LOGIC,” U.S. patent application Ser. No. 14/756,993, filed Nov. 4, 2015, entitled “METHOD FOR ROBOTIC VEHICLE COMMUNICATION WITH AN EXTERNAL ENVIRONMENT VIA ACOUSTIC BEAM FORMING,” U.S. patent application Ser. No. 14/756,991, filed Nov. 4, 2015, (now U.S. Pat. No. 9,720,415) entitled “SENSOR-BASED OBJECT-DETECTION OPTIMIZATION FOR AUTONOMOUS VEHICLES,” U.S. patent application Ser. No. 14/932,948, filed Nov. 4, 2015, (now U.S. Pat. No. 9,804,599) entitled “ACTIVE LIGHTING CONTROL FOR COMMUNICATING A STATE OF AN AUTONOMOUS VEHICLE TO ENTITIES IN A SURROUNDING ENVIRONMENT,” U.S. patent application Ser. No. 14/932,958, filed Nov. 4, 2015, (now U.S. Pat. No. 9,494,940) entitled “QUADRANT CONFIGURATION OF ROBOTIC VEHICLES,” and U.S. patent application Ser. No. 14/932,962, filed Nov. 4, 2015, entitled “ROBOTIC VEHICLE ACTIVE SAFETY SYSTEMS AND METHODS,” all of which are hereby incorporated by reference in their entirety for all purposes.

Field

[0002] Various embodiments relate generally to autonomous vehicles and associated mechanical, electrical and electronic hardware, computer software and systems, and wired and wireless network communications to calibrate sensors for autonomous vehicle operation.

Background

[0003] A variety of approaches to developing driverless vehicles focus predominately on automating conventional vehicles (e.g., manually-driven automotive vehicles) with an aim toward producing driverless vehicles for consumer purchase. For example, a number of automotive companies and affiliates are modifying conventional automobiles and control mechanisms, such as steering, to provide consumers with an ability to own a vehicle that may operate without a driver. In some approaches, a conventional driverless vehicle performs safety-critical driving functions in some conditions, but requires a driver to assume control (e.g., steering, etc.) should the vehicle controller fail to resolve certain issues that might jeopardize the safety of the occupants.

[0004] Although functional, conventional driverless vehicles typically have a number of drawbacks. For example, a large number of driverless cars under development have evolved from vehicles requiring manual (i.e., human-controlled) steering and other like automotive functions. Therefore, a majority of driverless cars are based on a paradigm that a vehicle is to be designed to accommodate a licensed driver, for which a specific seat or location is reserved within the vehicle. As such, driverless vehicles are designed sub-optimally and generally forego opportunities to simplify vehicle design and conserve resources (e.g., reducing costs of producing a driverless vehicle). Other drawbacks are also present in conventional driverless vehicles.

[0005] Other drawbacks are also present in conventional transportation services, which are not well-suited for managing, for example, inventory of vehicles effectively due to the common approaches of providing conventional transportation and ride-sharing services. In one conventional approach, passengers are required to access a mobile application to request transportation services via a centralized service that assigns a human driver and vehicle (e.g., under private ownership) to a passenger. With the use of differently-owned vehicles, maintenance of private vehicles and safety systems generally go unchecked. In another conventional approach, some entities enable ride-sharing for a group of vehicles by allowing drivers, who enroll as members, access to vehicles that are shared among the members. This approach is not well-suited to provide for convenient transportation services as drivers need to pick up and drop off shared vehicles at specific locations, which typically are rare and sparse in city environments, and require access to relatively expensive real estate (i.e., parking lots) at which to park ride-shared vehicles. In the above-described conventional approaches, the traditional vehicles used to provide transportation services are generally under-utilized, from an inventory perspective, as the vehicles are rendered immobile once a driver departs. Further, ride-sharing approaches (as well as individually-owned vehicle transportation services) generally are not well-suited to rebalance inventory to match demand of transportation services to accommodate usage and typical travel patterns. Note, too, that some conventionally-described vehicles having limited self-driving automation capabilities also are not well-suited to rebalance inventories as a human driver generally may be required. Examples of vehicles having limited self-driving automation capabilities are vehicles designated as Level 3 (“L3”) vehicles, according to the U.S. Department of Transportation's National Highway Traffic Safety Administration (“NHTSA”).

[0006] As another drawback, typical approaches to driverless vehicles are generally not well-suited to detect and navigate vehicles relative to interactions (e.g., social interactions) between a vehicle-in-travel and other drivers of vehicles or individuals. For example, some conventional approaches are not sufficiently able to identify pedestrians, cyclists, etc., and associated interactions, such as eye contact, gesturing, and the like, for purposes of addressing safety risks to occupants of a driverless vehicles, as well as drivers of other vehicles, pedestrians, etc.

[0007] Thus, what is needed is a solution for facilitating an implementation of autonomous vehicles, without the limitations of conventional techniques.

Brief Description of the Drawings

[0008] Various embodiments or examples (“examples”) of the invention are disclosed in the following detailed description and the accompanying drawings:

[0009] FIG. 1 is a diagram depicting implementation of a fleet of autonomous vehicles that are communicatively networked to an autonomous vehicle service platform, according to some embodiments;

[0010] FIG. 2 is an example of a flow diagram to monitor a fleet of autonomous vehicles, according to some embodiments;

[0011] FIG. 3A is a diagram depicting examples of sensors and other autonomous vehicle components, according to some examples;

[0012] FIGS. 3B to 3E are diagrams depicting examples of sensor field redundancy and autonomous vehicle adaption to a loss of a sensor field, according to some examples;

[0013] FIG. 4 is a functional block diagram depicting a system including an autonomous vehicle service platform that is communicatively coupled via a communication layer to an autonomous vehicle controller, according to some examples;

[0014] FIG. 5 is an example of a flow diagram to control an autonomous vehicle, according to some embodiments;

[0015] FIG. 6 is a diagram depicting an example of an architecture for an autonomous vehicle controller, according to some embodiments;

[0016] FIG. 7 is a diagram depicting an example of an autonomous vehicle service platform implementing redundant communication channels to maintain reliable communications with a fleet of autonomous vehicles, according to some embodiments;

[0017] FIG. 8 is a diagram depicting an example of a messaging application configured to exchange data among various applications, according to some embodiment;

[0018] FIG. 9 is a diagram depicting types of data for facilitating teleoperations using a communications protocol described in FIG. 8, according to some examples;

[0019] FIG. 10 is a diagram illustrating an example of a teleoperator interface with which a teleoperator may influence path planning, according to some embodiments;

[0020] FIG. 11 is a diagram depicting an example of a planner configured to invoke teleoperations, according to some examples;

[0021] FIG. 12 is an example of a flow diagram configured to control an autonomous vehicle, according to some embodiments;

[0022] FIG. 13 depicts an example in which a planner may generate a trajectory, according to some examples;

[0023] FIG. 14 is a diagram depicting another example of an autonomous vehicle service platform, according to some embodiments;

[0024] FIG. 15 is an example of a flow diagram to control an autonomous vehicle, according to some embodiments;

[0025] FIG. 16 is a diagram of an example of an autonomous vehicle fleet manager implementing a fleet optimization manager, according to some examples;

[0026] FIG. 17 is an example of a flow diagram for managing a fleet of autonomous vehicles, according to some embodiments;

[0027] FIG. 18 is a diagram illustrating an autonomous vehicle fleet manager implementing an autonomous vehicle communications link manager, according to some embodiments;

[0028] FIG. 19 is an example of a flow diagram to determine actions for autonomous vehicles during an event, according to some embodiments;

[0029] FIG. 20 is a diagram depicting an example of a localizer, according to some embodiments;

[0030] FIG. 21 is an example of a flow diagram to generate local pose data based on integrated sensor data, according to some embodiments;

[0031] FIG. 22 is a diagram depicting another example of a localizer, according to some embodiments;

[0032] FIG. 23 is a diagram depicting an example of a perception engine, according to some embodiments;

[0033] FIG. 24 is an example of a flow chart to generate perception engine data, according to some embodiments;

[0034] FIG. 25 is an example of a segmentation processor, according to some embodiments;

[0035] FIG. 26A is a diagram depicting examples of an object tracker and a classifier, according to various embodiments;

[0036] FIG. 26B is a diagram depicting another example of an object tracker according to at least some examples;

[0037] FIG. 27 is an example of front-end processor for a perception engine, according to some examples;

[0038] FIG. 28 is a diagram depicting a simulator configured to simulate an autonomous vehicle in a synthetic environment, according to various embodiments;

[0039] FIG. 29 is an example of a flow chart to simulate various aspects of an autonomous vehicle, according to some embodiments;

[0040] FIG. 30 is an example of a flow chart to generate map data, according to some embodiments;

[0041] FIG. 31 is a diagram depicting an architecture of a mapping engine, according to some embodiments;

[0042] FIG. 32 is a diagram depicting an autonomous vehicle application, according to some examples;

[0043] FIGS. 33 to 35 illustrate examples of various computing platforms configured to provide various functionalities to components of an autonomous vehicle service, according to various embodiments;

[0044] FIGS. 36A to 36C illustrate high-level block diagrams depicting an autonomous vehicle system having a sensor calibration issue while in operation, according to various embodiments;

[0045] FIG. 37 illustrates a high-level block diagram of calibration for autonomous vehicle operation, according to various embodiments;

[0046] FIG. 38 is a network diagram of a system for calibration for autonomous vehicle operation, showing a block diagram of an autonomous vehicle management system, according to an embodiment;

[0047] FIG. 39 is a high-level flow diagram illustrating a process for calibration for autonomous vehicle operation, according to some examples; and

[0048] FIGS. 40 and 41 illustrate exemplary computing platforms disposed in devices configured to calibrate autonomous vehicles in accordance with various embodiments.

Detailed Description

[0049] Various embodiments or examples may be implemented in numerous ways, including as a system, a process, an apparatus, a user interface, or a series of program instructions on a computer readable medium such as a computer readable storage medium or a computer network where the program instructions are sent over optical, electronic, or wireless communication links. In general, operations of disclosed processes may be performed in an arbitrary order, unless otherwise provided in the claims.

[0050] A detailed description of one or more examples is provided below along with accompanying figures. The detailed description is provided in connection with such examples, but is not limited to any particular example. The scope is limited only by the claims, and numerous alternatives, modifications, and equivalents thereof. Numerous specific details are set forth in the following description in order to provide a thorough understanding. These details are provided for the purpose of example and the described techniques may be practiced according to the claims without some or all of these specific details. For clarity, technical material that is known in the technical fields related to the examples has not been described in detail to avoid unnecessarily obscuring the description.

[0051] FIG. 1 is a diagram depicting an implementation of a fleet of autonomous vehicles that are communicatively networked to an autonomous vehicle service platform, according to some embodiments. Diagram 100 depicts a fleet of autonomous vehicles 109 (e.g., one or more of autonomous vehicles 109a to 109e) operating as a service, each autonomous vehicle 109 being configured to self-drive a road network 110 and establish a communication link 192 with an autonomous vehicle service platform 101. In examples in which a fleet of autonomous vehicles 109 constitutes a service, a user 102 may transmit a request 103 for autonomous transportation via one or more networks 106 to autonomous vehicle service platform 101. In response, autonomous vehicle service platform 101 may dispatch one of autonomous vehicles 109 to transport user 102 autonomously from geographic location 119 to geographic location 111. Autonomous vehicle service platform 101 may dispatch an autonomous vehicle from a station 190 to geographic location 119, or may divert an autonomous vehicle 109c, already in transit (e.g., without occupants), to service the transportation request for user 102. Autonomous vehicle service platform 101 may be further configured to divert an autonomous vehicle 109c in transit, with passengers, responsive to a request from user 102 (e.g., as a passenger). In addition, autonomous vehicle service platform 101 may be configured to reserve an autonomous vehicle 109c in transit, with passengers, for diverting to service a request of user 102 subsequent to dropping off existing passengers. Note that multiple autonomous vehicle service platforms 101 (not shown) and one or more stations 190 may be implemented to service one or more autonomous vehicles 109 in connection with road network 110. One or more stations 190 may be configured to store, service, manage, and/or maintain an inventory of autonomous vehicles 109 (e.g., station 190 may include one or more computing devices implementing autonomous vehicle service platform 101).

[0052] According to some examples, at least some of autonomous vehicles 109a to 109e are configured as bidirectional autonomous vehicles, such as bidirectional autonomous vehicle (“AV”) 130. Bidirectional autonomous vehicle 130 may be configured to travel in either direction principally along, but not limited to, a longitudinal axis 131. Accordingly, bidirectional autonomous vehicle 130 may be configured to implement active lighting external to the vehicle to alert others (e.g., other drivers, pedestrians, cyclists, etc.) in the adjacent vicinity, and a direction in which bidirectional autonomous vehicle 130 is traveling. For example, active sources of light 136 may be implemented as active lights 138a when traveling in a first direction, or may be implemented as active lights 138b when traveling in a second direction. Active lights 138a may be implemented using a first subset of one or more colors, with optional animation (e.g., light patterns of variable intensities of light or color that may change over time). Similarly, active lights 138b may be implemented using a second subset of one or more colors and light patterns that may be different than those of active lights 138a. For example, active lights 138a may be implemented using white-colored lights as “headlights,” whereas active lights 138b may be implemented using red-colored lights as “taillights.” Active lights 138a and 138b, or portions thereof, may be configured to provide other light-related functionalities, such as provide “turn signal indication” functions (e.g., using yellow light). According to various examples, logic in autonomous vehicle 130 may be configured to adapt active lights 138a and 138b to comply with various safety requirements and traffic regulations or laws for any number of jurisdictions.

[0053] In some embodiments, bidirectional autonomous vehicle 130 may be configured to have similar structural elements and components in each quad portion, such as quad portion 194. The quad portions are depicted, at least in this example, as portions of bidirectional autonomous vehicle 130 defined by the intersection of a plane 132 and a plane 134, both of which pass through the vehicle to form two similar halves on each side of planes 132 and 134. Further, bidirectional autonomous vehicle 130 may include an autonomous vehicle controller 147 that includes logic (e.g., hardware or software, or as combination thereof) that is configured to control a predominate number of vehicle functions, including driving control (e.g., propulsion, steering, etc.) and active sources 136 of light, among other functions. Bidirectional autonomous vehicle 130 also includes a number of sensors 139 disposed at various locations on the vehicle (other sensors are not shown).

[0054] Autonomous vehicle controller 147 may be further configured to determine a local pose (e.g., local position) of an autonomous vehicle 109 and to detect external objects relative to the vehicle. For example, consider that bidirectional autonomous vehicle 130 is traveling in the direction 119 in road network 110. A localizer (not shown) of autonomous vehicle controller 147 can determine a local pose at the geographic location 111. As such, the localizer may use acquired sensor data, such as sensor data associated with surfaces of buildings 115 and 117, which can be compared against reference data, such as map data (e.g., 3D map data, including reflectance data) to determine a local pose. Further, a perception engine (not shown) of autonomous vehicle controller 147 may be configured to detect, classify, and predict the behavior of external objects, such as external object 112 (a “tree”) and external object 114 (a “pedestrian”). Classification of such external objects may broadly classify objects as static objects, such as external object 112, and dynamic objects, such as external object 114. The localizer and the perception engine, as well as other components of the AV controller 147, collaborate to cause autonomous vehicles 109 to drive autonomously.

[0055] According to some examples, autonomous vehicle service platform 101 is configured to provide teleoperator services should an autonomous vehicle 109 request teleoperation. For example, consider that an autonomous vehicle controller 147 in autonomous vehicle 109d detects an object 126 obscuring a path 124 on roadway 122 at point 191, as depicted in inset 120. If autonomous vehicle controller 147 cannot ascertain a path or trajectory over which vehicle 109d may safely transit with a relatively high degree of certainty, then autonomous vehicle controller 147 may transmit request message 105 for teleoperation services. In response, a teleoperator computing device 104 may receive instructions from a teleoperator 108 to perform a course of action to successfully (and safely) negotiate obstacles 126. Response data 107 then can be transmitted back to autonomous vehicle 109d to cause the vehicle to, for example, safely cross a set of double lines as it transits along the alternate path 121. In some examples, teleoperator computing device 104 may generate a response identifying geographic areas to exclude from planning a path. In particular, rather than provide a path to follow, a teleoperator 108 may define areas or locations that the autonomous vehicle must avoid.

[0056] In view of the foregoing, the structures and/or functionalities of autonomous vehicle 130 and/or autonomous vehicle controller 147, as well as their components, can perform real-time (or near real-time) trajectory calculations through autonomous-related operations, such as localization and perception, to enable autonomous vehicles 109 to self-drive.

[0057] In some cases, the bidirectional nature of bidirectional autonomous vehicle 130 provides for a vehicle that has quad portions 194 (or any other number of symmetric portions) that are similar or are substantially similar to each other. Such symmetry reduces complexity of design and decreases relatively the number of unique components or structures, thereby reducing inventory and manufacturing complexities. For example, a drivetrain and wheel system may be disposed in any of the quad portions 194. Further, autonomous vehicle controller 147 is configured to invoke teleoperation services to reduce the likelihood that an autonomous vehicle 109 is delayed in transit while resolving an event or issue that may otherwise affect the safety of the occupants. In some cases, the visible portion of road network 110 depicts a geo-fenced region that may limit or otherwise control the movement of autonomous vehicles 109 to the road network shown in FIG. 1. According to various examples, autonomous vehicle 109, and a fleet thereof, may be configurable to operate as a level 4 (“full self-driving automation,” or L4) vehicle that can provide transportation on demand with the convenience and privacy of point-to-point personal mobility while providing the efficiency of shared vehicles. In some examples, autonomous vehicle 109, or any autonomous vehicle described herein, may be configured to omit a steering wheel or any other mechanical means of providing manual (i.e., human-controlled) steering for autonomous vehicle 109. Further, autonomous vehicle 109, or any autonomous vehicle described herein, may be configured to omit a seat or location reserved within the vehicle for an occupant to engage a steering wheel.

[0058] FIG. 2 is an example of a flow diagram to monitor a fleet of autonomous vehicles, according to some embodiments. At 202, flow 200 begins when a fleet of autonomous vehicles are monitored. At least one autonomous vehicle includes an autonomous vehicle controller configured to cause the vehicle to autonomously transit from a first geographic region to a second geographic region. At 204, data representing an event associated with a calculated confidence level for a vehicle is detected. An event may be a condition or situation affecting operation, or potentially affecting operation, of an autonomous vehicle. The events may be internal to an autonomous vehicle, or external. For example, an obstacle obscuring a roadway may be viewed as an event, as well as a reduction or loss of communication. An event may include traffic conditions or congestion, as well as unexpected or unusual numbers or types of external objects (or tracks) that are perceived by a perception engine. An event may include weather-related conditions (e.g., loss of friction due to ice or rain) or the angle at which the sun is shining (e.g., at sunset), such as low angle to the horizon that cause sun to shine brightly in the eyes of human drivers of other vehicles. These and other conditions may be viewed as events that cause invocation of the teleoperator service or for the vehicle to execute a safe-stop trajectory.

[0059] At 206, data representing a subset of candidate trajectories may be received from an autonomous vehicle responsive to the detection of the event. For example, a planner of an autonomous vehicle controller may calculate and evaluate large numbers of trajectories (e.g., thousands or greater) per unit time, such as a second. In some embodiments, candidate trajectories are a subset of the trajectories that provide for relatively higher confidence levels that an autonomous vehicle may move forward safely in view of the event (e.g., using an alternate path provided by a teleoperator). Note that some candidate trajectories may be ranked or associated with higher degrees of confidence than other candidate trajectories. According to some examples, subsets of candidate trajectories may originate from any number of sources, such as a planner, a teleoperator computing device (e.g., teleoperators can determine and provide approximate paths), etc., and may be combined as a superset of candidate trajectories. At 208, path guidance data may be identified at one or more processors. The path guidance data may be configured to assist a teleoperator in selecting a guided trajectory from one or more of the candidate trajectories. In some instances, the path guidance data specifies a value indicative of a confidence level or probability that indicates the degree of certainty that a particular candidate trajectory may reduce or negate the probability that the event may impact operation of an autonomous vehicle. A guided trajectory, as a selected candidate trajectory, may be received at 210, responsive to input from a teleoperator (e.g., a teleoperator may select at least one candidate trajectory as a guided trajectory from a group of differently-ranked candidate trajectories). The selection may be made via an operator interface that lists a number of candidate trajectories, for example, in order from highest confidence levels to lowest confidence levels. At 212, the selection of a candidate trajectory as a guided trajectory may be transmitted to the vehicle, which, in turn, implements the guided trajectory for resolving the condition by causing the vehicle to perform a teleoperator-specified maneuver. As such, the autonomous vehicle may transition from a non-normative operational state.

[0060] FIG. 3A is a diagram depicting examples of sensors and other autonomous vehicle components, according to some examples. Diagram 300 depicts an interior view of a bidirectional autonomous vehicle 330 that includes sensors, signal routers 345, drive trains 349, removable batteries 343, audio generators 344 (e.g., speakers or transducers), and autonomous vehicle (“AV”) control logic 347. Sensors shown in diagram 300 include image capture sensors 340 (e.g., light capture devices or cameras of any type), audio capture sensors 342 (e.g., microphones of any type), radar devices 348, sonar devices 341 (or other like sensors, including ultrasonic sensors or acoustic-related sensors), and Lidar devices 346, among other sensor types and modalities (some of which are not shown, such inertial measurement units, or “IMUs,” global positioning system (“GPS”) sensors, sonar sensors, etc.). Note that quad portion 350 is representative of the symmetry of each of four “quad portions” of bidirectional autonomous vehicle 330 (e.g., each quad portion 350 may include a wheel, a drivetrain 349, similar steering mechanisms, similar structural support and members, etc. beyond that which is depicted). As depicted in FIG. 3A, similar sensors may be placed in similar locations in each quad portion 350, however any other configuration may implemented. Each wheel may be steerable individually and independent of the others. Note, too, that removable batteries 343 may be configured to facilitate being swapped in and swapped out rather than charging in situ, thereby ensuring reduced or negligible downtimes due to the necessity of charging batteries 343. While autonomous vehicle controller 347a is depicted as being used in a bidirectional autonomous vehicle 330, autonomous vehicle controller 347a is not so limited and may be implemented in unidirectional autonomous vehicles or any other type of vehicle, whether on land, in air, or at sea. Note that the depicted and described positions, locations, orientations, quantities, and types of sensors shown in FIG. 3A are not intended to be limiting, and, as such, there may be any number and type of sensor, and any sensor may be located and oriented anywhere on autonomous vehicle 330.

[0061] According to some embodiments, portions of the autonomous vehicle (“AV”) control logic 347 may be implemented using clusters of graphics processing units (“GPUs”) implementing a framework and programming model suitable for programming the clusters of GPUs. For example, a compute unified device architecture (“CUDA”)-compatible programming language and application programming interface (“API”) model may be used to program the GPUs. CUDA™ is produced and maintained by NVIDIA of Santa Clara, Calif. Note that other programming languages may be implemented, such as OpenCL, or any other parallel programming language.

[0062] According to some embodiments, autonomous vehicle control logic 347 may be implemented in hardware and/or software as autonomous vehicle controller 347a, which is shown to include a motion controller 362, a planner 364, a perception engine 366, and a localizer 368. As shown, autonomous vehicle controller 347a is configured to receive camera data 340a, Lidar data 346a, and radar data 348a, or any other range-sensing or localization data, including sonar data 341a or the like. Autonomous vehicle controller 347a is also configured to receive positioning data, such as GPS data 352, IMU data 354, and other position-sensing data (e.g., wheel-related data, such as steering angles, angular velocity, etc.). Further, autonomous vehicle controller 347a may receive any other sensor data 356, as well as reference data 339. In some cases, reference data 339 includes map data (e.g., 3D map data, 2D map data, 4D map data (e.g., including Epoch Determination)) and route data (e.g., road network data, including, but not limited to, RNDF data (or similar data), MDF data (or similar data), etc.

[0063] Localizer 368 is configured to receive sensor data from one or more sources, such as GPS data 352, wheel data, IMU data 354, Lidar data 346a, camera data 340a, radar data 348a, and the like, as well as reference data 339 (e.g., 3D map data and route data). Localizer 368 integrates (e.g., fuses the sensor data) and analyzes the data by comparing sensor data to map data to determine a local pose (or position) of bidirectional autonomous vehicle 330. According to some examples, localizer 368 may generate or update the pose or position of any autonomous vehicle in real-time or near real-time. Note that localizer 368 and its functionality need not be limited to “bi-directional” vehicles and can be implemented in any vehicle of any type. Therefore, localizer 368 (as well as other components of AV controller 347a) may be implemented in a “uni-directional” vehicle or any non-autonomous vehicle. According to some embodiments, data describing a local pose may include one or more of an x-coordinate, a y-coordinate, a z-coordinate (or any coordinate of any coordinate system, including polar or cylindrical coordinate systems, or the like), a yaw value, a roll value, a pitch value (e.g., an angle value), a rate (e.g., velocity), altitude, and the like.

[0064] Perception engine 366 is configured to receive sensor data from one or more sources, such as Lidar data 346a, camera data 340a, radar data 348a, and the like, as well as local pose data. Perception engine 366 may be configured to determine locations of external objects based on sensor data and other data. External objects, for instance, may be objects that are not part of a drivable surface. For example, perception engine 366 may be able to detect and classify external objects as pedestrians, bicyclists, dogs, other vehicles, etc. (e.g., perception engine 366 is configured to classify the objects in accordance with a type of classification, which may be associated with semantic information, including a label). Based on the classification of these external objects, the external objects may be labeled as dynamic objects or static objects. For example, an external object classified as a tree may be labeled as a static object, while an external object classified as a pedestrian may be labeled as a static object. External objects labeled as static may or may not be described in map data. Examples of external objects likely to be labeled as static include traffic cones, cement barriers arranged across a roadway, lane closure signs, newly-placed mailboxes or trash cans adjacent a roadway, etc. Examples of external objects likely to be labeled as dynamic include bicyclists, pedestrians, animals, other vehicles, etc. If the external object is labeled as dynamic, and further data about the external object may indicate a typical level of activity and velocity, as well as behavior patterns associated with the classification type. Further data about the external object may be generated by tracking the external object. As such, the classification type can be used to predict or otherwise determine the likelihood that an external object may, for example, interfere with an autonomous vehicle traveling along a planned path. For example, an external object that is classified as a pedestrian may be associated with some maximum speed, as well as an average speed (e.g., based on tracking data). The velocity of the pedestrian relative to the velocity of an autonomous vehicle can be used to determine if a collision is likely. Further, perception engine 364 may determine a level of uncertainty associated with a current and future state of objects. In some examples, the level of uncertainty may be expressed as an estimated value (or probability).

[0065] Planner 364 is configured to receive perception data from perception engine 366, and may also include localizer data from localizer 368. According to some examples, the perception data may include an obstacle map specifying static and dynamic objects located in the vicinity of an autonomous vehicle, whereas the localizer data may include a local pose or position. In operation, planner 364 generates numerous trajectories, and evaluates the trajectories, based on at least the location of the autonomous vehicle against relative locations of external dynamic and static objects. Planner 364 selects an optimal trajectory based on a variety of criteria over which to direct the autonomous vehicle in way that provides for collision-free travel. In some examples, planner 364 may be configured to calculate the trajectories as probabilistically-determined trajectories. Further, planner 364 may transmit steering and propulsion commands (as well as decelerating or braking commands) to motion controller 362. Motion controller 362 subsequently may convert any of the commands, such as a steering command, a throttle or propulsion command, and a braking command, into control signals (e.g., for application to actuators or other mechanical interfaces) to implement changes in steering or wheel angles 351 and/or velocity 353.

[0066] FIGS. 3B to 3E are diagrams depicting examples of sensor field redundancy and autonomous vehicle adaption to a loss of a sensor field, according to some examples. Diagram 391 of FIG. 3B depicts a sensor field 301a in which sensor 310a detects objects (e.g., for determining range or distance, or other information). While sensor 310a may implement any type of sensor or sensor modality, sensor 310a and similarly-described sensors, such as sensors 310b, 310c, and 310d, may include Lidar devices. Therefore, sensor fields 301a, 301b, 301c, and 301d each includes a field into which lasers extend. Diagram 392 of FIG. 3C depicts four overlapping sensor fields each of which is generated by a corresponding Lidar sensor 310 (not shown). As shown, portions 301 of the sensor fields include no overlapping sensor fields (e.g., a single Lidar field), portions 302 of the sensor fields include two overlapping sensor fields, and portions 303 include three overlapping sensor fields, whereby such sensors provide for multiple levels of redundancy should a Lidar sensor fail.

[0067] FIG. 3D depicts a loss of a sensor field due to failed operation of Lidar 309, according to some examples. Sensor field 302 of FIG. 3C is transformed into a single sensor field 305, one of sensor fields 301 of FIG. 3C is lost to a gap 304, and three of sensor fields 303 of FIG. 3C are transformed into sensor fields 306 (i.e., limited to two overlapping fields). Should autonomous car 330c be traveling in the direction of travel 396, the sensor field in front of the moving autonomous vehicle may be less robust than the one at the trailing end portion. According to some examples, an autonomous vehicle controller (not shown) is configured to leverage the bidirectional nature of autonomous vehicle 330c to address the loss of sensor field at the leading area in front of the vehicle. FIG. 3E depicts a bidirectional maneuver for restoring a certain robustness of the sensor field in front of autonomous vehicle 330d. As shown, a more robust sensor field 302 is disposed at the rear of the vehicle 330d coextensive with taillights 348. When convenient, autonomous vehicle 330d performs a bidirectional maneuver by pulling into a driveway 397 and switches its directionality such that taillights 348 actively switch to the other side (e.g., the trailing edge) of autonomous vehicle 330d. As shown, autonomous vehicle 330d restores a robust sensor field 302 in front of the vehicle as it travels along direction of travel 398. Further, the above-described bidirectional maneuver obviates a requirement for a more complicated maneuver that requires backing up into a busy roadway.

[0068] FIG. 4 is a functional block diagram depicting a system including an autonomous vehicle service platform that is communicatively coupled via a communication layer to an autonomous vehicle controller, according to some examples. Diagram 400 depicts an autonomous vehicle controller (“AV”) 447 disposed in an autonomous vehicle 430, which, in turn, includes a number of sensors 470 coupled to autonomous vehicle controller 447. Sensors 470 include one or more Lidar devices 472, one or more cameras 474, one or more radars 476, one or more global positioning system (“GPS”) data receiver-sensors, one or more inertial measurement units (“IMUs”) 475, one or more odometry sensors 477 (e.g., wheel encoder sensors, wheel speed sensors, and the like), and any other suitable sensors 478, such as infrared cameras or sensors, hyperspectral-capable sensors, ultrasonic sensors (or any other acoustic energy-based sensor), radio frequency-based sensors, etc. In some cases, wheel angle sensors configured to sense steering angles of wheels may be included as odometry sensors 477 or suitable sensors 478. In a non-limiting example, autonomous vehicle controller 447 may include four or more Lidars 472, sixteen or more cameras 474 and four or more radar units 476. Further, sensors 470 may be configured to provide sensor data to components of autonomous vehicle controller 447 and to elements of autonomous vehicle service platform 401. As shown in diagram 400, autonomous vehicle controller 447 includes a planner 464, a motion controller 462, a localizer 468, a perception engine 466, and a local map generator 440. Note that elements depicted in diagram 400 of FIG. 4 may include structures and/or functions as similarly-named elements described in connection to one or more other drawings.

[0069] Localizer 468 is configured to localize autonomous vehicle (i.e., determine a local pose) relative to reference data, which may include map data, route data (e.g., road network data, such as RNDF-like data), and the like. In some cases, localizer 468 is configured to identify, for example, a point in space that may represent a location of autonomous vehicle 430 relative to features of a representation of an environment. Localizer 468 is shown to include a sensor data integrator 469, which may be configured to integrate multiple subsets of sensor data (e.g., of different sensor modalities) to reduce uncertainties related to each individual type of sensor. According to some examples, sensor data integrator 469 is configured to fuse sensor data (e.g., Lidar data, camera data, radar data, etc.) to form integrated sensor data values for determining a local pose. According to some examples, localizer 468 retrieves reference data originating from a reference data repository 405, which includes a map data repository 405a for storing 2D map data, 3D map data, 4D map data, and the like. Localizer 468 may be configured to identify at least a subset of features in the environment to match against map data to identify, or otherwise confirm, a pose of autonomous vehicle 430. According to some examples, localizer 468 may be configured to identify any amount of features in an environment, such that a set of features can one or more features, or all features. In a specific example, any amount of Lidar data (e.g., most or substantially all Lidar data) may be compared against data representing a map for purposes of localization. Generally, non-matched objects resulting from the comparison of the environment features and map data may be a dynamic object, such as a vehicle, bicyclist, pedestrian, etc. Note that detection of dynamic objects, including obstacles, may be performed with or without map data. In particular, dynamic objects may be detected and tracked independently of map data (i.e., in the absence of map data). In some instances, 2D map data and 3D map data may be viewed as “global map data” or map data that has been validated at a point in time by autonomous vehicle service platform 401. As map data in map data repository 405a may be updated and/or validated periodically, a deviation may exist between the map data and an actual environment in which the autonomous vehicle is positioned. Therefore, localizer 468 may retrieve locally-derived map data generated by local map generator 440 to enhance localization. Local map generator 440 is configured to generate local map data in real-time or near real-time. Optionally, local map generator 440 may receive static and dynamic object map data to enhance the accuracy of locally generated maps by, for example, disregarding dynamic objects in localization. According to at least some embodiments, local map generator 440 may be integrated with, or formed as part of, localizer 468. In at least one case, local map generator 440, either individually or in collaboration with localizer 468, may be configured to generate map and/or reference data based on simultaneous localization and mapping (“SLAM”) or the like. Note that localizer 468 may implement a “hybrid” approach to using map data, whereby logic in localizer 468 may be configured to select various amounts of map data from either map data repository 405a or local map data from local map generator 440, depending on the degrees of reliability of each source of map data. Therefore, localizer 468 may still use out-of-date map data in view of locally-generated map data.

[0070] Perception engine 466 is configured to, for example, assist planner 464 in planning routes and generating trajectories by identifying objects of interest in a surrounding environment in which autonomous vehicle 430 is transiting. Further, probabilities may be associated with each of the object of interest, whereby a probability may represent a likelihood that an object of interest may be a threat to safe travel (e.g., a fast-moving motorcycle may require enhanced tracking rather than a person sitting at a bus stop bench while reading a newspaper). As shown, perception engine 466 includes an object detector 442 and an object classifier 444. Object detector 442 is configured to distinguish objects relative to other features in the environment, and object classifier 444 may be configured to classify objects as either dynamic or static objects and track the locations of the dynamic and the static objects relative to autonomous vehicle 430 for planning purposes. Further, perception engine 466 may be configured to assign an identifier to a static or dynamic object that specifies whether the object is (or has the potential to become) an obstacle that may impact path planning at planner 464. Although not shown in FIG. 4, note that perception engine 466 may also perform other perception-related functions, such as segmentation and tracking, examples of which are described below.

[0071] Planner 464 is configured to generate a number of candidate trajectories for accomplishing a goal to reaching a destination via a number of paths or routes that are available. Trajectory evaluator 465 is configured to evaluate candidate trajectories and identify which subsets of candidate trajectories are associated with higher degrees of confidence levels of providing collision-free paths to the destination. As such, trajectory evaluator 465 can select an optimal trajectory based on relevant criteria for causing commands to generate control signals for vehicle components 450 (e.g., actuators or other mechanisms). Note that the relevant criteria may include any number of factors that define optimal trajectories, the selection of which need not be limited to reducing collisions. For example, the selection of trajectories may be made to optimize user experience (e.g., user comfort) as well as collision-free trajectories that comply with traffic regulations and laws. User experience may be optimized by moderating accelerations in various linear and angular directions (e.g., to reduce jerking-like travel or other unpleasant motion). In some cases, at least a portion of the relevant criteria can specify which of the other criteria to override or supersede, while maintain optimized, collision-free travel. For example, legal restrictions may be temporarily lifted or deemphasized when generating trajectories in limited situations (e.g., crossing double yellow lines to go around a cyclist or travelling at higher speeds than the posted speed limit to match traffic flows). As such, the control signals are configured to cause propulsion and directional changes at the drivetrain and/or wheels. In this example, motion controller 462 is configured to transform commands into control signals (e.g., velocity, wheel angles, etc.) for controlling the mobility of autonomous vehicle 430. In the event that trajectory evaluator 465 has insufficient information to ensure a confidence level high enough to provide collision-free, optimized travel, planner 464 can generate a request to teleoperator 404 for teleoperator support.

[0072] Autonomous vehicle service platform 401 includes teleoperator 404 (e.g., a teleoperator computing device), reference data repository 405, a map updater 406, a vehicle data controller 408, a calibrator 409, and an off-line object classifier 410. Note that each element of autonomous vehicle service platform 401 may be independently located or distributed and in communication with other elements in autonomous vehicle service platform 401. Further, element of autonomous vehicle service platform 401 may independently communicate with the autonomous vehicle 430 via the communication layer 402. Map updater 406 is configured to receive map data (e.g., from local map generator 440, sensors 460, or any other component of autonomous vehicle controller 447), and is further configured to detect deviations, for example, of map data in map data repository 405a from a locally-generated map. Vehicle data controller 408 can cause map updater 406 to update reference data within repository 405 and facilitate updates to 2D, 3D, and/or 4D map data. In some cases, vehicle data controller 408 can control the rate at which local map data is received into autonomous vehicle service platform 408 as well as the frequency at which map updater 406 performs updating of the map data.

[0073] Calibrator 409 is configured to perform calibration of various sensors of the same or different types. Calibrator 409 may be configured to determine the relative poses of the sensors (e.g., in Cartesian space (x, y, z)) and orientations of the sensors (e.g., roll, pitch and yaw). The pose and orientation of a sensor, such a camera, Lidar sensor, radar sensor, etc., may be calibrated relative to other sensors, as well as globally relative to the vehicle's reference frame. Off-line self-calibration can also calibrate or estimate other parameters, such as vehicle inertial tensor, wheel base, wheel radius or surface road friction. Calibration can also be done online to detect parameter change, according to some examples. Note, too, that calibration by calibrator 409 may include intrinsic parameters of the sensors (e.g., optical distortion, beam angles, etc.) and extrinsic parameters. In some cases, calibrator 409 may be performed by maximizing a correlation between depth discontinuities in 3D laser data and edges of image data, as an example. Off-line object classification 410 is configured to receive data, such as sensor data, from sensors 470 or any other component of autonomous vehicle controller 447. According to some embodiments, an off-line classification pipeline of off-line object classification 410 may be configured to pre-collect and annotate objects (e.g., manually by a human and/or automatically using an offline labeling algorithm), and may further be configured to train an online classifier (e.g., object classifier 444), which can provide real-time classification of object types during online autonomous operation.

[0074] FIG. 5 is an example of a flow diagram to control an autonomous vehicle, according to some embodiments. At 502, flow 500 begins when sensor data originating from sensors of multiple modalities at an autonomous vehicle is received, for example, by an autonomous vehicle controller. One or more subsets of sensor data may be integrated for generating fused data to improve, for example, estimates. In some examples, a sensor stream of one or more sensors (e.g., of same or different modalities) may be fused to form fused sensor data at 504. In some examples, subsets of Lidar sensor data and camera sensor data may be fused at 504 to facilitate localization. At 506, data representing objects based on the least two subsets of sensor data may be derived at a processor. For example, data identifying static objects or dynamic objects may be derived (e.g., at a perception engine) from at least Lidar and camera data. At 508, a detected object is determined to affect a planned path, and a subset of trajectories are evaluated (e.g., at a planner) responsive to the detected object at 510. A confidence level is determined at 512 to exceed a range of acceptable confidence levels associated with normative operation of an autonomous vehicle. Therefore, in this case, a confidence level may be such that a certainty of selecting an optimized path is less likely, whereby an optimized path may be determined as a function of the probability of facilitating collision-free travel, complying with traffic laws, providing a comfortable user experience (e.g., comfortable ride), and/or generating candidate trajectories on any other factor. As such, a request for an alternate path may be transmitted to a teleoperator computing device at 514. Thereafter, the teleoperator computing device may provide a planner with an optimal trajectory over which an autonomous vehicle made travel. In situations, the vehicle may also determine that executing a safe-stop maneuver is the best course of action (e.g., safely and automatically causing an autonomous vehicle to a stop at a location of relatively low probabilities of danger). Note that the order depicted in this and other flow charts herein are not intended to imply a requirement to linearly perform various functions as each portion of a flow chart may be performed serially or in parallel with any one or more other portions of the flow chart, as well as independent or dependent on other portions of the flow chart.

[0075] FIG. 6 is a diagram depicting an example of an architecture for an autonomous vehicle controller, according to some embodiments. Diagram 600 depicts a number of processes including a motion controller process 662, a planner processor 664, a perception process 666, a mapping process 640, and a localization process 668, some of which may generate or receive data relative to other processes. Other processes, such as such as processes 670 and 650 may facilitate interactions with one or more mechanical components of an autonomous vehicle. For example, perception process 666, mapping process 640, and localization process 668 are configured to receive sensor data from sensors 670, whereas planner process 664 and perception process 666 are configured to receive guidance data 606, which may include route data, such as road network data. Further to diagram 600, localization process 668 is configured to receive map data 605a (i.e., 2D map data), map data 605b (i.e., 3D map data), and local map data 642, among other types of map data. For example, localization process 668 may also receive other forms of map data, such as 4D map data, which may include, for example, an epoch determination. Localization process 668 is configured to generate local position data 641 representing a local pose. Local position data 641 is provided to motion controller process 662, planner process 664, and perception process 666. Perception process 666 is configured to generate static and dynamic object map data 667, which, in turn, may be transmitted to planner process 664. In some examples, static and dynamic object map data 667 may be transmitted with other data, such as semantic classification information and predicted object behavior. Planner process 664 is configured to generate trajectories data 665, which describes a number of trajectories generated by planner 664. Motion controller process uses trajectories data 665 to generate low-level commands or control signals for application to actuators 650 to cause changes in steering angles and/or velocity.

[0076] FIG. 7 is a diagram depicting an example of an autonomous vehicle service platform implementing redundant communication channels to maintain reliable communications with a fleet of autonomous vehicles, according to some embodiments. Diagram 700 depicts an autonomous vehicle service platform 701 including a reference data generator 705, a vehicle data controller 702, an autonomous vehicle fleet manager 703, a teleoperator manager 707, a simulator 740, and a policy manager 742. Reference data generator 705 is configured to generate and modify map data and route data (e.g., RNDF data). Further, reference data generator 705 may be configured to access 2D maps in 2D map data repository 720, access 3D maps in 3D map data repository 722, and access route data in route data repository 724. Other map representation data and repositories may be implemented in some examples, such as 4D map data including Epoch Determination. Vehicle data controller 702 may be configured to perform a variety of operations. For example, vehicle data controller 702 may be configured to change a rate that data is exchanged between a fleet of autonomous vehicles and platform 701 based on quality levels of communication over channels 770. During bandwidth-constrained periods, for example, data communications may be prioritized such that teleoperation requests from autonomous vehicle 730 are prioritized highly to ensure delivery. Further, variable levels of data abstraction may be transmitted per vehicle over channels 770, depending on bandwidth available for a particular channel. For example, in the presence of a robust network connection, full Lidar data (e.g., substantially all Lidar data, but also may be less) may be transmitted, whereas in the presence of a degraded or low-speed connection, simpler or more abstract depictions of the data may be transmitted (e.g., bounding boxes with associated metadata, etc.). Autonomous vehicle fleet manager 703 is configured to coordinate the dispatching of autonomous vehicles 730 to optimize multiple variables, including an efficient use of battery power, times of travel, whether or not an air-conditioning unit in an autonomous vehicle 730 may be used during low charge states of a battery, etc., any or all of which may be monitored in view of optimizing cost functions associated with operating an autonomous vehicle service. An algorithm may be implemented to analyze a variety of variables with which to minimize costs or times of travel for a fleet of autonomous vehicles. Further, autonomous vehicle fleet manager 703 maintains an inventory of autonomous vehicles as well as parts for accommodating a service schedule in view of maximizing up-time of the fleet.

[0077] Teleoperator manager 707 is configured to manage a number of teleoperator computing devices 704 with which teleoperators 708 provide input. Simulator 740 is configured to simulate operation of one or more autonomous vehicles 730, as well as the interactions between teleoperator manager 707 and an autonomous vehicle 730. Simulator 740 may also simulate operation of a number of sensors (including the introduction of simulated noise) disposed in autonomous vehicle 730. Further, an environment, such as a city, may be simulated such that a simulated autonomous vehicle can be introduced to the synthetic environment, whereby simulated sensors may receive simulated sensor data, such as simulated laser returns. Simulator 740 may provide other functions as well, including validating software updates and/or map data. Policy manager 742 is configured to maintain data representing policies or rules by which an autonomous vehicle ought to behave in view of a variety of conditions or events that an autonomous vehicle encounters while traveling in a network of roadways. In some cases, updated policies and/or rules may be simulated in simulator 740 to confirm safe operation of a fleet of autonomous vehicles in view of changes to a policy. Some of the above-described elements of autonomous vehicle service platform 701 are further described hereinafter.

[0078] Communication channels 770 are configured to provide networked communication links among a fleet of autonomous vehicles 730 and autonomous vehicle service platform 701. For example, communication channel 770 includes a number of different types of networks 771, 772, 773, and 774, with corresponding subnetworks (e.g., 771a to 771n), to ensure a certain level of redundancy for operating an autonomous vehicle service reliably. For example, the different types of networks in communication channels 770 may include different cellular network providers, different types of data networks, etc., to ensure sufficient bandwidth in the event of reduced or lost communications due to outages in one or more networks 771, 772, 773, and 774.

[0079] FIG. 8 is a diagram depicting an example of a messaging application configured to exchange data among various applications, according to some embodiments. Diagram 800 depicts an teleoperator application 801 disposed in a teleoperator manager, and an autonomous vehicle application 830 disposed in an autonomous vehicle, whereby teleoperator applications 801 and autonomous vehicle application 830 exchange message data via a protocol that facilitates communications over a variety of networks, such as network 871, 872, and other networks 873. According to some examples, the communication protocol is a middleware protocol implemented as a Data Distribution Service™ having a specification maintained by the Object Management Group consortium. In accordance with the communications protocol, teleoperator application 801 and autonomous vehicle application 830 may include a message router 854 disposed in a message domain, the message router being configured to interface with the teleoperator API 852. In some examples, message router 854 is a routing service. In some examples, message domain 850a in teleoperator application 801 may be identified by a teleoperator identifier, whereas message domain 850b be may be identified as a domain associated with a vehicle identifier. Teleoperator API 852 in teleoperator application 801 is configured to interface with teleoperator processes 803a to 803c, whereby teleoperator process 803b is associated with an autonomous vehicle identifier 804, and teleoperator process 803c is associated with an event identifier 806 (e.g., an identifier that specifies an intersection that may be problematic for collision-free path planning). Teleoperator API 852 in autonomous vehicle application 830 is configured to interface with an autonomous vehicle operating system 840, which includes sensing application 842, a perception application 844, a localization application 846, and a control application 848. In view of the foregoing, the above-described communications protocol may facilitate data exchanges to facilitate teleoperations as described herein. Further, the above-described communications protocol may be adapted to provide secure data exchanges among one or more autonomous vehicles and one or more autonomous vehicle service platforms. For example, message routers 854 may be configured to encrypt and decrypt messages to provide for secured interactions between, for example, a teleoperator process 803 and an autonomous vehicle operation system 840.

[0080] FIG. 9 is a diagram depicting types of data for facilitating teleoperations using a communications protocol described in FIG. 8, according to some examples. Diagram 900 depicts a teleoperator 908 interfacing with a teleoperator computing device 904 coupled to a teleoperator application 901, which is configured to exchange data via a data-centric messaging bus 972 implemented in one or more networks 971. Data-centric messaging bus 972 provides a communication link between teleoperator application 901 and autonomous vehicle application 930. Teleoperator API 962 of teleoperator application 901 is configured to receive message service configuration data 964 and route data 960, such as road network data (e.g., RNDF-like data), mission data (e.g., MDF-data), and the like. Similarly, a messaging service bridge 932 is also configured to receive messaging service configuration data 934. Messaging service configuration data 934 and 964 provide configuration data to configure the messaging service between teleoperator application 901 and autonomous vehicle application 930. An example of messaging service configuration data 934 and 964 includes quality of service (“QoS”) configuration data implemented to configure a Data Distribution Service™ application.

[0081] An example of a data exchange for facilitating teleoperations via the communications protocol is described as follows. Consider that obstacle data 920 is generated by a perception system of an autonomous vehicle controller. Further, planner options data 924 is generated by a planner to notify a teleoperator of a subset of candidate trajectories, and position data 926 is generated by the localizer. Obstacle data 920, planner options data 924, and position data 926 are transmitted to a messaging service bridge 932, which, in accordance with message service configuration data 934, generates telemetry data 940 and query data 942, both of which are transmitted via data-centric messaging bus 972 into teleoperator application 901 as telemetry data 950 and query data 952. Teleoperator API 962 receives telemetry data 950 and inquiry data 952, which, in turn are processed in view of Route data 960 and message service configuration data 964. The resultant data is subsequently presented to a teleoperator 908 via teleoperator computing device 904 and/or a collaborative display (e.g., a dashboard display visible to a group of collaborating teleoperators 908). Teleoperator 908 reviews the candidate trajectory options that are presented on the display of teleoperator computing device 904, and selects a guided trajectory, which generates command data 982 and query response data 980, both of which are passed through teleoperator API 962 as query response data 954 and command data 956. In turn, query response data 954 and command data 956 are transmitted via data-centric messaging bus 972 into autonomous vehicle application 930 as query response data 944 and command data 946. Messaging service bridge 932 receives query response data 944 and command data 946 and generates teleoperator command data 928, which is configured to generate a teleoperator-selected trajectory for implementation by a planner. Note that the above-described messaging processes are not intended to be limiting, and other messaging protocols may be implemented as well.

[0082] FIG. 10 is a diagram illustrating an example of a teleoperator interface with which a teleoperator may influence path planning, according to some embodiments. Diagram 1000 depicts examples of an autonomous vehicle 1030 in communication with an autonomous vehicle service platform 1001, which includes a teleoperator manager 1007 configured to facilitate teleoperations. In a first example, teleoperator manager 1007 receives data that requires teleoperator 1008 to preemptively view a path of an autonomous vehicle approaching a potential obstacle or an area of low planner confidence levels so that teleoperator 1008 may be able to address an issue in advance. To illustrate, consider that an intersection that an autonomous vehicle is approaching may be tagged as being problematic. As such, user interface 1010 displays a representation 1014 of a corresponding autonomous vehicle 1030 transiting along a path 1012, which has been predicted by a number of trajectories generated by a planner. Also displayed are other vehicles 1011 and dynamic objects 1013, such as pedestrians, that may cause sufficient confusion at the planner, thereby requiring teleoperation support. User interface 1010 also presents to teleoperator 1008 a current velocity 1022, a speed limit 1024, and an amount of charge 1026 presently in the batteries. According to some examples, user interface 1010 may display other data, such as sensor data as acquired from autonomous vehicle 1030. In a second example, consider that planner 1064 has generated a number of trajectories that are coextensive with a planner-generated path 1044 regardless of a detected unidentified object 1046. Planner 1064 may also generate a subset of candidate trajectories 1040, but in this example, the planner is unable to proceed given present confidence levels. If planner 1064 fails to determine an alternative path, a teleoperation request may be transmitted. In this case, a teleoperator may select one of candidate trajectories 1040 to facilitate travel by autonomous vehicle 1030 that is consistent with teleoperator-based path 1042.

[0083] FIG. 11 is a diagram depicting an example of a planner configured to invoke teleoperations, according to some examples. Diagram 1100 depicts a planner 1164 including a topography manager 1110, a route manager 1112, a path generator 1114, a trajectory evaluator 1120, and a trajectory tracker 1128. Topography manager 1110 is configured to receive map data, such as 3D map data or other like map data that specifies topographic features. Topography manager 1110 is further configured to identify candidate paths based on topographic-related features on a path to a destination. According to various examples, topography manager 1110 receives 3D maps generated by sensors associated with one or more autonomous vehicles in the fleet. Route manager 1112 is configured to receive environmental data 1103, which may include traffic-related information associated with one or more routes that may be selected as a path to the destination. Path generator 1114 receives data from topography manager 1110 and route manager 1112, and generates one or more paths or path segments suitable to direct autonomous vehicle toward a destination. Data representing one or more paths or path segments is transmitted into trajectory evaluator 1120.

[0084] Trajectory evaluator 1120 includes a state and event manager 1122, which, in turn, may include a confidence level generator 1123. Trajectory evaluator 1120 further includes a guided trajectory generator 1126 and a trajectory generator 1124. Further, planner 1164 is configured to receive policy data 1130, perception engine data 1132, and localizer data 1134.

[0085] Policy data 1130 may include criteria with which planner 1164 uses to determine a path that has a sufficient confidence level with which to generate trajectories, according to some examples. Examples of policy data 1130 include policies that specify that trajectory generation is bounded by stand-off distances to external objects (e.g., maintaining a safety buffer of 3 feet from a cyclist, as possible), or policies that require that trajectories must not cross a center double yellow line, or policies that require trajectories to be limited to a single lane in a 4-lane roadway (e.g., based on past events, such as typically congregating at a lane closest to a bus stop), and any other similar criteria specified by policies. Perception engine data 1132 includes maps of locations of static objects and dynamic objects of interest, and localizer data 1134 includes at least a local pose or position.

[0086] State and event manager 1122 may be configured to probabilistically determine a state of operation for an autonomous vehicle. For example, a first state of operation (i.e., “normative operation”) may describe a situation in which trajectories are collision-free, whereas a second state of operation (i.e., “non-normative operation”) may describe another situation in which the confidence level associated with possible trajectories are insufficient to guarantee collision-free travel. According to some examples, state and event manager 1122 is configured to use perception data 1132 to determine a state of autonomous vehicle that is either normative or non-normative. Confidence level generator 1123 may be configured to analyze perception data 1132 to determine a state for the autonomous vehicle. For example, confidence level generator 1123 may use semantic information associated with static and dynamic objects, as well as associated probabilistic estimations, to enhance a degree of certainty that planner 1164 is determining safe course of action. For example, planner 1164 may use perception engine data 1132 that specifies a probability that an object is either a person or not a person to determine whether planner 1164 is operating safely (e.g., planner 1164 may receive a degree of certainty that an object has a 98% probability of being a person, and a probability of 2% that the object is not a person).

[0087] Upon determining a confidence level (e.g., based on statistics and probabilistic determinations) is below a threshold required for predicted safe operation, a relatively low confidence level (e.g., single probability score) may trigger planner 1164 to transmit a request 1135 for teleoperation support to autonomous vehicle service platform 1101. In some cases, telemetry data and a set of candidate trajectories may accompany the request. Examples of telemetry data include sensor data, localization data, perception data, and the like. A teleoperator 1108 may transmit via teleoperator computing device 1104 a selected trajectory 1137 to guided trajectory generator 1126. As such, selected trajectory 1137 is a trajectory formed with guidance from a teleoperator. Upon confirming there is no change in the state (e.g., a non-normative state is pending), guided trajectory generator 1126 passes data to trajectory generator 1124, which, in turn, causes trajectory tracker 1128, as a trajectory tracking controller, to use the teleop-specified trajectory for generating control signals 1170 (e.g., steering angles, velocity, etc.). Note that planner 1164 may trigger transmission of a request 1135 for teleoperation support prior to a state transitioning to a non-normative state. In particular, an autonomous vehicle controller and/or its components can predict that a distant obstacle may be problematic and preemptively cause planner 1164 to invoke teleoperations prior to the autonomous vehicle reaching the obstacle. Otherwise, the autonomous vehicle may cause a delay by transitioning to a safe state upon encountering the obstacle or scenario (e.g., pulling over and off the roadway). In another example, teleoperations may be automatically invoked prior to an autonomous vehicle approaching a particular location that is known to be difficult to navigate. This determination may optionally take into consideration other factors, including the time of day, the position of the sun, if such situation is likely to cause a disturbance to the reliability of sensor readings, and traffic or accident data derived from a variety of sources.

[0088] FIG. 12 is an example of a flow diagram configured to control an autonomous vehicle, according to some embodiments. At 1202, flow 1200 begins. Data representing a subset of objects that are received at a planner in an autonomous vehicle, the subset of objects including at least one object associated with data representing a degree of certainty for a classification type. For example, perception engine data may include metadata associated with objects, whereby the metadata specifies a degree of certainty associated with a specific classification type. For instance, a dynamic object may be classified as a “young pedestrian” with an 85% confidence level of being correct. At 1204, localizer data may be received (e.g., at a planner). The localizer data may include map data that is generated locally within the autonomous vehicle. The local map data may specify a degree of certainty (including a degree of uncertainty) that an event at a geographic region may occur. An event may be a condition or situation affecting operation, or potentially affecting operation, of an autonomous vehicle. The events may be internal (e.g., failed or impaired sensor) to an autonomous vehicle, or external (e.g., roadway obstruction). Examples of events are described herein, such as in FIG. 2 as well as in other figures and passages. A path coextensive with the geographic region of interest may be determined at 1206. For example, consider that the event is the positioning of the sun in the sky at a time of day in which the intensity of sunlight impairs the vision of drivers during rush hour traffic. As such, it is expected or predicted that traffic may slow down responsive to the bright sunlight. Accordingly, a planner may preemptively invoke teleoperations if an alternate path to avoid the event is less likely. At 1208, a local position is determined at a planner based on local pose data. At 1210, a state of operation of an autonomous vehicle may be determined (e.g., probabilistically), for example, based on a degree of certainty for a classification type and a degree of certainty of the event, which is may be based on any number of factors, such as speed, position, and other state information. To illustrate, consider an example in which a young pedestrian is detected by the autonomous vehicle during the event in which other drivers' vision likely will be impaired by the sun, thereby causing an unsafe situation for the young pedestrian. Therefore, a relatively unsafe situation can be detected as a probabilistic event that may be likely to occur (i.e., an unsafe situation for which teleoperations may be invoked). At 1212, a likelihood that the state of operation is in a normative state is determined, and based on the determination, a message is transmitted to a teleoperator computing device requesting teleoperations to preempt a transition to a next state of operation (e.g., preempt transition from a normative to non-normative state of operation, such as an unsafe state of operation).

[0089] FIG. 13 depicts an example in which a planner may generate a trajectory, according to some examples. Diagram 1300 includes a trajectory evaluator 1320 and a trajectory generator 1324. Trajectory evaluator 1320 includes a confidence level generator 1322 and a teleoperator query messenger 1329. As shown, trajectory evaluator 1320 is coupled to a perception engine 1366 to receive static map data 1301, and current and predicted object state data 1303. Trajectory evaluator 1320 also receives local pose data 1305 from localizer 1368 and plan data 1307 from a global planner 1369. In one state of operation (e.g., non-normative), confidence level generator 1322 receives static map data 1301 and current and predicted object state data 1303. Based on this data, confidence level generator 1322 may determine that detected trajectories are associated with unacceptable confidence level values. As such, confidence level generator 1322 transmits detected trajectory data 1309 (e.g., data including candidate trajectories) to notify a teleoperator via teleoperator query messenger 1329, which, in turn, transmits a request 1370 for teleoperator assistance.

[0090] In another state of operation (e.g., a normative state), static map data 1301, current and predicted object state data 1303, local pose data 1305, and plan data 1307 (e.g., global plan data) are received into trajectory calculator 1325, which is configured to calculate (e.g., iteratively) trajectories to determine an optimal one or more paths. Next, at least one path is selected and is transmitted as selected path data 1311. According to some embodiments, trajectory calculator 1325 is configured to implement re-planning of trajectories as an example. Nominal driving trajectory generator 1327 is configured to generate trajectories in a refined approach, such as by generating trajectories based on receding horizon control techniques. Nominal driving trajectory generator 1327 subsequently may transmit nominal driving trajectory path data 1372 to, for example, a trajectory tracker or a vehicle controller to implement physical changes in steering, acceleration, and other components.

[0091] FIG. 14 is a diagram depicting another example of an autonomous vehicle service platform, according to some embodiments. Diagram 1400 depicts an autonomous vehicle service platform 1401 including a teleoperator manager 1407 that is configured to manage interactions and/or communications among teleoperators 1408, teleoperator computing devices 1404, and other components of autonomous vehicle service platform 1401. Further to diagram 1400, autonomous vehicle service platform 1401 includes a simulator 1440, a repository 1441, a policy manager 1442, a reference data updater 1438, a 2D map data repository 1420, a 3D map data repository 1422, and a route data repository 1424. Other map data, such as 4D map data (e.g., using epoch determination), may be implemented and stored in a repository (not shown).

[0092] Teleoperator action recommendation controller 1412 includes logic configured to receive and/or control a teleoperation service request via autonomous vehicle (“AV”) planner data 1472, which can include requests for teleoperator assistance as well as telemetry data and other data. As such, planner data 1472 may include recommended candidate trajectories or paths from which a teleoperator 1408 via teleoperator computing device 1404 may select. According to some examples, teleoperator action recommendation controller 1412 may be configured to access other sources of recommended candidate trajectories from which to select an optimum trajectory. For example, candidate trajectories contained in autonomous vehicle planner data 1472 may, in parallel, be introduced into simulator 1440, which is configured to simulate an event or condition being experienced by an autonomous vehicle requesting teleoperator assistance. Simulator 1440 can access map data and other data necessary for performing a simulation on the set of candidate trajectories, whereby simulator 1440 need not exhaustively reiterate simulations to confirm sufficiency. Rather, simulator 1440 may provide either confirm the appropriateness of the candidate trajectories, or may otherwise alert a teleoperator to be cautious in their selection.

[0093] Teleoperator interaction capture analyzer 1416 may be configured to capture numerous amounts of teleoperator transactions or interactions for storage in repository 1441, which, for example, may accumulate data relating to a number of teleoperator transactions for analysis and generation of policies, at least in some cases. According to some embodiments, repository 1441 may also be configured to store policy data for access by policy manager 1442. Further, teleoperator interaction capture analyzer 1416 may apply machine learning techniques to empirically determine how best to respond to events or conditions causing requests for teleoperation assistance. In some cases, policy manager 1442 may be configured to update a particular policy or generate a new policy responsive to analyzing the large set of teleoperator interactions (e.g., subsequent to applying machine learning techniques). Policy manager 1442 manages policies that may be viewed as rules or guidelines with which an autonomous vehicle controller and its components operate under to comply with autonomous operations of a vehicle. In some cases, a modified or updated policy may be applied to simulator 1440 to confirm the efficacy of permanently releasing or implementing such policy changes.

[0094] Simulator interface controller 1414 is configured to provide an interface between simulator 1440 and teleoperator computing devices 1404. For example, consider that sensor data from a fleet of autonomous vehicles is applied to reference data updater 1438 via autonomous (“AV”) fleet data 1470, whereby reference data updater 1438 is configured to generate updated map and route data 1439. In some implementations, updated map and route data 1439 may be preliminarily released as an update to data in map data repositories 1420 and 1422, or as an update to data in route data repository 1424. In this case, such data may be tagged as being a “beta version” in which a lower threshold for requesting teleoperator service may be implemented when, for example, a map tile including preliminarily updated information is used by an autonomous vehicle. Further, updated map and route data 1439 may be introduced to simulator 1440 for validating the updated map data. Upon full release (e.g., at the close of beta testing), the previously lowered threshold for requesting a teleoperator service related to map tiles is canceled. User interface graphics controller 1410 provides rich graphics to teleoperators 1408, whereby a fleet of autonomous vehicles may be simulated within simulator 1440 and may be accessed via teleoperator computing device 1404 as if the simulated fleet of autonomous vehicles were real.

[0095] FIG. 15 is an example of a flow diagram to control an autonomous vehicle, according to some embodiments. At 1502, flow 1500 begins. Message data may be received at a teleoperator computing device for managing a fleet of autonomous vehicles. The message data may indicate event attributes associated with a non-normative state of operation in the context of a planned path for an autonomous vehicle. For example, an event may be characterized as a particular intersection that becomes problematic due to, for example, a large number of pedestrians, hurriedly crossing the street against a traffic light. The event attributes describe the characteristics of the event, such as, for example, the number of people crossing the street, the traffic delays resulting from an increased number of pedestrians, etc. At 1504, a teleoperation repository may be accessed to retrieve a first subset of recommendations based on simulated operations of aggregated data associated with a group of autonomous vehicles. In this case, a simulator may be a source of recommendations with which a teleoperator may implement. Further, the teleoperation repository may also be accessed to retrieve a second subset of recommendations based on an aggregation of teleoperator interactions responsive to similar event attributes. In particular, a teleoperator interaction capture analyzer may apply machine learning techniques to empirically determine how best to respond to events having similar attributes based on previous requests for teleoperation assistance. At 1506, the first subset and the second subset of recommendations are combined to form a set of recommended courses of action for the autonomous vehicle. At 1508, representations of the set of recommended courses of actions may be presented visually on a display of a teleoperator computing device. At 1510, data signals representing a selection (e.g., by teleoperator) of a recommended course of action may be detected.

[0096] FIG. 16 is a diagram of an example of an autonomous vehicle fleet manager implementing a fleet optimization manager, according to some examples. Diagram 1600 depicts an autonomous vehicle fleet manager that is configured to manage a fleet of autonomous vehicles 1630 transiting within a road network 1650. Autonomous vehicle fleet manager 1603 is coupled to a teleoperator 1608 via a teleoperator computing device 1604, and is also coupled to a fleet management data repository 1646. Autonomous vehicle fleet manager 1603 is configured to receive policy data 1602 and environmental data 1606, as well as other data. Further to diagram 1600, fleet optimization manager 1620 is shown to include a transit request processor 1631, which, in turn, includes a fleet data extractor 1632 and an autonomous vehicle dispatch optimization calculator 1634. Transit request processor 1631 is configured to process transit requests, such as from a user 1688 who is requesting autonomous vehicle service. Fleet data extractor 1632 is configured to extract data relating to autonomous vehicles in the fleet. Data associated with each autonomous vehicle is stored in repository 1646. For example, data for each vehicle may describe maintenance issues, scheduled service calls, daily usage, battery charge and discharge rates, and any other data, which may be updated in real-time, may be used for purposes of optimizing a fleet of autonomous vehicles to minimize downtime. Autonomous vehicle dispatch optimization calculator 1634 is configured to analyze the extracted data and calculate optimized usage of the fleet so as to ensure that the next vehicle dispatched, such as from station 1652, provides for the least travel times and/or costs—in the aggregate—for the autonomous vehicle service.

[0097] Fleet optimization manager 1620 is shown to include a hybrid autonomous vehicle/non-autonomous vehicle processor 1640, which, in turn, includes an AV/non-AV optimization calculator 1642 and a non-AV selector 1644. According to some examples, hybrid autonomous vehicle/non-autonomous vehicle processor 1640 is configured to manage a hybrid fleet of autonomous vehicles and human-driven vehicles (e.g., as independent contractors). As such, autonomous vehicle service may employ non-autonomous vehicles to meet excess demand, or in areas, such as non-AV service region 1690, that may be beyond a geo-fence or in areas of poor communication coverage. AV/non-AV optimization calculator 1642 is configured to optimize usage of the fleet of autonomous and to invite non-AV drivers into the transportation service (e.g., with minimal or no detriment to the autonomous vehicle service). Non-AV selector 1644 includes logic for selecting a number of non-AV drivers to assist based on calculations derived by AV/non-AV optimization calculator 1642.

[0098] FIG. 17 is an example of a flow diagram to manage a fleet of autonomous vehicles, according to some embodiments. At 1702, flow 1700 begins. At 1702, policy data is received. The policy data may include parameters that define how best apply to select an autonomous vehicle for servicing a transit request. At 1704, fleet management data from a repository may be extracted. The fleet management data includes subsets of data for a pool of autonomous vehicles (e.g., the data describes the readiness of vehicles to service a transportation request). At 1706, data representing a transit request is received. For exemplary purposes, the transit request could be for transportation from a first geographic location to a second geographic location. At 1708, attributes based on the policy data are calculated to determine a subset of autonomous vehicles that are available to service the request. For example, attributes may include a battery charge level and time until next scheduled maintenance. At 1710, an autonomous vehicle is selected as transportation from the first geographic location to the second geographic location, and data is generated to dispatch the autonomous vehicle to a third geographic location associated with the origination of the transit request.

[0099] FIG. 18 is a diagram illustrating an autonomous vehicle fleet manager implementing an autonomous vehicle communications link manager, according to some embodiments. Diagram 1800 depicts an autonomous vehicle fleet manager that is configured to manage a fleet of autonomous vehicles 1830 transiting within a road network 1850 that coincides with a communication outage at an area identified as “reduced communication region” 1880. Autonomous vehicle fleet manager 1803 is coupled to a teleoperator 1808 via a teleoperator computing device 1804. Autonomous vehicle fleet manager 1803 is configured to receive policy data 1802 and environmental data 1806, as well as other data. Further to diagram 1800, an autonomous vehicle communications link manager 1820 is shown to include an environment event detector 1831, a policy adaption determinator 1832, and a transit request processor 1834. Environment event detector 1831 is configured to receive environmental data 1806 specifying a change within the environment in which autonomous vehicle service is implemented. For example, environmental data 1806 may specify that region 1880 has degraded communication services, which may affect the autonomous vehicle service. Policy adaption determinator 1832 may specify parameters with which to apply when receiving transit requests during such an event (e.g., during a loss of communications). Transit request processor 1834 is configured to process transit requests in view of the degraded communications. In this example, a user 1888 is requesting autonomous vehicle service. Further, transit request processor 1834 includes logic to apply an adapted policy for modifying the way autonomous vehicles are dispatched so to avoid complications due to poor communications.

[0100] Communication event detector 1840 includes a policy download manager 1842 and communications-configured (“COMM-configured”) AV dispatcher 1844. Policy download manager 1842 is configured to provide autonomous vehicles 1830 an updated policy in view of reduced communications region 1880, whereby the updated policy may specify routes to quickly exit region 1880 if an autonomous vehicle enters that region. For example, autonomous vehicle 1864 may receive an updated policy moments before driving into region 1880. Upon loss of communications, autonomous vehicle 1864 implements the updated policy and selects route 1866 to drive out of region 1880 quickly. COMM-configured AV dispatcher 1844 may be configured to identify points 1865 at which to park autonomous vehicles that are configured as relays to establishing a peer-to-peer network over region 1880. As such, COMM-configured AV dispatcher 1844 is configured to dispatch autonomous vehicles 1862 (without passengers) to park at locations 1865 for the purposes of operating as communication towers in a peer-to-peer ad hoc network.

[0101] FIG. 19 is an example of a flow diagram to determine actions for autonomous vehicles during an event, such as degraded or lost communications, according to some embodiments. At 1901, flow 1900 begins. Policy data is received, whereby the policy data defines parameters with which to apply to transit requests in a geographical region during an event. At 1902, one or more of the following actions may be implemented: (1) dispatch a subset of autonomous vehicles to geographic locations in the portion of the geographic location, the subset of autonomous vehicles being configured to either park at specific geographic locations and each serve as a static communication relay, or transit in a geographic region to each serve as a mobile communication relay, (2) implement peer-to-peer communications among a portion of the pool of autonomous vehicles associated with the portion of the geographic region, (3) provide to the autonomous vehicles an event policy that describes a route to egress the portion of the geographic region during an event, (4) invoke teleoperations, and (5) recalculate paths so as to avoid the geographic portion. Subsequent to implementing the action, the fleet of autonomous vehicles is monitored at 1914.

[0102] FIG. 20 is a diagram depicting an example of a localizer, according to some embodiments. Diagram 2000 includes a localizer 2068 configured to receive sensor data from sensors 2070, such as Lidar data 2072, camera data 2074, radar data 2076, and other data 2078. Further, localizer 2068 is configured to receive reference data 2020, such as 2D map data 2022, 3D map data 2024, and 3D local map data. According to some examples, other map data, such as 4D map data 2025 and semantic map data (not shown), including corresponding data structures and repositories, may also be implemented. Further to diagram 2000, localizer 2068 includes a positioning system 2010 and a localization system 2012, both of which are configured to receive sensor data from sensors 2070 as well as reference data 2020. Localization data integrator 2014 is configured to receive data from positioning system 2010 and data from localization system 2012, whereby localization data integrator 2014 is configured to integrate or fuse sensor data from multiple sensors to form local pose data 2052.

[0103] FIG. 21 is an example of a flow diagram to generate local pose data based on integrated sensor data, according to some embodiments. At 2101, flow 2100 begins. At 2102, reference data is received, the reference data including three dimensional map data. In some examples, reference data, such as 3-D or 4-D map data, may be received via one or more networks. At 2104, localization data from one or more localization sensors is received and placed into a localization system. At 2106, positioning data from one or more positioning sensors is received into a positioning system. At 2108, the localization and positioning data are integrated. At 2110, the localization data and positioning data are integrated to form local position data specifying a geographic position of an autonomous vehicle.

[0104] FIG. 22 is a diagram depicting another example of a localizer, according to some embodiments. Diagram 2200 includes a localizer 2268, which, in turn, includes a localization system 2210 and a relative localization system 2212 to generate positioning-based data 2250 and local location-based data 2251, respectively. Localization system 2210 includes a projection processor 2254a for processing GPS data 2273, a GPS datum 2211, and 3D Map data 2222, among other optional data (e.g., 4D map data). Localization system 2210 also includes an odometry processor 2254b to process wheel data 2275 (e.g., wheel speed), vehicle model data 2213 and 3D map data 2222, among other optional data. Further yet, localization system 2210 includes an integrator processor 2254c to process IMU data 2257, vehicle model data 2215, and 3D map data 2222, among other optional data. Similarly, relative localization system 2212 includes a Lidar localization processor 2254d for processing Lidar data 2272, 2D tile map data 2220, 3D map data 2222, and 3D local map data 2223, among other optional data. Relative localization system 2212 also includes a visual registration processor 2254e to process camera data 2274, 3D map data 2222, and 3D local map data 2223, among other optional data. Further yet, relative localization system 2212 includes a radar return processor 2254f to process radar data 2276, 3D map data 2222, and 3D local map data 2223, among other optional data. Note that in various examples, other types of sensor data and sensors or processors may be implemented, such as sonar data and the like.

[0105] Further to diagram 2200, localization-based data 2250 and relative localization-based data 2251 may be fed into data integrator 2266a and localization data integrator 2266, respectively. Data integrator 2266a and localization data integrator 2266 may be configured to fuse corresponding data, whereby localization-based data 2250 may be fused at data integrator 2266a prior to being fused with relative localization-based data 2251 at localization data integrator 2266. According to some embodiments, data integrator 2266a is formed as part of localization data integrator 2266, or is absent. Regardless, a localization-based data 2250 and relative localization-based data 2251 can be both fed into localization data integrator 2266 for purposes of fusing data to generate local position data 2252. Localization-based data 2250 may include unary-constrained data (and uncertainty values) from projection processor 2254a, as well as binary-constrained data (and uncertainty values) from odometry processor 2254b and integrator processor 2254c. Relative localization-based data 2251 may include unary-constrained data (and uncertainty values) from localization processor 2254d and visual registration processor 2254e, and optionally from radar return processor 2254f. According to some embodiments, localization data integrator 2266 may implement non-linear smoothing functionality, such as a Kalman filter (e.g., a gated Kalman filter), a relative bundle adjuster, pose-graph relaxation, particle filter, histogram filter, or the like.

[0106] FIG. 23 is a diagram depicting an example of a perception engine, according to some embodiments. Diagram 2300 includes a perception engine 2366, which, in turn, includes a segmentation processor 2310, an object tracker 2330, and a classifier 2360. Further, perception engine 2366 is configured to receive a local position data 2352, Lidar data 2372, camera data 2374, and radar data 2376, for example. Note that other sensor data, such as sonar data, may be accessed to provide functionalities of perception engine 2366. Segmentation processor 2310 is configured to extract ground plane data and/or to segment portions of an image to distinguish objects from each other and from static imagery (e.g., background). In some cases, 3D blobs may be segmented to distinguish each other. In some examples, a blob may refer to a set of features that identify an object in a spatially-reproduced environment and may be composed of elements (e.g., pixels of camera data, points of laser return data, etc.) having similar characteristics, such as intensity and color. In some examples, a blob may also refer to a point cloud (e.g., composed of colored laser return data) or other elements constituting an object. Object tracker 2330 is configured to perform frame-to-frame estimations of motion for blobs, or other segmented image portions. Further, data association is used to associate a blob at one location in a first frame at time, t1, to a blob in a different position in a second frame at time, t2. In some examples, object tracker 2330 is configured to perform real-time probabilistic tracking of 3-D objects, such as blobs. Classifier 2360 is configured to identify an object and to classify that object by classification type (e.g., as a pedestrian, cyclist, etc.) and by energy/activity (e.g. whether the object is dynamic or static), whereby data representing classification is described by a semantic label. According to some embodiments, probabilistic estimations of object categories may be performed, such as classifying an object as a vehicle, bicyclist, pedestrian, etc. with varying confidences per object class. Perception engine 2366 is configured to determine perception engine data 2354, which may include static object maps and/or dynamic object maps, as well as semantic information so that, for example, a planner may use this information to enhance path planning. According to various examples, one or more of segmentation processor 2310, object tracker 2330, and classifier 2360 may apply machine learning techniques to generate perception engine data 2354.

[0107] FIG. 24 is an example of a flow chart to generate perception engine data, according to some embodiments. Flow chart 2400 begins at 2402, at which data representing a local position of an autonomous vehicle is retrieved. At 2404, localization data from one or more localization sensors is received, and features of an environment in which the autonomous vehicle is disposed are segmented at 2406 to form segmented objects. One or more portions of the segmented object are tracked spatially at 2408 to form at least one tracked object having a motion (e.g., an estimated motion). At 2410, a tracked object is classified at least as either being a static object or a dynamic object. In some cases, a static object or a dynamic object may be associated with a classification type. At 2412, data identifying a classified object is generated. For example, the data identifying the classified object may include semantic information.

[0108] FIG. 25 is an example of a segmentation processor, according to some embodiments. Diagram 2500 depicts a segmentation processor 2510 receiving Lidar data from one or more Lidars 2572 and camera image data from one or more cameras 2574. Local pose data 2552, Lidar data, and camera image data are received into meta spin generator 2521. In some examples, meta spin generator is configured to partition an image based on various attributes (e.g., color, intensity, etc.) into distinguishable regions (e.g., clusters or groups of a point cloud), at least two or more of which may be updated at the same time or about the same time. Meta spin data 2522 is used to perform object segmentation and ground segmentation at segmentation processor 2523, whereby both meta spin data 2522 and segmentation-related data from segmentation processor 2523 are applied to a scanned differencing processor 2513. Scanned differencing processor 2513 is configured to predict motion and/or relative velocity of segmented image portions, which can be used to identify dynamic objects at 2517. Data indicating objects with detected velocity at 2517 are optionally transmitted to the planner to enhance path planning decisions. Additionally, data from scanned differencing processor 2513 may be used to approximate locations of objects to form mapping of such objects (as well as optionally identifying a level of motion). In some examples, an occupancy grid map 2515 may be generated. Data representing an occupancy grid map 2515 may be transmitted to the planner to further enhance path planning decisions (e.g., by reducing uncertainties). Further to diagram 2500, image camera data from one or more cameras 2574 are used to classify blobs in blob classifier 2520, which also receives blob data 2524 from segmentation processor 2523. Segmentation processor 2510 also may receive raw radar returns data 2512 from one or more radars 2576 to perform segmentation at a radar segmentation processor 2514, which generates radar-related blob data 2516. Further to FIG. 25, segmentation processor 2510 may also receive and/or generate tracked blob data 2518 related to radar data. Blob data 2516, tracked blob data 2518, data from blob classifier 2520, and blob data 2524 may be used to track objects or portions thereof. According to some examples, one or more of the following may be optional: scanned differencing processor 2513, blob classification 2520, and data from radar 2576.

[0109] FIG. 26A is a diagram depicting examples of an object tracker and a classifier, according to various embodiments. Object tracker 2630 of diagram 2600 is configured to receive blob data 2516, tracked blob data 2518, data from blob classifier 2520, blob data 2524, and camera image data from one or more cameras 2676. Image tracker 2633 is configured to receive camera image data from one or more cameras 2676 to generate tracked image data, which, in turn, may be provided to data association processor 2632. As shown, data association processor 2632 is configured to receive blob data 2516, tracked blob data 2518, data from blob classifier 2520, blob data 2524, and track image data from image tracker 2633, and is further configured to identify one or more associations among the above-described types of data. Data association processor 2632 is configured to track, for example, various blob data from one frame to a next frame to, for example, estimate motion, among other things. Further, data generated by data association processor 2632 may be used by track updater 2634 to update one or more tracks, or tracked objects. In some examples, track updater 2634 may implement a Kalman Filter, or the like, to form updated data for tracked objects, which may be stored online in track database (“DB”) 2636. Feedback data may be exchanged via path 2699 between data association processor 2632 and track database 2636. In some examples, image tracker 2633 may be optional and may be excluded. Object tracker 2630 may also use other sensor data, such as radar or sonar, as well as any other types of sensor data, for example.

[0110] FIG. 26B is a diagram depicting another example of an object tracker according to at least some examples. Diagram 2601 includes an object tracker 2631 that may include structures and/or functions as similarly-named elements described in connection to one or more other drawings (e.g., FIG. 26A). As shown, object tracker 2631 includes an optional registration portion 2699 that includes a processor 2696 configured to perform object scan registration and data fusion. Processor 2696 is further configured to store the resultant data in 3D object database 2698.

[0111] Referring back to FIG. 26A, diagram 2600 also includes classifier 2660, which may include a track classification engine 2662 for generating static obstacle data 2672 and dynamic obstacle data 2674, both of which may be transmitted to the planner for path planning purposes. In at least one example, track classification engine 2662 is configured to determine whether an obstacle is static or dynamic, as well as another classification type for the object (e.g., whether the object is a vehicle, pedestrian, tree, cyclist, dog, cat, paper bag, etc.). Static obstacle data 2672 may be formed as part of an obstacle map (e.g., a 2D occupancy map), and dynamic obstacle data 2674 may be formed to include bounding boxes with data indicative of velocity and classification type. Dynamic obstacle data 2674, at least in some cases, includes 2D dynamic obstacle map data.

[0112] FIG. 27 is an example of front-end processor for a perception engine, according to some examples. Diagram 2700 includes a ground segmentation processor 2723a for performing ground segmentation, and an over segmentation processor 2723b for performing “over-segmentation,” according to various examples. Processors 2723a and 2723b are configured to receive optionally colored Lidar data 2775. Over segmentation processor 2723b generates data 2710 of a first blob type (e.g., a relatively small blob), which is provided to an aggregation classification and segmentation engine 2712 that generates data 2714 of a second blob type. Data 2714 is provided to data association processor 2732, which is configured to detect whether data 2714 resides in track database 2736. A determination is made at 2740 whether data 2714 of the second blob type (e.g., a relatively large blob, which may include one or more smaller blobs) is a new track. If so, a track is initialized at 2742, otherwise, the tracked object data stored in track database 2736 and the track may be extended or updated by track updater 2742. Track classification engine 2762 is coupled to track database 2736 to identify and update/modify tracks by, for example, adding, removing or modifying track-related data.

[0113] FIG. 28 is a diagram depicting a simulator configured to simulate an autonomous vehicle in a synthetic environment, according to various embodiments. Diagram 2800 includes a simulator 2840 that is configured to generate a simulated environment 2803. As shown, simulator 2840 is configured to use reference data 2822 (e.g., 3D map data and/or other map or route data including RNDF data or similar road network data) to generate simulated geometries, such as simulated surfaces 2892a and 2892b, within simulated environment 2803. Simulated surfaces 2892a and 2892b may simulate walls or front sides of buildings adjacent a roadway. Simulator 2840 may also pre-generated or procedurally generated use dynamic object data 2825 to simulate dynamic agents in a synthetic environment. An example of a dynamic agent is simulated dynamic object 2801, which is representative of a simulated cyclist having a velocity. The simulated dynamic agents may optionally respond to other static and dynamic agents in the simulated environment, including the simulated autonomous vehicle. For example, simulated object 2801 may slow down for other obstacles in simulated environment 2803 rather than follow a preset trajectory, thereby creating a more realistic simulation of actual dynamic environments that exist in the real world.

[0114] Simulator 2840 may be configured to generate a simulated autonomous vehicle controller 2847, which includes synthetic adaptations of a perception engine 2866, a localizer 2868, a motion controller 2862, and a planner 2864, each of which may have functionalities described herein within simulated environment 2803. Simulator 2840 may also generate simulated interfaces (“I/F”) 2849 to simulate the data exchanges with different sensors modalities and different sensor data formats. As such, simulated interface 2849 may simulate a software interface for packetized data from, for example, a simulated Lidar sensor 2872. Further, simulator 2840 may also be configured to generate a simulated autonomous vehicle 2830 that implements simulated AV controller 2847. Simulated autonomous vehicle 2830 includes simulated Lidar sensors 2872, simulated camera or image sensors 2874, and simulated radar sensors 2876. In the example shown, simulated Lidar sensor 2872 may be configured to generate a simulated laser consistent with ray trace 2892, which causes generation of simulated sensor return 2891. Note that simulator 2840 may simulate the addition of noise or other environmental effects on sensor data (e.g., added diffusion or reflections that affect simulated sensor return 2891, etc.). Further yet, simulator 2840 may be configured to simulate a variety of sensor defects, including sensor failure, sensor miscalibration, intermittent data outages, and the like.

[0115] Simulator 2840 includes a physics processor 2850 for simulating the mechanical, static, dynamic, and kinematic aspects of an autonomous vehicle for use in simulating behavior of simulated autonomous vehicle 2830. For example, physics processor 2850 includes a content mechanics module 2851 for simulating contact mechanics, a collision detection module 2852 for simulating the interaction between simulated bodies, and a multibody dynamics module 2854 to simulate the interaction between simulated mechanical interactions.

[0116] Simulator 2840 also includes a simulator controller 2856 configured to control the simulation to adapt the functionalities of any synthetically-generated element of simulated environment 2803 to determine cause-effect relationship, among other things. Simulator 2840 includes a simulator evaluator 2858 to evaluate the performance synthetically-generated element of simulated environment 2803. For example, simulator evaluator 2858 may analyze simulated vehicle commands 2880 (e.g., simulated steering angles and simulated velocities) to determine whether such commands are an appropriate response to the simulated activities within simulated environment 2803. Further, simulator evaluator 2858 may evaluate interactions of a teleoperator 2808 with the simulated autonomous vehicle 2830 via teleoperator computing device 2804. Simulator evaluator 2858 may evaluate the effects of updated reference data 2827, including updated map tiles and route data, which may be added to guide the responses of simulated autonomous vehicle 2830. Simulator evaluator 2858 may also evaluate the responses of simulator AV controller 2847 when policy data 2829 is updated, deleted, or added. The above-description of simulator 2840 is not intended to be limiting. As such, simulator 2840 is configured to perform a variety of different simulations of an autonomous vehicle relative to a simulated environment, which include both static and dynamic features. For example, simulator 2840 may be used to validate changes in software versions to ensure reliability. Simulator 2840 may also be used to determine vehicle dynamics properties and for calibration purposes. Further, simulator 2840 may be used to explore the space of applicable controls and resulting trajectories so as to effect learning by self-simulation.

[0117] FIG. 29 is an example of a flow chart to simulate various aspects of an autonomous vehicle, according to some embodiments. Flow chart 2900 begins at 2902, at which reference data including three dimensional map data is received into a simulator. Dynamic object data defining motion patterns for a classified object may be retrieved at 2904. At 2906, a simulated environment is formed based on at least three dimensional (“3D”) map data and the dynamic object data. The simulated environment may include one or more simulated surfaces. At 2908, an autonomous vehicle is simulated that includes a simulated autonomous vehicle controller that forms part of a simulated environment. The autonomous vehicle controller may include a simulated perception engine and a simulated localizer configured to receive sensor data. At 2910, simulated sensor data are generated based on data for at least one simulated sensor return, and simulated vehicle commands are generated at 2912 to cause motion (e.g., vectored propulsion) by a simulated autonomous vehicle in a synthetic environment. At 2914, simulated vehicle commands are evaluated to determine whether the simulated autonomous vehicle behaved consistent with expected behaviors (e.g., consistent with a policy).

[0118] FIG. 30 is an example of a flow chart to generate map data, according to some embodiments. Flow chart 3000 begins at 3002, at which trajectory data is retrieved. The trajectory data may include trajectories captured over a duration of time (e.g., as logged trajectories). At 3004, at least localization data may be received. The localization data may be captured over a duration of time (e.g., as logged localization data). At 3006, a camera or other image sensor may be implemented to generate a subset of the localization data. As such, the retrieved localization data may include image data. At 3008, subsets of localization data are aligned to identifying a global position (e.g., a global pose). At 3010, three dimensional (“3D”) map data is generated based on the global position, and at 3012, the 3 dimensional map data is available for implementation by, for example, a manual route data editor (e.g., including a manual road network data editor, such as an RNDF editor), an automated route data generator (e.g., including an automatic road network generator, including an automatic RNDF generator), a fleet of autonomous vehicles, a simulator, a teleoperator computing device, and any other component of an autonomous vehicle service.

[0119] FIG. 31 is a diagram depicting an architecture of a mapping engine, according to some embodiments. Diagram 3100 includes a 3D mapping engine that is configured to receive trajectory log data 3140, Lidar log data 3172, camera log data 3174, radar log data 3176, and other optional logged sensor data (not shown). Logic 3141 includes a loop-closure detector 3150 configured to detect whether sensor data indicates a nearby point in space has been previously visited, among other things. Logic 3141 also includes a registration controller 3152 for aligning map data, including 3D map data in some cases, relative to one or more registration points. Further, logic 3141 provides data 3142 representing states of loop closures for use by a global pose graph generator 3143, which is configured to generate pose graph data 3145. In some examples, pose graph data 3145 may also be generated based on data from registration refinement module 3146. Logic 3144 includes a 3D mapper 3154 and a Lidar self-calibration unit 3156. Further, logic 3144 receives sensor data and pose graph data 3145 to generate 3D map data 3120 (or other map data, such as 4D map data). In some examples, logic 3144 may implement a truncated sign distance function (“TSDF”) to fuse sensor data and/or map data to form optimal three-dimensional maps. Further, logic 3144 is configured to include texture and reflectance properties. 3D map data 3120 may be released for usage by a manual route data editor 3160 (e.g., an editor to manipulate Route data or other types of route or reference data), an automated route data generator 3162 (e.g., logic to configured to generate route data or other types of road network or reference data), a fleet of autonomous vehicles 3164, a simulator 3166, a teleoperator computing device 3168, and any other component of an autonomous vehicle service. Mapping engine 3110 may capture semantic information from manual annotation or automatically-generated annotation as well as other sensors, such as sonar or instrumented environment (e.g., smart stop-lights).

[0120] FIG. 32 is a diagram depicting an autonomous vehicle application, according to some examples. Diagram 3200 depicts a mobile computing device 3203 including an autonomous service application 3240 that is configured to contact an autonomous vehicle service platform 3201 to arrange transportation of user 3202 via an autonomous vehicle 3230. As shown, autonomous service application 3240 may include a transportation controller 3242, which may be a software application residing on a computing device (e.g., a mobile phone 3203, etc.). Transportation controller 3242 is configured to receive, schedule, select, or perform operations related to autonomous vehicles and/or autonomous vehicle fleets for which a user 3202 may arrange transportation from the user's location to a destination. For example, user 3202 may open up an application to request vehicle 3230. The application may display a map and user 3202 may drop a pin to indicate their destination within, for example, a geo-fenced region. Alternatively, the application may display a list of nearby pre-specified pick-up locations, or provide the user with a text entry field in which to type a destination either by address or by name.

[0121] Further to the example shown, autonomous vehicle application 3240 may also include a user identification controller 3246 that may be configured to detect that user 3202 is in a geographic region, or vicinity, near autonomous vehicle 3230, as the vehicle approaches. In some situations, user 3202 may not readily perceive or identify autonomous vehicle 3230 as it approaches for use by user 3203 (e.g., due to various other vehicles, including trucks, cars, taxis, and other obstructions that are typical in city environments). In one example, autonomous vehicle 3230 may establish a wireless communication link 3262 (e.g., via a radio frequency (“RF”) signal, such as WiFi or Bluetooth®, including BLE, or the like) for communicating and/or determining a spatial location of user 3202 relative to autonomous vehicle 3230 (e.g., using relative direction of RF signal and signal strength). In some cases, autonomous vehicle 3230 may detect an approximate geographic location of user 3202 using, for example, GPS data or the like. A GPS receiver (not shown) of mobile computing device 3203 may be configured to provide GPS data to autonomous vehicle service application 3240. Thus, user identification controller 3246 may provide GPS data via link 3260 to autonomous vehicle service platform 3201, which, in turn, may provide that location to autonomous vehicle 3230 via link 3261. Subsequently, autonomous vehicle 3230 may determine a relative distance and/or direction of user 3202 by comparing the user's GPS data to the vehicle's GPS-derived location.

[0122] Autonomous vehicle 3230 may also include additional logic to identify the presence of user 3202, such that logic configured to perform face detection algorithms to detect either user 3202 generally, or to specifically identify the identity (e.g., name, phone number, etc.) of user 3202 based on the user's unique facial characteristics. Further, autonomous vehicle 3230 may include logic to detect codes for identifying user 3202. Examples of such codes include specialized visual codes, such as QR codes, color codes, etc., specialized audio codes, such as voice activated or recognized codes, etc., and the like. In some cases, a code may be an encoded security key that may be transmitted digitally via link 3262 to autonomous vehicle 3230 to ensure secure ingress and/or egress. Further, one or more of the above-identified techniques for identifying user 3202 may be used as a secured means to grant ingress and egress privileges to user 3202 so as to prevent others from entering autonomous vehicle 3230 (e.g., to ensure third party persons do not enter an unoccupied autonomous vehicle prior to arriving at user 3202). According to various examples, any other means for identifying user 3202 and providing secured ingress and egress may also be implemented in one or more of autonomous vehicle service application 3240, autonomous vehicle service platform 3201, and autonomous vehicle 3230.

[0123] To assist user 3302 in identifying the arrival of its requested transportation, autonomous vehicle 3230 may be configured to notify or otherwise alert user 3202 to the presence of autonomous vehicle 3230 as it approaches user 3202. For example, autonomous vehicle 3230 may activate one or more light-emitting devices 3280 (e.g., LEDs) in accordance with specific light patterns. In particular, certain light patterns are created so that user 3202 may readily perceive that autonomous vehicle 3230 is reserved to service the transportation needs of user 3202. As an example, autonomous vehicle 3230 may generate light patterns 3290 that may be perceived by user 3202 as a “wink,” or other animation of its exterior and interior lights in such a visual and temporal way. The patterns of light 3290 may be generated with or without patterns of sound to identify to user 3202 that this vehicle is the one that they booked.

[0124] According to some embodiments, autonomous vehicle user controller 3244 may implement a software application that is configured to control various functions of an autonomous vehicle. Further, an application may be configured to redirect or reroute the autonomous vehicle during transit to its initial destination. Further, autonomous vehicle user controller 3244 may be configured to cause on-board logic to modify interior lighting of autonomous vehicle 3230 to effect, for example, mood lighting. Controller 3244 may also control a source of audio (e.g., an external source such as Spotify, or audio stored locally on the mobile computing device 3203), select a type of ride (e.g., modify desired acceleration and braking aggressiveness, modify active suspension parameters to select a set of “road-handling” characteristics to implement aggressive driving characteristics, including vibrations, or to select “soft-ride” qualities with vibrations dampened for comfort), and the like. For example, mobile computing device 3203 may be configured to control HVAC functions as well, like ventilation and temperature.

[0125] FIGS. 33 to 35 illustrate examples of various computing platforms configured to provide various functionalities to components of an autonomous vehicle service, according to various embodiments. In some examples, computing platform 3300 may be used to implement computer programs, applications, methods, processes, algorithms, or other software to perform the above-described techniques.

[0126] Note that various structures and/or functionalities of FIG. 33 are applicable to FIGS. 34 and 35, and, as such, some elements in those figures may be discussed in the context of FIG. 33.

[0127] In some cases, computing platform 3300 can be disposed in any device, such as a computing device 3390a, which may be disposed in an autonomous vehicle 3391, and/or mobile computing device 3390b.

[0128] Computing platform 3300 includes a bus 3302 or other communication mechanism for communicating information, which interconnects subsystems and devices, such as processor 3304, system memory 3306 (e.g., RAM, etc.), storage device 3308 (e.g., ROM, etc.), an in-memory cache (which may be implemented in RAM 3306 or other portions of computing platform 3300), a communication interface 3313 (e.g., an Ethernet or wireless controller, a Bluetooth controller, NFC logic, etc.) to facilitate communications via a port on communication link 3321 to communicate, for example, with a computing device, including mobile computing and/or communication devices with processors. Processor 3304 can be implemented with one or more graphics processing units (“GPUs”), with one or more central processing units (“CPUs”), such as those manufactured by Intel® Corporation, or one or more virtual processors, as well as any combination of CPUs and virtual processors. Computing platform 3300 exchanges data representing inputs and outputs via input-and-output devices 3301, including, but not limited to, keyboards, mice, audio inputs (e.g., speech-to-text devices), user interfaces, displays, monitors, cursors, touch-sensitive displays, LCD or LED displays, and other I/O-related devices.

[0129] According to some examples, computing platform 3300 performs specific operations by processor 3304 executing one or more sequences of one or more instructions stored in system memory 3306, and computing platform 3300 can be implemented in a client-server arrangement, peer-to-peer arrangement, or as any mobile computing device, including smart phones and the like. Such instructions or data may be read into system memory 3306 from another computer readable medium, such as storage device 3308. In some examples, hard-wired circuitry may be used in place of or in combination with software instructions for implementation. Instructions may be embedded in software or firmware. The term “computer readable medium” refers to any tangible medium that participates in providing instructions to processor 3304 for execution. Such a medium may take many forms, including but not limited to, non-volatile media and volatile media. Non-volatile media includes, for example, optical or magnetic disks and the like. Volatile media includes dynamic memory, such as system memory 3306.

[0130] Common forms of computer readable media includes, for example, floppy disk, flexible disk, hard disk, magnetic tape, any other magnetic medium, CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, RAM, PROM, EPROM, FLASH-EPROM, any other memory chip or cartridge, or any other medium from which a computer can read. Instructions may further be transmitted or received using a transmission medium. The term “transmission medium” may include any tangible or intangible medium that is capable of storing, encoding or carrying instructions for execution by the machine, and includes digital or analog communications signals or other intangible medium to facilitate communication of such instructions. Transmission media includes coaxial cables, copper wire, and fiber optics, including wires that comprise bus 3302 for transmitting a computer data signal.

[0131] In some examples, execution of the sequences of instructions may be performed by computing platform 3300. According to some examples, computing platform 3300 can be coupled by communication link 3321 (e.g., a wired network, such as LAN, PSTN, or any wireless network, including WiFi of various standards and protocols, Bluetooth®, NFC, Zig-Bee, etc.) to any other processor to perform the sequence of instructions in coordination with (or asynchronous to) one another. Computing platform 3300 may transmit and receive messages, data, and instructions, including program code (e.g., application code) through communication link 3321 and communication interface 3313. Received program code may be executed by processor 3304 as it is received, and/or stored in memory 3306 or other non-volatile storage for later execution.

[0132] In the example shown, system memory 3306 can include various modules that include executable instructions to implement functionalities described herein. System memory 3306 may include an operating system (“O/S”) 3332, as well as an application 3336 and/or logic module(s) 3359. In the example shown in FIG. 33, system memory 3306 includes an autonomous vehicle (“AV”) controller module 3350 and/or its components (e.g., a perception engine module, a localization module, a planner module, and/or a motion controller module), any of which, or one or more portions of which, can be configured to facilitate an autonomous vehicle service by implementing one or more functions described herein.

[0133] Referring to the example shown in FIG. 34, system memory 3306 includes an autonomous vehicle service platform module 3450 and/or its components (e.g., a teleoperator manager, a simulator, etc.), any of which, or one or more portions of which, can be configured to facilitate managing an autonomous vehicle service by implementing one or more functions described herein.

[0134] Referring to the example shown in FIG. 35, system memory 3306 includes an autonomous vehicle (“AV”) module and/or its components for use, for example, in a mobile computing device. One or more portions of module 3550 can be configured to facilitate delivery of an autonomous vehicle service by implementing one or more functions described herein.

[0135] Referring back to FIG. 33, the structures and/or functions of any of the above-described features can be implemented in software, hardware, firmware, circuitry, or a combination thereof. Note that the structures and constituent elements above, as well as their functionality, may be aggregated with one or more other structures or elements. Alternatively, the elements and their functionality may be subdivided into constituent sub-elements, if any. As software, the above-described techniques may be implemented using various types of programming or formatting languages, frameworks, syntax, applications, protocols, objects, or techniques. As hardware and/or firmware, the above-described techniques may be implemented using various types of programming or integrated circuit design languages, including hardware description languages, such as any register transfer language (“RTL”) configured to design field-programmable gate arrays (“FPGAs”), application-specific integrated circuits (“ASICs”), or any other type of integrated circuit. According to some embodiments, the term “module” can refer, for example, to an algorithm or a portion thereof, and/or logic implemented in either hardware circuitry or software, or a combination thereof. These can be varied and are not limited to the examples or descriptions provided.

[0136] In some embodiments, module 3350 of FIG. 33, module 3450 of FIG. 34, and module 3550 of FIG. 35, or one or more of their components, or any process or device described herein, can be in communication (e.g., wired or wirelessly) with a mobile device, such as a mobile phone or computing device, or can be disposed therein.

[0137] In some cases, a mobile device, or any networked computing device (not shown) in communication with one or more modules 3359 (module 3350 of FIG. 33, module 3450 of FIG. 34, and module 3550 of FIG. 35) or one or more of its components (or any process or device described herein), can provide at least some of the structures and/or functions of any of the features described herein. As depicted in the above-described figures, the structures and/or functions of any of the above-described features can be implemented in software, hardware, firmware, circuitry, or any combination thereof. Note that the structures and constituent elements above, as well as their functionality, may be aggregated or combined with one or more other structures or elements. Alternatively, the elements and their functionality may be subdivided into constituent sub-elements, if any. As software, at least some of the above-described techniques may be implemented using various types of programming or formatting languages, frameworks, syntax, applications, protocols, objects, or techniques. For example, at least one of the elements depicted in any of the figures can represent one or more algorithms. Or, at least one of the elements can represent a portion of logic including a portion of hardware configured to provide constituent structures and/or functionalities.

[0138] For example, module 3350 of FIG. 33, module 3450 of FIG. 34, and module 3550 of FIG. 35, or one or more of its components, or any process or device described herein, can be implemented in one or more computing devices (i.e., any mobile computing device, such as a wearable device, an audio device (such as headphones or a headset) or mobile phone, whether worn or carried) that include one or more processors configured to execute one or more algorithms in memory. Thus, at least some of the elements in the above-described figures can represent one or more algorithms. Or, at least one of the elements can represent a portion of logic including a portion of hardware configured to provide constituent structures and/or functionalities. These can be varied and are not limited to the examples or descriptions provided.

[0139] As hardware and/or firmware, the above-described structures and techniques can be implemented using various types of programming or integrated circuit design languages, including hardware description languages, such as any register transfer language (“RTL”) configured to design field-programmable gate arrays (“FPGAs”), application-specific integrated circuits (“ASICs”), multi-chip modules, or any other type of integrated circuit.

[0140] For example, module 3350 of FIG. 33, module 3450 of FIG. 34, and module 3550 of FIG. 35, or one or more of its components, or any process or device described herein, can be implemented in one or more computing devices that include one or more circuits. Thus, at least one of the elements in the above-described figures can represent one or more components of hardware. Or, at least one of the elements can represent a portion of logic including a portion of a circuit configured to provide constituent structures and/or functionalities.

[0141] According to some embodiments, the term “circuit” can refer, for example, to any system including a number of components through which current flows to perform one or more functions, the components including discrete and complex components. Examples of discrete components include transistors, resistors, capacitors, inductors, diodes, and the like, and examples of complex components include memory, processors, analog circuits, digital circuits, and the like, including field-programmable gate arrays (“FPGAs”), application-specific integrated circuits (“ASICs”). Therefore, a circuit can include a system of electronic components and logic components (e.g., logic configured to execute instructions, such that a group of executable instructions of an algorithm, for example, and, thus, is a component of a circuit). According to some embodiments, the term “module” can refer, for example, to an algorithm or a portion thereof, and/or logic implemented in either hardware circuitry or software, or a combination thereof (i.e., a module can be implemented as a circuit). In some embodiments, algorithms and/or the memory in which the algorithms are stored are “components” of a circuit. Thus, the term “circuit” can also refer, for example, to a system of components, including algorithms. These can be varied and are not limited to the examples or descriptions provided.

[0142] FIGS. 36A to 36C illustrate high-level block diagrams depicting an autonomous vehicle system experiencing a sensor calibration issue while in operation, according to various embodiments. An autonomous vehicle system 3602 may be operating and driving past various structures in a typical driving scenario. As mentioned above, with respect to FIG. 3A, the autonomous vehicle system 3602 may include many types of sensors or any quantity of sensors to facilitate perception, including image capture sensors, audio capture sensors, LIDAR, RADAR, SONAR, GPS, and IMU. As illustrated in FIG. 36A, LIDAR sensors 3604 may capture distance, image intensity, and 3D point cloud data based on laser returns 3608. As illustrated in FIG. 36A, LIDAR sensors 3604 may be arranged in an array. In other examples, LIDAR sensors 3604 may be arranged in pairs at different heights along corners of the autonomous vehicle system 3602. In another embodiment, other arrangements of sensors may be used, such as a single multi-beam LIDAR sensor, a single single-beam LIDAR sensor, a group, an array, as well as a distributed array of LIDAR sensors.

[0143] An example LIDAR sensor 3604 is a VELODYNE VLP-16 Real-Time 3D LiDAR Sensor, manufactured by Velodyne Acoustics, Inc. in Morgan Hill, Calif. A LIDAR Sensor 3604 creates 360 degree 3D images by using 16 laser/detector pairs mounted in a compact housing. The housing rapidly spins to scan the surrounding environment. The lasers fire thousands of times per second, providing a rich 3D point cloud in real time. A laser return 3608 provides data about the reflectivity of an object with 256-bit resolution independent of laser power and distance over a range from 1 meter to 100 meters. For example, laser returns 3608 from LIDAR sensors 3604 may be used to identify a labeled point 3626 on a landmark 3632 independently. Data representing laser returns 3608 may be stored and processed in the autonomous vehicle system 3602, as described above. A LIDAR sensor 3604 may be synchronized with GPS data based on GPS-supplied time pulses, enabling the ability to determine the exact GPS location at the time of each firing time of each laser. This enables data representing laser returns 3608 to be geo-referenced in real-time. In another embodiment, a LIDAR sensor may be a single-beam LIDAR sensor having various beam angles, beam intensities, and beam strengths. The rotational speed of LIDAR sensors may be adjusted, but may range between 5-20 rotations per second, in one embodiment.

[0144] A LIDAR sensor 3604 measures the reflectivity of an object with 256-bit resolution, independent of laser power and distance over a range of 1 to 100 meters. Different types of objects provide different reflectivity values, such as diffuse reflectors which report values from 0-100 for reflectivity of 0-100%. Retro-reflectors, devices or surfaces that reflect light back to its source with a minimum of scattering, report values from 101 to 255 with 255 being the reported reflectivity for an ideal retro-reflector, and with 101 to 254 being the reported reflectivity for partially obstructed or imperfect retro-reflectors.

[0145] A LIDAR sensor 3604 outputs data packets via Ethernet that includes azimuth (horizontal angle), elevation angle (vertical angle), distance to the object, and time stamp. Coordinates are reported in spherical coordinates such that X, Y, Z coordinates may be calculated for each measured point. The calculated X, Y, Z data may be stored for additional processing and/or displayed on a computer as a series of point clouds. A point cloud is a set of data points in a coordinate system intended to represent the external surface of an object.

[0146] Traditionally, a LIDAR sensor may be calibrated manually by using various objects, such as a checkerboard object, to ensure that the reflectivity values are correctly calibrated. A manual calibration process for an autonomous vehicle is not feasible due to the number of beams that may need to be calibrated. Additionally, other types of sensors, including GPS, IMU, RADAR, SONAR, and cameras, may be relied upon while an autonomous vehicle is in operation to correct a miscalibrated sensor, such as a LIDAR sensor. In this way, the sensors 3610 included on an autonomous vehicle system 3602 are jointly calibrated against each other. No one sensor is “sacrosanct,” meaning that each sensor may experience drift, or miscalibration over time. In one embodiment, the autonomous vehicle system 3602 may track the drift of each sensor over time.

[0147] As described above, an autonomous vehicle system 3602 uses various sensors, including LIDAR sensors 3604, in a perception system to aid in localizing the autonomous vehicle system 3602. For safety reasons and operational efficiency, the autonomous vehicle system 3602 determines its location and the surrounding environment, such as static objects like lane markings and curbs as well as dynamic objects like moving vehicles in real-time and continuously. The autonomous vehicle system 3602 relies on sensor data to update map tile information, such as new lane markings, as well as current localized information, such as other moving vehicles also traveling in the same direction of travel 3630 as the autonomous vehicle 3602.

[0148] While in operation, a LIDAR sensor 3604 may become out of alignment due to various causes, such as a bird striking the sensor, a person bumping into the sensor, or a myriad of other reasons. Various sensors may be relied upon to determine an extrinsic calibration of the misaligned LIDAR sensor, such as an X, Y, Z offset in relation to the physical frame of the autonomous vehicle system 3602. For example, a RADAR sensor or group of RADAR sensors may be used to determine the position of the misaligned LIDAR sensor. The initial physical position of the LIDAR sensor may be stored in the AV system 3602 such that the current position of the LIDAR sensor may be compared after being determined by the RADAR sensor(s).

[0149] Similarly, SONAR sensors may be used to track dynamic and static objects that are within close proximity to the AV system 3602 as well as objects that are low to the ground. SONAR sensors may be used to aid the AV system 3602 in parking. The SONAR sensors may also, in one embodiment, be used to help track the position of one or more sensors that may be arranged in the lower half of the AV system 3602, such as RADAR sensors, audio capture sensors, and odometry sensors located on the tires of the AV system 3602. In this way, SONAR sensors may also be used to identify any translational offset of sensors that may have occurred that could affect the sensor measurements of those sensors.

[0150] Cameras may be used to create grid maps of the current environment of the AV system 3602 to aid in localizing the AV system 3602. Camera data may be fused with LIDAR sensor data such that the edges detected by the LIDAR sensors may be aligned with edges of objects in the images captured by one or more cameras positioned on the AV system 3602 in various locations. In this way, a new fused data stream that includes camera data and LIDAR sensor data may be maintained and continuously calibrated. Further, this fused data stream may also be aligned with external map data files, such as RNDF files and mission data files, by various systems of the AV system 3602, including the localizer and planner. Discrepancies may be detected based on the fusing of the data, such as changes in lane markings from a downloaded RNDF file, as well as new buildings and structures being erected. By calibrating the sensors jointly and against each other, these discrepancies may be resolved.

[0151] Cameras may also, over time, experience drift that results in fuzzier images. A miscalibrated camera may be calibrated by using LIDAR data to help identify edges of objects, thus enabling the camera to focus and adjust lens properties to sharpen images. By aligning detected edges from laser returns of LIDAR sensors with the same edges of objects within captured images, a miscalibrated camera may be calibrated, on the fly in real time by onboard processors on the AV system 3602.

[0152] Inertial measurement units (IMUs) may also be used to help calibrate LIDAR sensors and cameras. Images captured by cameras do not include a relative coordinate system, but IMU data may be used to establish the direction of gravity within these images. As a result, a coordinate system may be assigned to images. Further, IMU data may provide the relative velocity of the AV system 3602 at the time stamp of a data point, such as a point cloud captured by a LIDAR sensor. Thus, in a situation where a sensor measurement may not be expected, the current velocity may be captured to identify other sensor data previously captured by the same sensor and other sensors to aid in detecting whether a sensor needs to be calibrated.

[0153] Illustrated in FIG. 36A, two LIDAR sensor arrays 3604 are mounted on an AV system 3602, where each array includes five LIDAR sensors operating in redundancy, and each array position at the front and back ends of the autonomous vehicle system 3602, in relation to the direction of travel 3630. In other embodiments, additional or fewer LIDAR sensors 3604 may be used, and in various other arrangements, such as a curved array of LIDAR sensors 3604, and/or two LIDAR sensors 3604 at each corner of the autonomous vehicle 3602. However, as illustrated in FIG. 36A, LIDAR sensors 3604 are arranged in an array on each end of the roof of the AV system 3602. The LIDAR sensors 3604 were calibrated at the factory such that the laser returns 3608 are captured independently, in real-time with synchronized GPS data.

[0154] In this example, one of the LIDAR sensors returns sensor data that appears anomalous and/or incorrect. The other LIDAR sensors generated laser returns 3608 that indicated a labeled point 3626 having a distance, reflectivity value, and position in relation to the AV system 3602. However, the LIDAR sensor with the abnormal sensor data returned an unidentified point 3628 that had a completely different range, reflectivity value, and/or position in relation to the AV system 3602. As illustrated in FIG. 36A, the landmark 3632 is a speed limit sign having a very reflective surface. Thus, the labeled point 3626 may have an expected reflectivity value around 100 because of the reflectivity of the surface of the speed limit sign. However, the unidentified point 3628 of the miscalibrated LIDAR sensor may return a reflectivity value of a building, which may not be as reflective as the speed limit sign. Additionally, the distance of the building may be much further than the speed limit sign. Further, the location of the unidentified point 3628 may be reported by the miscalibrated LIDAR sensor as the same location as the labeled point 3626, but in reality, the location may include a large offset in roll, pitch, and/or yaw. As a result, the AV system 3602 may become aware that the LIDAR sensor reporting the abnormal sensor measurement may need calibration.

[0155] Mounted cameras 3606 on the AV system 3602 may be used to aid in perception of the environment by the AV system 3602. Mounted cameras 3606 may be arranged such that a camera is located at each corner of every side of the AV system 3602, including the front, rear, and sides, relative to the direction of travel 3630. As mentioned above, images may be fused with LIDAR sensor data by aligning detected edges from laser returns with edges identified in the images. In the example above, mounted cameras 3606 may be used to capture images that may be fused with the LIDAR sensor data, including the miscalibrated sensor. Part of the fusion process may include generating a probabilistic model to label and identify objects within the captured images based on the LIDAR sensor data. Here, the probability that the labeled point 3626 would not be 100% because the miscalibrated LIDAR sensor returned differing reflectivity and distance values. However, a separate process may identify the speed limit sign as a landmark 3632 that has been captured and identified before based on historical log files, download map data, and the like, such that the labeled point 3626 is identified with 100% probability. This separate process may then confirm that the LIDAR sensor returning the differing reflectivity and distance values requires calibration, in one embodiment.

[0156] FIG. 36B illustrates a captured camera frame 3640 of the scenario presented in FIG. 36A, except the captured camera frame 3640 is captured a few seconds after the miscalibrated laser beam 3624 is captured by the LIDAR sensor. The AV system 3602 may rely on various data points captured by various sensors to aid in calibrating the sensor requiring calibration. Here, the camera frame 3640 depicts, from the vantage point of the mounted camera(s) 3606 located on the side of the AV system 3602 closest to the unidentified point 3628 and the labeled point 3626. This camera frame 3640 illustrates the landmark 3632 that includes the labeled point 3626 as well as the unidentified point 3628. The direction of travel 3630 may be computed based on odometry sensor data 3636 as well as IMU data. This enables the AV system 3602 to determine and identify data points at the timestamp when the miscalibrated laser beam 3624 was captured, as illustrated in FIG. 36A.

[0157] Further, the current pose 3638 may be determined based on GPS data and fused data including currently captured LIDAR data, IMU and camera data. A previous pose 3634 may be calculated to retrieve data points that may be useful in calibrating the LIDAR sensor reporting the miscalibrated laser beam 3624. The AV system 3602 may only store the last thirty (30) seconds of sensor data due to the massive amounts of sensor data being captured by the various sensors 3610 onboard the AV system 3602. Thus, determining the previous pose 3634 helps identify relevant data points to be used in calibrating the detected miscalibrated sensor.

[0158] FIG. 36C illustrates a high-level conceptual block diagram of the computational problem presented by the scenario presented in FIGS. 36A and 36B. An expected laser return 3642 may be computed based on a generative probabilistic model using the fused data stream of the camera and LIDAR sensor data, in one embodiment. In other embodiments, the expected laser return 3642 may be computed based on the other LIDAR sensors included in the array 3604. In a further embodiment, the expected laser return 3642 may be computed based on a heuristic identifying the local maximum probability of what the miscalibrated LIDAR sensor 3646 should have returned, using the sensor data from the various sensors captured at the time of the miscalibrated laser beam 3624. The expected laser return 3642 may also be computed based on the position, distance, and reflectivity value of the labeled point 3626, in an embodiment.

[0159] Using trigonometric methods, an error offset 3650 may be computed for the miscalibrated LIDAR sensor 3646. The error offset 3650 may have angular dimensions in roll, pitch, and/or yaw, as well as a range offset dimension. The recorded velocity and distance traveled, captured by the IMU and odometry sensors, may be used by the trigonometric methods to calculate the error offset 3650, in an embodiment. For example, given the position of the miscalibrated LIDAR sensor 3646 and the velocity of the AV system 3602 at the time of the miscalibrated laser beam 3624, the error offset 3650 may be computed based on the expected laser return 3642.

[0160] FIG. 37 illustrates a high-level block diagram of calibration for autonomous vehicle operation, according to various embodiments. In one embodiment, the autonomous vehicle system 3602 (“AV system”) may include a perception system 3702 that includes an intrinsic sensor calibration module 3704, an extrinsic sensor calibration module 3706, a data transform module 3708, a generative model module 3710, a heuristics engine module 3712, and a sensor drift tracking module 3714.

[0161] Upon receiving an indication of an anomaly in a sensor measurement from a sensor of the sensors 3610 included in the AV system 3602, which may include LIDAR sensors 3604, RADAR sensors 3620, other sensors 3660, IMUs 3612, cameras 3614, odometry sensors 3616, GPS 3618, and SONAR sensors 3622, a perception system 3702 may utilize a calibration detection module 3730 to process the anomalous sensor measurements. The AV system 3602 may rely on the sensors 3610 and “fuse” the data generated by the heterogeneous types of sensors, such as data from LIDAR sensors 3604 and motion data from IMUs 3612, in one embodiment. A localizer, not pictured, may generate a probabilistic map of the current environment, assigning probability scores to labeled objects in the field of perception. Probability scores may be lower based on the uncertainty of the data received from the miscalibrated sensor.

[0162] A calibration detection module 3730 may identify a miscalibrated sensor of the sensors 3610 based on an expected data stream not matching the measured data received from the miscalibrated sensor. The calibration detection module 3730 may identify a miscalibrated sensor based on anomalous data being received from the sensor, in one embodiment. Once a sensor is identified as potentially miscalibrated, log file data may be retrieved from a log file store 3716 to assist in calibrating the identified sensor.

[0163] An intrinsic sensor calibration module 3704 may be used to determine intrinsic calibration parameters for a sensor. For example, a LIDAR sensor may require an intrinsic calibration of reflectivity values captured by laser returns. As described above, other types of sensors may provide data to assist with the intrinsic calibration of LIDAR sensors. Cameras, on the other hand, may have other intrinsic calibration parameters, such as color mapping, focal length, image positioning, scaling/skew factors, and lens distortion that may affect the imaging process. LIDAR sensor data, such as 3D point clouds, may be used to calibrate cameras, in an embodiment. In other embodiments, other types of sensors, such as an IMU, may help to calibrate a camera, in conjunction with LIDAR sensor data, to correct image positioning such that the image is correctly centered on the AV system 3602, for example. The intrinsic sensor calibration module 3704 may operate in conjunction with a data transform module 3708 and a generative model module 3710 to perform the computations necessary to converge on the optimal intrinsic calibration parameters for the miscalibrated sensor, in an embodiment.

[0164] An extrinsic sensor calibration module 3706 may be used to identify a physical translation of a sensor in X, Y, Z space and/or roll, pitch, and yaw coordinate system in relation to the frame of the AV system 3602. For example, a bird strike or other impact may have caused a sensor 3610 to have a translational offset. RADAR sensors 3620 may be used to detect the translational offset through radio waves detecting the physical location of the miscalibrated sensor. As a result, the extrinsic sensor calibration module 3706 may determine that the miscalibrated sensor was horizontally rotated by 9 degrees and vertically rotated by 1 degree, for example. The translational offset may then be stored in association with the identified sensor such that future data retrieved from the sensor is correctly calibrated based on the translational offset. In this way, the AV system 3602 has been self-calibrated while in operation, without having to stop and interrupt the user experience. Similar to the intrinsic sensor calibration module 3704, the extrinsic sensor calibration module 3706 may rely on a data transform module 3708 and/or a generative model module 3710 to perform the computations necessary to converge on optimal extrinsic calibration parameters using other sensor data, log file data, and/or map tile data.

[0165] A data transform module 3708 may perform trigonometric computations to transform spherical coordinates to X, Y, Z coordinates, for example, as well as other computations, such as transforming an image captured by a camera that has no coordinate system to the coordinate system of the AV system 3602 such that the image is aligned with the X, Y, Z coordinates of the AV system 3602 and its current pose. In one embodiment, the camera image data may also be transformed by the data transform module 3708 to align the image data with received RNDF tiles and/or other map tile data retrieved from a map tile store 3718. Other computations may be performed by a data transform module 3708, such as performing trigonometric computations to determine an error offset for calibrating LIDAR sensors, calculating angular offsets of laser beams, and so forth. In one embodiment, a data transform module 3708 may be used to align the AV system 3602 with received GPS coordinates for localization with RNDF and/or other map data retrieved from a map tile store 3718.

[0166] A generative model module 3710 may generate various probabilistic models for different calibration scenarios. An expectation model may be generated by the generative model module 3710 to calibrate an errant LIDAR sensor based on other LIDAR sensor data, IMU data, camera data, RADAR data, SONAR data, and/or odometry data. In one embodiment, an expectation model may be generated for fused data that includes camera data and LIDAR sensor data, where the 3D point cloud data has been fused with image data. In other embodiments, two models may be generated for determining calibration parameters for a miscalibrated sensor, where a first model is generated for an expected sensor measurement based on other similar sensors and a second model is generated for an expected sensor measurement based on other types of data, such as other types of sensors, log file data reporting previously measured data points, and/or map data.

[0167] For example, the AV system 3602 may use log file data retrieved from the log file store 3716 and/or sensor data being generated from sensors 3610 to assign a probability score of the likelihood of an object being correctly labeled within the field of perception, such as labeled point 3626 being correctly correlated with the landmark 3632. Probability scores such as this may be generated from past sensor measurements and current sensor measurements, including the potentially miscalibrated sensor, in one embodiment. In another embodiment, a localizer, as described above with respect to FIG. 4, may generate probabilistic map tiles that assign probabilities to each map tile, which may be 16 cm×16 cm, in one embodiment. Map tiles may then be compared with map tile data retrieved from a map tile store 3718, in an embodiment. A generative model may be generated by the generative model module 3710 to determine the probabilities of the correctness of each stored map tile based on the currently measured sensor data, in an embodiment. A new map tile may be generated and stored in the map tile store 3718 based on the measured sensor data, in another embodiment. The AV system 3602 is able to compute these probabilities in real-time, as the vehicle is in operation, because of processing power of on-board computers, super-fast processors, and low-latency log file retrieval and storage.

[0168] A heuristics engine module 3712 may provide one or more heuristic rules for optimizing various searches for optimal calibration parameters. A typical method for calibration involves an initial guess that converges towards an optimal calibration parameter through a series of step-wise increments, using generative probabilistic models provided by the generative model module 3710. Because a search is computationally expensive, a heuristic rule may be used to arrive at a calibration parameter that is acceptable, but not the most optimal calibration parameter. Various predetermined thresholds may be used by the heuristics engine module 3712 in the heuristic rules to ensure that the AV system 3602 is operating within acceptable safe operating parameters.

[0169] A sensor drift tracking module 3714 may track sensor drift for one or more sensors 3610 of the AV system 3602. Sensor drift may be a known offset or error of the sensor. Rather than calibrate a sensor that has sensor drift, which may cost processing time and/or computational effort by various modules of the AV system 3602, the drift may be tracked and stored in association with the identified sensor 3610. The identified sensor 3610 may then be calibrated offline, in one embodiment, by an offline calibration system 3720 to coincide when the AV system 3602 is charging its batteries and/or is parked for service. Sensor calibrations may be simulated offline using a simulator that replicates real driving conditions in a simulated world, in an embodiment.

[0170] In one embodiment, a quantifiable measure of how the perception system 3702 is affected by a miscalibrated sensor may be determined by the sensor drift tracking module 3714. The AV system 3602 may continue operating in a sub-optimal mode of operation that remains within safe levels of operation, in one embodiment. In this sub-optimal mode of operation, other sensors may be used and relied upon to ensure a safe level of operation, such as using camera data in conjunction with IMU motion data to compensate for blind spots caused by a miscalibrated sensor.

[0171] A teleoperator system 3750 may be contacted by an AV system 3602, optionally, in relation to a calibration issue. For example, a calibration detection module 3730 may determine that a sensor is reporting anomalous data and that calibration may be required. However, the sensor may also be experiencing sensor drift that can be tracked and covered by other sensors. The AV system 3602 may present the teleoperator system 3750 with the data received by the identified sensor 3610 and request guidance as to whether the sensor should be calibrated online or offline, for example. The teleoperator system 3750 may be presented with the current trajectory and/or route of the AV system 3602 as well as other operational parameters, such as current battery level, computational processing efficiency, and/or storage capacity, for example. Through an interface provided on the teleoperator system 3750, a teleoperator may select a course of action and provide guidance to the AV system 3602 as to when to calibrate the sensor, in one embodiment.

[0172] FIG. 38 is a network diagram of a system for calibration for autonomous vehicle operation, showing a block diagram of an autonomous vehicle management system, according to an embodiment. The system environment includes one or more AV systems 3602, teleoperator systems 3750, user devices 3806, an autonomous vehicle (“AV”) management system 3800, and a network 3804. In alternative configurations, different and/or additional modules can be included in the system.

[0173] The user devices 3806 may include one or more computing devices that can receive user input and can transmit and receive data via the network 3804. In one embodiment, the user device 3806 is a conventional computer system executing, for example, a Microsoft Windows-compatible operating system (OS), Apple OS X, and/or a Linux distribution. In another embodiment, the user device 3806 can be a device having computer functionality, such as a personal digital assistant (PDA), mobile telephone, smart-phone, wearable device, etc. The user device 3806 is configured to communicate via network 3804. The user device 3806 can execute an application, for example, a browser application that allows a user of the user device 3806 to interact with the AV management system 3800. In another embodiment, the user device 3806 interacts with the AV management system 3800 through an application programming interface (API) that runs on the native operating system of the user device 3806, such as iOS and ANDROID.

[0174] In one embodiment, the network 3804 uses standard communications technologies and/or protocols. Thus, the network 3804 can include links using technologies such as Ethernet, 802.11, worldwide interoperability for microwave access (WiMAX), 3G, 4G, CDMA, digital subscriber line (DSL), etc. Similarly, the networking protocols used on the network 3804 can include multiprotocol label switching (MPLS), the transmission control protocol/Internet protocol (TCP/IP), the User Datagram Protocol (UDP), the hypertext transport protocol (HTTP), the simple mail transfer protocol (SMTP), and the file transfer protocol (FTP). The data exchanged over the network 3804 can be represented using technologies and/or formats including the hypertext markup language (HTML) and the extensible markup language (XML). In addition, all or some of links can be encrypted using conventional encryption technologies such as secure sockets layer (SSL), transport layer security (TLS), and Internet Protocol security (IPsec).

[0175] FIG. 38 contains a block diagram of the AV management system 3800. The AV management system 3800 includes a map store 3802, a web server 3810, an API management module 3808, an offline sensor calibration module 3812, and a teleoperator sensor interface module 3814. In other embodiments, the AV management system 3800 may include additional, fewer, or different modules for various applications. Conventional components such as network interfaces, security functions, load balancers, failover servers, management and network operations consoles, and the like are not shown so as to not obscure the details of the system.

[0176] The web server 3810 links the AV management system 3800 via the network 3804 to one or more user devices 3806; the web server 3810 serves web pages, as well as other web-related content, such as Java, Flash, XML, and so forth. The web server 3810 may provide the functionality of receiving and routing messages between the AV management system 3800 and the user devices 3806, for example, instant messages, queued messages (e.g., email), text and SMS (short message service) messages, or messages sent using any other suitable messaging technique. The user can send a request to the web server 3810 to provide information, for example, images or videos that are stored in the AV management system 3800 for viewing by the user device(s) 3806. Additionally, the web server 3810 may provide API functionality to send data directly to native user device operating systems, such as iOS, ANDROID, webOS, and RIM.

[0177] An API management module 3808 may manage one or more adapters needed for the AV management system 3800 to communicate with various systems, such as teleoperator systems 3750 and user devices 3806. Application programming interfaces (APIs), or adapters, may be used to push data to external tools, websites, and user devices 3806. Adapters may also be used to receive data from the external systems. In one embodiment, the API management module 3808 manages the amount of connections to the external systems needed to operate efficiently.

[0178] A teleoperator sensor interface module 3814 may provide an interface for teleoperator systems 3750 to interact with and provide guidance to AV systems 3602. In conjunction with the offline sensor calibration module 3812, the teleoperator sensor interface module 3814 may provide one or more selected strategies to a teleoperator system 3750 that has been requested to provide assistance to an AV system 3602 experiencing a sensor anomaly and/or miscalibration. For example, an AV system 3602 may detect that a sensor is not operating within normal parameters based on data generated from the sensor not corroborating with other sensor data, such as a lane marking appearing in an unexpected location based on map tile information. In presenting the sensor information from the AV system 3602 to the teleoperator system 3750, the teleoperator sensor interface module 3814 may retrieve log data and/or other information related to the sensor miscalibration. As a result, the teleoperator system 3750 may confirm, through the teleoperator interface provided by the teleoperator sensor interface module 3814, that the sensor is miscalibrated. A suggested course of action, such as whether to allow the AV system 3602 to self-calibrate using the methods described above, whether to continue operation and track the sensor drift, or whether to end the current route and bring the AV system 3602 in for service and offline calibration. The course of action may be selected through the teleoperator interface, in one embodiment. In another embodiment, the AV system 3602 may automatically identify a course of action based on the confirmed sensor miscalibration.

[0179] An offline sensor calibration module 3812 may analyze and provide calibration parameters computed offline to AV systems 3602 based on received log data and/or indications from the AV systems 3602 that one or more sensors are miscalibrated, in one embodiment. The offline sensor calibration module 3812 may process the data gathered from multiple AV systems 3602 offline and may rely on various probabilistic techniques, Bayesian inference, and machine learning algorithms to identify optimal strategies in responding to sensor calibration issues over time. A map store 3802 may store updated map tiles generated by AV systems 3602. This map tile data may be used by the offline sensor calibration module 3812 in calibrating one or more sensors of an AV system 3602 that require calibration. This “fused” map data may be used in generating probabilistic models to aid in generating optimal calibration parameters in a similar way as described above, but in an offline mode while the AV system 3602 is not in operation.

[0180] FIG. 39 is a high-level flow diagram illustrating a process for calibration for autonomous vehicle operation, according to some examples. Data associated with a sensor measurement of a perceived object is received 3900. At least one label associated with the perceived object is determined 3902 based on an initial calibration. Log data associated with the label is retrieved 3904. A calibration parameter associated with the sensor measurement is determined 3906 based on the retrieved log file data. The calibration parameter is then stored 3908 in association with a sensor associated with the sensor measurement.

[0181] FIGS. 40 and 41 illustrate exemplary computing platforms disposed in devices configured to calibrate autonomous vehicles in accordance with various embodiments. In some examples, computing platforms 4000 and 4100 may be used to implement computer programs, applications, methods, processes, algorithms, or other software to perform the above-described techniques.

[0182] In some cases, computing platform can be disposed in wearable device or implement, a mobile computing device 4090b or 4190b, or any other device, such as a computing device 4090a or 4190a.

[0183] Computing platform 4000 or 4100 includes a bus 4004 or 4104 or other communication mechanism for communicating information, which interconnects subsystems and devices, such as processor 4006 or 4106, system memory 4010 or 4110 (e.g., RAM, etc.), storage device 4008 or 4108 (e.g., ROM, etc.), a communication interface 4012 or 4112 (e.g., an Ethernet or wireless controller, a Bluetooth controller, etc.) to facilitate communications via a port on communication link 4014 or 4114 to communicate, for example, with a computing device, including mobile computing and/or communication devices with processors. Processor 4006 or 4106 can be implemented with one or more central processing units (“CPUs”), such as those manufactured by Intel® Corporation, or one or more virtual processors, as well as any combination of CPUs and virtual processors. Computing platform 4000 or 4100 exchanges data representing inputs and outputs via input-and-output devices 4002 or 4102, including, but not limited to, keyboards, mice, audio inputs (e.g., speech-to-text devices), user interfaces, displays, monitors, cursors, touch-sensitive displays, LCD or LED displays, and other I/O-related devices.

[0184] According to some examples, computing platform 4000 or 4100 performs specific operations by processor 4006 or 4106 executing one or more sequences of one or more instructions stored in system memory 4010 or 4110, and computing platform 4000 or 4100 can be implemented in a client-server arrangement, peer-to-peer arrangement, or as any mobile computing device, including smart phones and the like. Such instructions or data may be read into system memory 4010 or 4110 from another computer readable medium, such as storage device 4008 or 4108. In some examples, hard-wired circuitry may be used in place of or in combination with software instructions for implementation. Instructions may be embedded in software or firmware. The term “computer readable medium” refers to any tangible medium that participates in providing instructions to processor 4006 or 4106 for execution. Such a medium may take many forms, including but not limited to, non-volatile media and volatile media. Non-volatile media includes, for example, optical or magnetic disks and the like. Volatile media includes dynamic memory, such as system memory 4010 or 4110.

[0185] Common forms of computer readable media includes, for example, floppy disk, flexible disk, hard disk, magnetic tape, any other magnetic medium, CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with patterns of holes, RAM, PROM, EPROM, FLASH-EPROM, any other memory chip or cartridge, or any other medium from which a computer can read. Instructions may further be transmitted or received using a transmission medium. The term “transmission medium” may include any tangible or intangible medium that is capable of storing, encoding or carrying instructions for execution by the machine, and includes digital or analog communications signals or other intangible medium to facilitate communication of such instructions. Transmission media includes coaxial cables, copper wire, and fiber optics, including wires that comprise bus 4004 or 4104 for transmitting a computer data signal.

[0186] In some examples, execution of the sequences of instructions may be performed by computing platform 4000 or 4100. According to some examples, computing platform 4000 or 4100 can be coupled by communication link 4014 or 4114 (e.g., a wired network, such as LAN, PSTN, or any wireless network, including WiFi of various standards and protocols, Blue Tooth®, Zig-Bee, etc.) to any other processor to perform the sequence of instructions in coordination with (or asynchronous to) one another. Computing platform 4000 or 4100 may transmit and receive messages, data, and instructions, including program code (e.g., application code) through communication link 4014 or 4114 and communication interface 4012 or 4112. Received program code may be executed by processor 4006 or 4106 as it is received, and/or stored in memory 4010 or 4110 or other non-volatile storage for later execution.

[0187] In the example shown, system memory 4010 or 4110 can include various modules that include executable instructions to implement functionalities described herein. System memory 4010 or 4110 may include an operating system (“O/S”) 4030 or 4130, as well as an application 4032 or 4132 and/or logic module 4050 or 4150. In the example shown in FIG. 40, system memory 4010 includes an offline sensor calibration module 3812, an API management module 3808, and a teleoperator sensor interface module 3814. The system memory 4150 shown in FIG. 41 includes a perception system 3702 that includes a calibration detection module 3730, an intrinsic sensor calibration module 3704, an extrinsic sensor calibration module 3706, a data transform module 3708, a generative model module 3710, a heuristics engine module 3712, and a sensor drift tracking module 3714. One or more of the modules included in memory 4010 or 4110 can be configured to provide or consume outputs to implement one or more functions described herein.

[0188] In at least some examples, the structures and/or functions of any of the above-described features can be implemented in software, hardware, firmware, circuitry, or a combination thereof. Note that the structures and constituent elements above, as well as their functionality, may be aggregated with one or more other structures or elements. Alternatively, the elements and their functionality may be subdivided into constituent sub-elements, if any. As software, the above-described techniques may be implemented using various types of programming or formatting languages, frameworks, syntax, applications, protocols, objects, or techniques. As hardware and/or firmware, the above-described techniques may be implemented using various types of programming or integrated circuit design languages, including hardware description languages, such as any register transfer language (“RTL”) configured to design field-programmable gate arrays (“FPGAs”), application-specific integrated circuits (“ASICs”), or any other type of integrated circuit. According to some embodiments, the term “module” can refer, for example, to an algorithm or a portion thereof, and/or logic implemented in either hardware circuitry or software, or a combination thereof. These can be varied and are not limited to the examples or descriptions provided.

[0189] In some embodiments, an autonomous vehicle management system or one or more of its components, or any process or device described herein, can be in communication (e.g., wired or wirelessly) with a mobile device, such as a mobile phone or computing device, or can be disposed therein.

[0190] In some cases, a mobile device, or any networked computing device (not shown) in communication with an action alert controller or one or more of its components (or any process or device described herein), can provide at least some of the structures and/or functions of any of the features described herein. As depicted in the above-described figures, the structures and/or functions of any of the above-described features can be implemented in software, hardware, firmware, circuitry, or any combination thereof. Note that the structures and constituent elements above, as well as their functionality, may be aggregated or combined with one or more other structures or elements. Alternatively, the elements and their functionality may be subdivided into constituent sub-elements, if any. As software, at least some of the above-described techniques may be implemented using various types of programming or formatting languages, frameworks, syntax, applications, protocols, objects, or techniques. For example, at least one of the elements depicted in any of the figure can represent one or more algorithms. Or, at least one of the elements can represent a portion of logic including a portion of hardware configured to provide constituent structures and/or functionalities.

[0191] For example, a perception system 3702 or any of its one or more components, or any process or device described herein, can be implemented in one or more computing devices (i.e., any mobile computing device, such as a wearable device, an audio device (such as headphones or a headset) or mobile phone, whether worn or carried) that include one or more processors configured to execute one or more algorithms in memory. Thus, at least some of the elements in the above-described figures can represent one or more algorithms. Or, at least one of the elements can represent a portion of logic including a portion of hardware configured to provide constituent structures and/or functionalities. These can be varied and are not limited to the examples or descriptions provided.

[0192] As hardware and/or firmware, the above-described structures and techniques can be implemented using various types of programming or integrated circuit design languages, including hardware description languages, such as any register transfer language (“RTL”) configured to design field-programmable gate arrays (“FPGAs”), application-specific integrated circuits (“ASICs”), multi-chip modules, or any other type of integrated circuit.

[0193] For example, an AV management system, including one or more components, or any process or device described herein, can be implemented in one or more computing devices that include one or more circuits. Thus, at least one of the elements in the above-described figures can represent one or more components of hardware. Or, at least one of the elements can represent a portion of logic including a portion of circuit configured to provide constituent structures and/or functionalities.

[0194] According to some embodiments, the term “circuit” can refer, for example, to any system including a number of components through which current flows to perform one or more functions, the components including discrete and complex components. Examples of discrete components include transistors, resistors, capacitors, inductors, diodes, and the like, and examples of complex components include memory, processors, analog circuits, digital circuits, and the like, including field-programmable gate arrays (“FPGAs”), application-specific integrated circuits (“ASICs”). Therefore, a circuit can include a system of electronic components and logic components (e.g., logic configured to execute instructions, such that a group of executable instructions of an algorithm, for example, and, thus, is a component of a circuit). According to some embodiments, the term “module” can refer, for example, to an algorithm or a portion thereof, and/or logic implemented in either hardware circuitry or software, or a combination thereof (i.e., a module can be implemented as a circuit). In some embodiments, algorithms and/or the memory in which the algorithms are stored are “components” of a circuit. Thus, the term “circuit” can also refer, for example, to a system of components, including algorithms. These can be varied and are not limited to the examples or descriptions provided.

[0195] Although the foregoing examples have been described in some detail for purposes of clarity of understanding, the above-described inventive techniques are not limited to the details provided. There are many alternative ways of implementing the above-described invention techniques. The disclosed examples are illustrative and not restrictive.

[0196] The foregoing description of the embodiments of the invention has been presented for the purpose of illustration; it is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Persons skilled in the relevant art can appreciate that many modifications and variations are possible in light of the above disclosure.

[0197] Some portions of this description describe the embodiments of the invention in terms of algorithms and symbolic representations of operations on information. These algorithmic descriptions and representations are commonly used by those skilled in the data processing arts to convey the substance of their work effectively to others skilled in the art. These operations, while described functionally, computationally, or logically, are understood to be implemented by computer programs or equivalent electrical circuits, microcode, or the like. Furthermore, it has also proven convenient at times, to refer to these arrangements of operations as modules, without loss of generality. The described operations and their associated modules may be embodied in software, firmware, hardware, or any combinations thereof.

[0198] Any of the steps, operations, or processes described herein may be performed or implemented with one or more hardware or software modules, alone or in combination with other devices. In one embodiment, a software module is implemented with a computer program product comprising a computer-readable medium containing computer program code, which can be executed by a computer processor for performing any or all of the steps, operations, or processes described.

[0199] Embodiments of the invention may also relate to an apparatus for performing the operations herein. This apparatus may be specially constructed for the required purposes, and/or it may comprise a general-purpose computing device selectively activated or reconfigured by a computer program stored in the computer. Such a computer program may be stored in a non-transitory, tangible computer readable storage medium, or any type of media suitable for storing electronic instructions, which may be coupled to a computer system bus. Furthermore, any computing systems referred to in the specification may include a single processor or may be architectures employing multiple processor designs for increased computing capability.

[0200] Embodiments of the invention may also relate to a product that is produced by a computing process described herein. Such a product may comprise information resulting from a computing process, where the information is stored on a non-transitory, tangible computer readable storage medium and may include any embodiment of a computer program product or other data combination described herein.

[0201] Finally, the language used in the specification has been principally selected for readability and instructional purposes, and it may not have been selected to delineate or circumscribe the inventive subject matter. It is therefore intended that the scope of the invention be limited not by this detailed description, but rather by any claims that issue on an application based hereon. Accordingly, the disclosure of the embodiments of the invention is intended to be illustrative, but not limiting, of the scope of the invention, which is set forth in the following claims.

Claims

What is claimed:

1. A method comprising:

receiving data at an autonomous vehicle system comprising a plurality of sensors, the plurality of sensors including an inertial measurement unit (IMU) sensor, an odometry sensor, and a multi-beam LIDAR sensor, and the data representing measurements made using the plurality of sensors;

determining a distance traveled by the autonomous vehicle system based on one or more sensor measurements of the odometry sensor;

determining a velocity of the autonomous vehicle system based on one or more sensor measurements of the IMU sensor;

identifying an abnormal sensor measurement, made using the multi-beam LIDAR sensor, based at least in part on the data;

generating an initial calibration parameter, associated with the multi-beam LIDAR sensor, based at least in part on the abnormal sensor measurement;

generating an expected sensor measurement, associated with the multi-beam LIDAR sensor, based on the data;

modifying the initial calibration parameter based at least in part on the expected sensor measurement, wherein modifying the initial calibration parameter includes generating a modified calibration parameter associated with the multi-beam LIDAR sensor; and

calibrating the multi-beam LIDAR sensor based at least in part on the modified calibration parameter, the distance traveled by the autonomous vehicle system, and the velocity of the autonomous vehicle system.

2. The method of claim 1, wherein the plurality of sensors includes an array of LIDAR sensors arranged at one end of the autonomous vehicle system,

the array including the multi-beam LIDAR sensor.

3. The method of claim 2, further comprising:

determining an extrinsic calibration of the multi-beam LIDAR sensor based on a physical location of the multi-beam LIDAR sensor relative to the autonomous vehicle system.

4. The method of claim 2, further comprising:

determining an intrinsic calibration of the multi-beam LIDAR sensor based on data associated with at least one additional LIDAR sensor of the array disposed adjacent to the multi-beam LIDAR sensor.

5. The method of claim 1, further comprising:

retrieving one or more log files, including information indicative of past sensor measurements, based on the abnormal sensor measurement; and

generating the expected sensor measurement based, in part, on the information indicative of past sensor measurements.

6. The method of claim 1, wherein the expected sensor measurement is further based at least in part on a teleoperation interface selection made via a teleoperator system.

7. The method of claim 1, wherein the expected sensor measurement is further based at least in part on a map tile associated with a pose of the autonomous vehicle system at a time of the abnormal sensor measurement.

8. The method of claim 1, wherein the initial calibration parameter is generated based at least in part on a first probabilistic model, the method further comprising:

generating a second probabilistic model; and

generating an additional expected sensor measurement, associated with the multi-beam LIDAR sensor, using the second probabilistic model, wherein the additional expected sensor measurement is generated based on a measurement made using at least one additional sensor of the plurality of sensors.

9. A method comprising:

receiving, at an autonomous vehicle system comprising a plurality of sensors, a plurality of sensor measurements associated with an environmental feature, the plurality of sensors comprising an inertial measurement unit (IMU) sensor, an odometry sensor, and a multi-beam LIDAR sensor;

determining a distance traveled by the autonomous vehicle system based on one or more sensor measurements of the odometry sensor;

determining a velocity of the autonomous vehicle system based on one or more sensor measurements of the IMU sensor;

identifying an abnormal sensor measurement of the plurality of sensor measurements made using the multi-beam LIDAR sensor, wherein the abnormal sensor measurement is incongruent with one or more other measurements of the plurality of sensor measurements made using the multi-beam LIDAR sensor;

determining a plurality of calibration parameters associated with the multi-beam LIDAR sensor, wherein at least one parameter of the plurality of calibration parameters is determined based on the abnormal sensor measurement and a statistical computation associated with the environmental feature;

determining an expectation model associated with the multi-beam LIDAR sensor, wherein the expectation model is configured to output one or more expected sensor measurements;

generating an expected sensor measurement with the expectation model and using at least one measurement of the plurality of sensor measurements made using the multi-beam LIDAR sensor;

modifying the plurality of calibration parameters based at least in part on the expected sensor measurement, wherein modifying the plurality of calibration parameter includes generating a modified calibration parameter associated with the multi-beam LIDAR sensor; and

10. The method of claim 9, further comprising:

comparing the plurality of sensor measurements to a predetermined sensor measurement range; and

identifying the abnormal sensor measurement based at least in part on the comparing.

11. The method of claim 9, wherein the abnormal sensor measurement is incongruent with the one or more other measurements of the plurality of sensor measurements based on the abnormal sensor measurement not matching the one or more other measurements of the plurality of sensor measurements.

12. The method of claim 9, wherein the statistical computation associated with the environmental feature comprises an average of the plurality of sensor measurements associated with the environmental feature.

13. The method of claim 9, wherein the statistical computation associated with the environmental feature comprises a median of the plurality of sensor measurements associated with the environmental feature.

14. The method of claim 9, wherein modifying the plurality of calibration parameters comprises:

determining a heuristic rule to identify a subset of calibration parameters based on the expectation model and the plurality of calibration parameters;

identifying the subset of calibration parameters using the heuristic rule; and

replacing the plurality of calibration parameters with the subset of calibration parameters.

15. A method comprising:

receiving, at an autonomous vehicle system comprising a plurality of sensors, data associated with a measurement made by a multi-beam LIDAR sensor of the plurality of sensors, wherein the measurement indicates detection of an object by the multi-beam LIDAR sensor, and the plurality of sensors further comprises at least one RADAR sensor;

determining a current physical location of the multi-beam LIDAR sensor, relative to the autonomous vehicle system, using the at least one RADAR sensor;

comparing the current physical location to a stored indication of an initial position of the multi-beam LIDAR sensor;

determining a labeled data point associated with the object, wherein the labeled data point is determined based at least in part on an initial calibration of the multi-beam LIDAR sensor;

retrieving log file data associated with the labeled data point, the log file data comprising one or more contradictory sensor measurements from the multi-beam LIDAR sensor;

determining, based at least in part on the one or more contradictory sensor measurements, an expected sensor measurement of the multi-beam LIDAR sensor;

determining a calibration parameter associated with the multi-beam LIDAR sensor based at least in part on the retrieved log file data and the expected sensor measurement;

associating the calibration parameter with information, identifying the multi-beam LIDAR sensor, in a memory associated with the autonomous vehicle system; and

calibrating the multi-beam LIDAR sensor based at least in part on the calibration parameter and on the comparing.

16. The method of claim 15, wherein:

the calibration parameter is determined based at least in part on at least one generative probabilistic model,

the measurement is made by the multi-beam LIDAR sensor at a particular time, and

the one or more contradictory sensor measurements are made concurrently with the measurement of the object.

17. The method of claim 15, wherein:

the calibration parameter is determined based at least in part on a generative probabilistic model using a plurality of fused sensor measurements from one or more heterogeneous sensors,

the measurement is made by the multi-beam LIDAR sensor at a particular time, and

each measurement of the plurality of fused sensor measurements is made concurrently with the measurement of the object.

18. The method of claim 15, wherein:

the calibration parameter is determined based at least in part on a generative probabilistic model using map tile data,

the map tile data comprises information indicative of past measurements of the object, and

the past measurements were made using one or more heterogeneous sensors.

Patent Citations (100)

Patent	Date	Inventor	Cited By
US4280188(A)	1981-07-01	Weinstein	Examiner
US5558370(A)	1996-09-01	Behr	Applicant
US5959552(A)	1999-09-01	Cho	Applicant
US6301542(B1)	2001-10-01	Kirchberger et al.	Applicant
US6374168(B1)	2002-04-01	Fujii	Applicant
US6728616(B1)	2004-04-01	Tabe	Applicant
US7426429(B2)	2008-09-01	Tabe	Applicant
US8123642(B2)	2012-02-01	Ishikawa et al.	Applicant
US8126642(B2)	2012-02-01	Trepagnier	Applicant
US8392064(B2)	2013-03-01	Thrun et al.	Applicant
US8434919(B2)	2013-05-01	Schofield	Applicant
US8880272(B1)	2014-11-01	Ferguson et al.	Applicant
US8965578(B2)	2015-02-01	Versteeg et al.	Applicant
US8977007(B1)	2015-03-01	Ferguson et al.	Applicant
USD743978(S)	2015-11-01	Amin	Applicant
US9384443(B2)	2016-07-01	Passot et al.	Applicant
US9395727(B1)	2016-07-01	Smith et al.	Applicant
US9719801(B1)	2017-08-01	Ferguson	Examiner
US2004/0017073(A1)	2004-01-01	Pavlov et al.	Applicant
US2004/0068354(A1)	2004-04-01	Tabe	Applicant
US2005/0062615(A1)	2005-03-01	Braeuchle	Examiner
US2005/0216181(A1)	2005-09-01	Estkowski et al.	Applicant
US2006/0064202(A1)	2006-03-01	Gutmann et al.	Applicant
US2006/0175116(A1)	2006-08-01	Friedman et al.	Applicant
US2006/0207820(A1)	2006-09-01	Joshi et al.	Applicant
US2007/0096447(A1)	2007-05-01	Tabe	Applicant
US2007/0246927(A1)	2007-10-01	Okada	Applicant
US2008/0033645(A1)	2008-02-01	Levinson et al.	Applicant
US2008/0084283(A1)	2008-04-01	Kalik	Applicant
US2008/0097699(A1)	2008-04-01	Ono	Applicant
US2008/0320421(A1)	2008-12-01	Demaris et al.	Applicant
US2009/0276149(A1)	2009-11-01	Kauffman et al.	Applicant
US2010/0045482(A1)	2010-02-01	Strauss	Applicant
US2010/0235129(A1)	2010-09-01	Sharma et al.	Applicant
US2010/0275829(A1)	2010-11-01	Sporsheim	Applicant
US2011/0130111(A1)	2011-06-01	Crandall et al.	Applicant
US2011/0288684(A1)	2011-11-01	Farlow et al.	Applicant
US2012/0083960(A1)	2012-04-01	Zhu et al.	Applicant
US2012/0256448(A1)	2012-10-01	Yasui et al.	Applicant
US2012/0290169(A1)	2012-11-01	Zeng	Examiner
US2013/0054133(A1)	2013-02-01	Lewis et al.	Applicant
US2013/0246301(A1)	2013-09-01	Radhakrishnan et al.	Applicant
US2013/0268138(A1)	2013-10-01	Moughler et al.	Applicant
US2014/0032012(A1)	2014-01-01	Joshi et al.	Applicant
US2014/0032049(A1)	2014-01-01	Moshchuk et al.	Applicant
US2014/0129135(A1)	2014-05-01	Holden et al.	Applicant
US2014/0129302(A1)	2014-05-01	Amin et al.	Applicant
US2014/0129951(A1)	2014-05-01	Amin et al.	Applicant
US2014/0142830(A1)	2014-05-01	Bernzen et al.	Applicant
US2014/0152971(A1)	2014-06-01	James	Examiner
US2014/0185880(A1)	2014-07-01	Fairfield et al.	Applicant
US2014/0188347(A1)	2014-07-01	Tabe	Applicant
US2014/0195138(A1)	2014-07-01	Stelzig	Examiner
US2014/0214255(A1)	2014-07-01	Dolgov et al.	Applicant
US2014/0214260(A1)	2014-07-01	Eckert et al.	Applicant
US2014/0244151(A1)	2014-08-01	Matsubara et al.	Applicant
US2014/0257661(A1)	2014-09-01	Schulman et al.	Applicant
US2014/0303827(A1)	2014-10-01	Dolgov et al.	Applicant
US2014/0309833(A1)	2014-10-01	Ferguson et al.	Applicant
US2014/0333468(A1)	2014-11-01	Zhu et al.	Applicant
US2014/0336935(A1)	2014-11-01	Zhu et al.	Applicant
US2014/0358427(A1)	2014-12-01	Fuhrman	Applicant
US2015/0025708(A1)	2015-01-01	Anderson	Applicant
US2015/0039157(A1)	2015-02-01	Wolfe et al.	Applicant
US2015/0039391(A1)	2015-02-01	Hershkovitz et al.	Applicant
US2015/0057871(A1)	2015-02-01	Ono et al.	Applicant
US2015/0091374(A1)	2015-04-01	Lenius et al.	Applicant
US2015/0094850(A1)	2015-04-01	Passot et al.	Applicant
US2015/0127224(A1)	2015-05-01	Tabe	Applicant
US2015/0127239(A1)	2015-05-01	Breed et al.	Applicant
US2015/0149088(A1)	2015-05-01	Attard et al.	Applicant
US2015/0248131(A1)	2015-09-01	Fairfield et al.	Applicant
US2015/0254986(A1)	2015-09-01	Fairfield et al.	Applicant
US2015/0258928(A1)	2015-09-01	Goto et al.	Applicant
US2015/0268665(A1)	2015-09-01	Ludwick et al.	Applicant
US2015/0271290(A1)	2015-09-01	Tao et al.	Applicant
US2015/0292894(A1)	2015-10-01	Goddard et al.	Applicant
US2015/0298636(A1)	2015-10-01	Furst	Applicant
US2015/0336502(A1)	2015-11-01	Hillis et al.	Applicant
US2015/0336524(A1)	2015-11-01	Larner et al.	Applicant
US2015/0338852(A1)	2015-11-01	Ramanujam	Applicant
US2015/0346727(A1)	2015-12-01	Ramanujam	Applicant
US2015/0348221(A1)	2015-12-01	Pedersen et al.	Applicant
US2015/0356368(A1)	2015-12-01	Liu et al.	Applicant
US2015/0359032(A1)	2015-12-01	Menard et al.	Applicant
US2015/0375764(A1)	2015-12-01	Rajendran et al.	Applicant
US2016/0016315(A1)	2016-01-01	Kuffner, Jr. et al.	Applicant
US2016/0071278(A1)	2016-03-01	Leonard et al.	Applicant
US2016/0107703(A1)	2016-04-01	Briceno et al.	Applicant
US2016/0159402(A1)	2016-06-01	Nakaya et al.	Applicant
US2016/0159407(A1)	2016-06-01	Holmstrom et al.	Applicant
US2016/0165775(P1)	2016-06-01	Maillard et al.	Applicant
US2016/0165786(P1)	2016-06-01	Giesen	Applicant
US2016/0189544(A1)	2016-06-01	Ricci	Applicant
US2016/0209220(A1)	2016-07-01	Laetz	Applicant
US2016/0247106(A1)	2016-08-01	Dalloro et al.	Applicant
US2016/0247109(A1)	2016-08-01	Scicluna et al.	Applicant
US2016/0358475(A1)	2016-12-01	Prokhorov	Examiner
RU2269813	2005-10-01		Applicant
WO2009151781(A1)	2009-12-01		Applicant

Non-Patent Literature (71)

Bodensteiner et al., “Monocular Camera Trajectory Optimization using LiDAR Data”, IEEE International Conference on Computer Vision Workshops, 2011, 8 pages.Applicant
Office action for U.S. Appl. No. 14/932,962, dated Mar. 21, 2016, Kentley et al., “Robotic Vehicle Active Safety Systems and Methods”, 18 pages.Applicant
Office action for U.S. Appl. No. 14/932,963, dated Mar. 17, 2016, Levinson et al., “Adaptive Mapping to Navigate Autonomous Vehicles Responsive to Physical Environment Changes”, 26 pages.Applicant
Office action for U.S. Appl. No. 14/932,940, dated May 4, 2016, Levinson et al., “Automated Extraction of Semantic Information to Enhance Incremental Mapping Modifications for Robotic Vehicles”, 22 pages.Applicant
Combining 3D Shape, Color, and Motion for Robust Antime Tracking; Held, David, Levinson, Jesse, Thrun, Sebastian, Savarese, Silvio, Robotics: Science and Systems (RSS), Berkeley, California, USA (2014).Applicant
Group Induction; Teichman, Alex, Thrun, Sebastian, Proc. of the IEEE/RSJ Intl Conf on Intelligent Robotics and Systems (IROS) (2013).Applicant
Automatic Online Calibration of Cameras and Lasers; Levinson, Jesse, Thrun, Sebastian; Robotics: Science and Systems (RSS) (2013).Applicant
Precision Tracking With Sparse 3D and Dense Color 2D Data; Held, David, Levinson, Jesse, Thrun, Sebastian; International Conference on Robotics and Automation (ICRA) (2013).Applicant
Automatic Calibration of Cameras and Lasers in Arbitrary Scenes; Levinson, Jesse, Thrun, Sebastian; International Symposium on Experimental Robotics (ISER) (2012).Applicant
Online, Semi-Supervised Learning for Long-Term Interaction With Object Recognition Systems; Teichman, Alex, Thrun, Sebastian, RSS Workshop on Long-Term Operation of Autonomous Robotic Systems in Changing Environments (2012).Applicant
Tracking-Based Semi-Supervised Learning; Teichman, Alex, Thrun, Sebastian; International Journal of Robotics Research (IJRR); http://ijr.sagepub.com/content/31/7/804; (2012).Applicant
A Probabilistic Framework for Object Detection in Images Using Context and Scale; Held, David, Levinson, Jesse, Thrun, Sebastian; International Conference on Robotics and Automation (ICRA) (2012).Applicant
Practical Object Recognition in Autonomous Driving and Beyond; Teichman, Alex, Thrun, Sebastian; IEEE Workshop on Advanced Robotics and its Social Impacts (ARSO) (2011).Applicant
Tracking-Based Semi-Supervised Learning; Teichman, Alex, Thrun, Sebastian; Robotics: Science and Systems (RSS) (2011).Applicant
Towards 3D Object Recognition via Classification of Arbitrary Object Tracks; Teichman, Alex, Levinson, Jesse, Thrun, Sebastian; International Conference on Robotics and Automation (ICRA) (2011).Applicant
Towards Fully Autonomous Driving: Systems and Algorithms; Levinson, Jesse et al.; Intelligent Vehicles Symposium (2011).Applicant
Traffic Light Mapping, Localization, and State Detection for Autonomous Vehicles; Levison, Jesse, Askeland, Jake, Dolson, Jennifer, Thrun, Sebastian; International Conference on Robotics and Automation (ICRA) (2011).Applicant
Automatic Laser Calibration, Mapping, and Localization for Autonomous Vehicles, Levison, Jesse; Thesis (Ph D); Stanford University (2011).Applicant
Unsupervised Calibration for Multi-Beam Lasers; Levinson, Jesse, Thrun, Sebastian, International Symposium on Experimental Robotics (ISER) (2010).Applicant
Robust Vehicle Localization in Urban Environments Using Probabilistic Maps; Levinson, Jesse, Thrun, Sebastian; International Conference on Robotics and Automation (ICRA) (2010).Applicant
Exponential Family Sparse Coding With Application to Self-Taught Learning; Honglak, Lee, Raina, Rajat, Teichman, Alex, Ng, Andrew Y.; International Joint Conference on Artificial Intelligence (IJCAI) (2009).Applicant
Map-Based Precision Vehicle Localization in Urban Environments; Levinson, Jesse, Thrun, Sebastian; Robotics: Science and Systems (RSS) (2007).Applicant
Stanford Autonomous Driving Team website <http://driving.stanford.edu/papers.html>; Various; Stanford University 2014.Applicant
Combining Data-Driven and Model-Based Cues for Segmentation of Video Sequences; Eckes, Christian, Vorbruggen, Jan C.; Proceedings WCNN '96, San Diego, USA (1996).Applicant
A Real-Time Motion Planner With Trajectory Optimization for Autonomous Vehicles; Xu, Wenda et al.; Robotics and Automation (ICRA); Saint Paul, MN, USA (2012).Applicant
Dynamic Real-Time Replanning in Belief Space: An Experimental Study on Physical Mobile Robots; Agha-mohammadi, Ali-Akbar et al.; Technical Report TR 13-007; Texas A&M University, USA (2013).Applicant
An Evaluation of Dynamic Object Tracking With 3D Lidar; Morton, P., Douillard, B., Underwood, J.; Proceedings of Australasian Conference on Robotics and Automation; Dec. 7-9, 2011; Melbourne, Australia (2011).Applicant
Control of Robotic Mobility-on Demand Systems: A Queueing-Theoretical Perspective; Zhang, Rick, Pavone, Marco; Intl Journal of Robotics Research, pp. 1-18, Stanford, USA (2015).Applicant
Evaluation of Urban Vehicle Routing Algorithms; Kong, Linghe et al.; Intl Journal of Digital Content Technology and its Applications (JDCTA); vol. 6, No. 23, University of New Mexico, USA (2012).Applicant
Real-Time High Resolution Fusion of Depth Maps on GPU; Trifonov, Dmitry; Intl Conference on Computer-Aided Design and Computer Graphics (CAD/Graphics); Guangzhou, China (2013).Applicant
Real-Time Image Segmentation on a GPU; Abramov, Alexey et al.; Facing the Multi-Core Challenge, pp. 131-142, Berlin, German (2010).Applicant
A Tutorial on Graph-Based Slam; Grisetti, Giorgio et al.; Intelligent Transportation Systems Magazine, IEEE; pp. 31-43 (2010).Applicant
Sensor Fusion and Calibration of Inertial Sensors, Vision, Ultra-Wideband and GPS; Hol, Jeroen; Linkoping Studies in Science and Technology, Dissertations No. 1368; Department of Electrical Engineering; Linkoping University, SE-581 83 Linkoping, Sweden; (2011).Applicant
Large Scale Dense Visual Inertial Slam; Ma, Lu et al.; Field and Service Robotics (FSR); (2015).Applicant
Simultaneous Localization, Mapping, and Manipulation for Unsupervised Object Discovery; Ma, Lu et al.; IEEE International Conference on Robotics and Automation (ICRA); (2014).Applicant
Online Slam With Any-Time Self-Calibration and Automatic Change Detection; Nima Keivan and Gabe Sibley; IEEE International Conference on Robotics and Automation (ICRA); (2014).Applicant
Bayazit et al., “Swarming Behavior Using Probabilistic Roadmap Techniques”, Swarm Robotics WS 2004, LNCS, Springer-Verlag Berlin Heidelberg 2005, pp. 112-125.Applicant
Easymile (website), Retrieved from <<https://web.archive.org/web/20150723060050/http://easymile.com>> Jul. 2015, <<https://web.archive.org/web/201508012054107/http://easymile.com/mobility-soulition/>>, Aug. 2015, and <<http:www.youtube.com/watch?v=fijDBL76yDY>>, Feb. 2015, 13 pages.Applicant
Office action for U.S. Appl. No. 14/932,963, dated Aug. 15, 2016, Levinson et al., “Adaptive Mapping to Navigate Autonomous Vehicles Responsive to Physical Environment Changes”, 15 pages.Applicant
Office action for U.S. Appl. No. 14/932,959, dated Aug. 5, 2016, Kentley et al., “Autonomous Vehicle Fleet Service and System”, 16 pages.Applicant
Office action for U.S. Appl. No. 14/756,992, dated Aug. 25, 2016, Levinson et al., “Adaptive autonomous vehicle planner logic”, 9 pages.Applicant
Office action for U.S. Appl. No. 14/933,469, dated Aug. 30, 2016, Kentley et al., “Software Application to Request and Control an Autonomous Vehicle Service”, 11 pages.Applicant
Office Action for U.S. Appl. No. 14/932,959, dated Dec. 2, 2016, Kentley et al., “Autonomous Vehicle Fleet Service and System”, 21 pages.Applicant
U.S. Appl. No. 14/756,991, filed Nov. 4, 2015, Levinson et al., “Sensor-based object-detection optimization for autonomous vehicles”, 127 pages.Applicant
U.S. Appl. No. 14/756,992, filed Nov. 4, 2015, Levinson et al., “Adaptive autonomous vehicle planner logic”, 119 pages.Applicant
U.S. Appl. No. 14/756,993, filed Nov. 4, 2015, Kentley et al., “Method for robotic vehicle communication with an external environment via acoustic beam forming”, 130 pages.Applicant
U.S. Appl. No. 14/756,994, filed Nov. 4, 2015, Kentley et al., “System of configuring active lighting to indicate directionality of an autonomous vehicle”, 141 pages.Applicant
U.S. Appl. No. 14/756,995, filed Nov. 4, 2015, Kentley et al., “Coordination of dispatching and maintaining fleet of autonomous vehicles”, 134 pages.Applicant
U.S. Appl. No. 14/757,015, filed Nov. 5, 2015, Levinson et al., “Independent steering, power torque control and transfer in autonomous vehicles”, 27 pages.Applicant
U.S. Appl. No. 14/757,016, filed Nov. 5, 2015, Levinson et al., “Simulation system and methods for autonomous vehicles”, 127 pages.Applicant
U.S. Appl. No. 14/932,940, filed Nov. 4, 2015, Levinson et al., “Automated Extraction of Semantic Information to Enhance Incremental Mapping Modifications for Robotic Vehicles”, 130 pages.Applicant
U.S. Appl. No. 14/932,948, filed Nov. 4, 2015, Kentley et al., “Active Lighting Control for Communicating a State of an Autonomous Vehicle to Entities in a Surrounding Environment”, 123 pages.Applicant
U.S. Appl. No. 14/932,952, filed Nov. 4, 2015, Kentley et al., “Resilient Safety System for a Robotic Vehicle”, 125 pages.Applicant
U.S. Appl. No. 14/932,954, filed Nov. 4, 2015, Kentley et al., “Internal Safety Systems for Robotic Vehicles”, 127 pages.Applicant
U.S. Appl. No. 14/932,958, filed Nov. 4, 2015, Kentley, “Quadrant Configuration of Robotic Vehicles”, 57 pages.Applicant
U.S. Appl. No. 14/932,959, filed Nov. 4, 2015, Kentley et al., Titled “Autonomous Vehicle Fleet Service and System”, 65 pages.Applicant
U.S. Appl. No. 14/932,962, filed Nov. 4, 2015, Kently et al., “Robotic Vehicle Active Safety Systems and Methods”, 116 pages.Applicant
U.S. Appl. No. 14/932,966, filed Nov. 4, 2015, Levinson et al., “Teleoperation System and Method for Trajectory Modification of Autonomous Vehicles”, 130 pages.Applicant
U.S. Appl. No. 14/933,469, filed Nov. 5, 2015, Kentley eta al., “Software Application to Request and Control an Autonomous Vehicle Service”, 146 pages.Applicant
U.S. Appl. No. 14/933,602, filed Nov. 5, 2015, Levinson et al., “Machine-Learning Systems and Techniques to Optimize Teleoperation and/or Planner Decisions”, 123 pages.Applicant
U.S. Appl. No. 14/933,665, filed Nov. 5, 2015, Kentley et al., “Software Application and Logic to Modify Configuration of an Autonomous Vehicle”, 144 pages.Applicant
U.S. Appl. No. 14/933,706, filed Nov. 5, 2015, Kentley et al., “Interactive Autonomous Vehicle Command Controller”, 145 pages.Applicant
Stanford Autonomous Driving Team website <http://driving.stanford.edu/papers.html>; Various; Stanford University 2014, 1 page.Applicant
Sensor Fusion and Calibration of Inertial Sensors, Vision, Ultra-Wideband and GPS; Hol, Jeroen; Linkoping Studies in Science and Technology, Dissertations No. 1368; Department of Electrical Engineering; Linkoping University, SE-581 83 Linkoping, Sweden; (2011 ).Applicant
Office Action for U.S. Appl. No. 14/932,948, dated Oct. 14, 2016, Kentley et al., “Active Lighting Control for Communicating a State of an Autonomous Vehicle to Entities in a Surrounding Environment”, 15 pages.Applicant
Office Action for U.S. Appl. No. 14/756,995, dated Oct. 31, 2016, Kentley et al., “Coordination of dispatching and maintaining fleet of autonomous vehicles”, 35 pages.Applicant
Office Action for U.S. Appl. No. 14/932,954, dated Mar. 29, 2016, Kentley et al., “Internal Safety Systems for Robotic Vehicles”, 17 pages.Applicant
Office Action for U.S. Appl. No. 14/933,602, dated Nov. 5, 2015, Levinson et al., “Machine-Learning Systems and Techniques to Optimize Teleoperation and/or Planner Decisions”, 11 pages.Applicant
Office Action for U.S. Appl. No. 14/932,952, dated Jun. 24, 2016, Kentley et al., “Resilient Safety System for a Robotic Vehicle”, 11 pages.Applicant
“GPS Logger”, retrieved on Jan. 18, 2017 at <<http://www.gpslogger.ru/odometer>>, Sputnikovy monitoring transporta, kontrol raskhoda topliva, Vologda, Sposoby opredeleniya probega avtotransporta. May 31, 2012.Applicant
PCT Search Report and Written Opinion dated Mar. 30, 2017 for PCT Application No. PCT/US16/60118, 8 pages.Applicant

Abstract