Mobile Asset Data Recorder and Transmitter

[0001] This application claims priority to U.S. Provisional Patent Application Ser. No. 61/624,142, filed on Apr. 13, 2012, U.S. Non-provisional patent application Ser. No. 13/861,826, filed Apr. 12, 2013, and U.S. Non-provisional patent application Ser. No. 14/608,423, filed Jan. 29, 2015, to the extent allowed by law.

[0002] This disclosure relates to equipment used in high value assets and particularly, to event and data recorder systems used in high value assets.

[0003] High value assets such as locomotives, mining, cargo, marine and military vehicles and vessels typically employ an onboard data acquisition and logging device, similar to a “black box” on airplanes. A typical onboard data acquisition and logging device, or an event/data recorder, comprises digital and analog inputs as well as pressure switches and pressure transducers which record data from various onboard sensor devices. These event/data recorders log a variety of system parameters used for incident investigation, crew performance evaluation, fuel efficiency analysis, maintenance planning, and predictive diagnostics. Recorded data may include such parameters as speed, distance traveled, location, fuel level, engine revolutions per minute (RPM), fluid levels, operator controls, pressures, and ambient conditions. In addition to the basic event and operational data, video and audio event/data recording capabilities are also deployed on many of these same mobile assets.

[0004] The primary objects of the present disclosure are to provide automatic collision detection, to reduce loss due to rough switching and train operations, to reduce fuel costs by eliminating excess idle, to enhance positioning accuracy, to improve over-the-road fuel accuracy, and to provide real-time track monitoring.

[0005] A further object of the present disclosure is remote accident alerting combining collision detection, roll-over detection, video and logging of operator actions before and after incidents to provide complete incident analysis.

[0006] Another object of the present disclosure is using accelerometer-based pitch and roll measurements to provide accurate liquid levels in mobile fuel storage tanks.

[0007] A further object of the present disclosure is the use of a non-intrusive accelerometer to determine the running/not-running state of an engine and auxiliary loads on that engine.

[0008] Another object of the present disclosure is the use of accelerometer data combined with logged operator actions and GPS location data to precisely locate rough operating environment, such as bad track, rough seas, and poor roads.

[0009] Yet another object of the present disclosure is the use of a high accuracy GPS, a 3-axis digital accelerometer, a digital compass and a 3-axis digital gyroscope combined to provide dead reckoning based arrival and departure notifications for mobile assets in conditions where no GPS signal is available, such as under canopies or overhangs at stations and docks.

[0010] Disclosed herein is an embodiment of an acceleration-based mobile asset data recorder and transmitter used on locomotives that comprises the operational integration of nine components. The components are an event recorder similar to a black box on airplanes, a locomotive digital video recorder, a fuel level sensor, fuel level sensor software, a wireless processing unit, an inertial navigation sensor board, firmware, system software, and the system encompassing these components. The inertial navigation sensor board includes a 3-axis digital gyroscope, a 3-axis digital magnetometer, a 3-axis digital accelerometer, and a micro-controller. The gyroscope is used for measuring the angular acceleration and deceleration of the asset, the magnetometer is used for measuring magnetic fields, the accelerometer is used for measuring linear accelerations and decelerations, and the micro-controller is used for processing data and communicating between the sensors and the wireless processing unit.

[0011] The mobile asset data recorder and transmitter performs seven functions: automatic orientation, automatic compass calibration, fuel compensation with pitch and roll, emergency brake with impact detection, rough operating condition detection, engine running detection and inertial navigation (dead reckoning).

[0012] Automatic collision detection alerts appropriate personnel when an emergency brake application occurs and can instantly determine if a collision coincides with the braking event. The mobile asset data recorder and transmitter provides immediate notification of collision severity including an indication of locomotive derailment or rollover event.

[0013] Rough operating condition detection reduces loss due to rough switching and train operations. It provides alerts and summary reports when high energy impacts are detected during switching operations. It also detects excessive slack-action, allowing supervisors to continuously assess and improve train operations. This enables the reduction of lading and equipment damage by identifying unsafe trends and allowing users to take immediate corrective action. Continuous monitoring of track conditions and over the road monitoring of vibration levels alert track maintenance personnel to the precise location of rough track or switches which may need inspection and repair.

[0014] Accelerometer-based engine running detection may be used as a backup source if the engine running signal is not already accessible from other onboard systems, as a means of reducing fuel costs by eliminating excess idle. It also improves over the road fuel accuracy by compensating for locomotive tilt due to grade and super elevation.

[0015] Fuel compensation with pitch and roll improves fuel reporting accuracy. It provides a simple, universal and non-intrusive method of determining if the engine is running while the locomotive is stopped. Increased accuracy provides enhanced real-time business intelligence to support strategic initiatives such as smart fueling, burn-rate analysis, fuel reconciliation and emissions monitoring.

[0016] Inertial navigation, or dead reckoning, enhances positioning accuracy. It augments the wireless processing unit's high accuracy differential GPS with sophisticated dead reckoning when inside shop buildings, stations, tunnels or any location where GPS signals are not available. This provides highly accurate station arrival and departure times, and the precise positioning and locomotive orientation within shop areas increases operational efficiency by improving shop planning and work flow.

[0017] These and other aspects of the present disclosure are disclosed in the following detailed description of the embodiments, the appended claims and the accompanying figures.

[0025] The mobile asset data recorder and transmitter system of the present disclosure and its components are shown in FIG. 6. The mobile asset data recorder and transmitter system 200 consists of ten interrelated components: an event data recorder 38, a locomotive digital video recorder (DVR) 52, a fuel level sensor 210, fuel level sensor software 212, a WPU 202, an inertial navigation sensor board 214, global positioning system (GPS) 106, firmware 224, system software 226, and the system 200 itself. Installing the WPU 202 onto an asset, such as a locomotive, consists of mounting the WPU 202 and connecting it externally to an event data recorder 38, a locomotive digital video recorder 208 and any additional available condition sensing devices.

[0026] The event data recorder 38, similar to a black-box on airplanes, is an onboard data logging device for locomotives. A typical event data recorder 38 consists of digital and analog inputs as well as pressure switches and pressure transducers which record data from various onboard devices, such as throttle position, wheel speed, and emergency brake application. The WPU 202 receives and processes data from the event data recorder 38 once per second over an external serial connection.

[0027] The locomotive digital video recorder (DVR) 52, similar to a television DVR, is an onboard video recording device. The DVR 52 comes equipped with a forward facing camera and a microphone. The camera is mounted at such orientation that it sees and records what the engineer sees. The WPU 202 accesses the locomotive's DVR 52 via an external Ethernet connection to download the video from the hard drive before, during, and after an event.

[0028] The fuel level sensor 210 is a sensor that is used to measure the amount of fuel inside the fuel tank. The fuel level sensor 210 used in the present disclosure is an ultrasonic level sensor which uses ultrasonic acoustic waves to determine the distance between the sensor head and the fuel level. The sensor 210 is mounted on top of the fuel tank with known dimensions and mounting location. The WPU 202 accesses this data via an external serial connection.

[0029] The fuel level sensor software 212 takes the distance from the fuel level to the sensor 210 with fuel tank geometry and converts this data into a steady fuel volume. This is done by applying mathematical filtering to reduce noise from sloshing and ultrasonic behaviors of the tank. The software 226 also uses smart algorithms to determine refuel and fuel drop events.

[0030] The WPU 202 of the illustrated embodiment is a ruggedized onboard computer running Windows XP embedded specifically for industrial applications. It has many different features that can be installed to customize the product for specific customer needs. The WPU 202 has the ability to communicate with a wide variety of onboard systems, including, but not limited to, vehicle control systems, event data recorders, DVRs, fuel level sensors, and engine controllers. The WPU 202 has the ability to communicate over a wide variety of protocols, including, but not limited to, RS 232, RS 422, RS 485, CAN Bus, LAN, WiFi, cellular, and satellite.

[0031] The inertial navigation sensor board (Board) 214 is a hardware upgrade for the WPU 202. It is installed internally and communicates with the WPU 202 via an internal serial port. The board 214 consists of four components: a 3-axis gyroscope 216, a 3-axis magnetometer 100, a 3-axis accelerometer 20, and a microcontroller 222. The gyroscope 216 is used for measuring angular accelerations, the magnetometer 100 is used for measuring magnetic fields, the accelerometer 20 is used for measuring linear accelerations and decelerations, and the microcontroller 222 is used for processing data and communicating between the sensors and the WPU 202.

[0032] The firmware 224 runs on the Board's 214 microcontroller 222. The firmware 224 constantly calculates pitch and roll using the 3-axis acceleration 20 data. By comparing the 3-axis acceleration data to programmatically defined thresholds and durations, the firmware 224 can determine if a trigger event occurs and if so, sends a trigger event message to the WPU 202. Every second, the firmware 224 sends a periodic data message containing a predefined set of values to the WPU 202. This data is used for, but not limited to, determining heading, internal ambient temperature, and angular accelerations.

[0033] The system software 226 is an application running on the WPU 202. This application talks directly to the GPS 106 and Board 214 to gather related data. In addition to this data, the system software 226, like all other applications on the WPU 202, uses a standard inter-process communication protocol to gather data from other software applications. These other software applications are running on the WPU 202 and communicate to other devices (DVR 52, event data recorder 38, etc.) which are physically connected to the WPU 202. By using all the data gathered, the system software 226 can compare the data to predefined thresholds and durations to determine if specific events have occurred.

[0034] The system 200 consists of a WPU 202 with a Board 214, firmware 224, and system software 226 installed and an event data recorder 38, a DVR 52, and a fuel level sensor 210. The system software 226 runs on the WPU 202, constantly correcting fuel levels and checking for event messages from the Board 214 or event data recorder 38 to take action.

[0035] The mobile asset data recorder and transmitter system 200 (FIG. 6) of the present disclosure performs seven functions: automatic orientation, automatic compass calibration, emergency brake with impact detection, fuel compensation with pitch and roll, rough operating condition detection, engine running detection and inertial navigation (dead reckoning). Each of these seven functions factors in signals generated by the 3-axis accelerometer 20.

[0036] Auto orientation is used to correlate the axes of the WPU 202 to the axes of the locomotive so that the values measured by the sensors correspond to the locomotive's axes. This process is accomplished by the software 226 and firmware 224. Due to different electronic environments on locomotives, the compass needs to be calibrated on a per locomotive basis. The software uses the WPU's 202 GPS 106 (FIG. 5, 6) to determine the heading of the locomotive. It then takes measurements from the magnetometer 100 and stores them in the corresponding position of an array. The array consists of 360 positions, one for every degree of heading. Using these values, the WPU's 202 software 226 can correct for the locomotive's own magnetic fields and only detect the change due to the earth's magnetic field.

[0037] FIG. 1 depicts a flow diagram of a method application for emergency brake with impact detection. The WPU 202 (FIG. 6) software 226 (FIG. 6) sends initialization commands to the firmware 224 (FIG. 6) to establish acceleration durations in each axis (Adx, Ady, Adz) 14 to be used for triggering events. These durations are stored onboard in the device embodying system 200. The WPU 202 software 226 also sends initialization commands to the firmware 224 to establish acceleration thresholds in each axis (Atx, Aty, Atz) 16 to be used for triggering events. These thresholds are stored onboard in the device embodying system 200 (FIG. 6). The microcontroller 222 (FIG. 6) pulls the raw 3-axis acceleration (Ax, Ay, Az) 18 data from the accelerometer 20 at a rate of 100 Hz. A low pass filter 22 is applied to the raw acceleration values (Ax, Ay, Az) 18, which results in filtered acceleration values (Afx, Afy, Afz) 24. The Board 214 (FIG. 6) axes of the filtered acceleration values (Afx, Afy, Afz) 24 are translated to asset axes (Af′x, Af′y, Af′z) 26. The Board 214 values of the raw values (Ax, Ay, Az) 18 are translated to asset axes (A′x, A′y, A′z) 28. The filtered values of the asset axes (Af′x, Af′y, Af′z) 26 are added to the established thresholds for each axis (Atx, Aty, Atz) 16, and this added threshold (Af′tx, Af′ty, Af′tz) 32 is then continually compared 29 to the raw acceleration in the asset axes (A′x, A′y, A′z) 28. When the raw values (A′x, A′y, A′z) 28 exceed the thresholds 32 in one or more axes, a timer is activated 30. When a raw value 28 no longer exceeds the thresholds 32 in a specific axis 30, the duration that the raw value 28 exceeded the thresholds 32 is evaluated to determine if the duration exceeds the specified duration for that axis (Adx, Ady, Adz) 14. If the event duration was longer than 34 the duration established (Adx, Ady, Adz) 14, a trigger event is stored 36, including specifics on which axis, duration of the event, and time of the trigger event. In parallel with this monitoring, the onboard software 226 (FIG. 6) is receiving periodic data messages 40 from an onboard event data recorder 38, which is monitoring real-time status of various input sensors. The onboard software 226 monitors the periodic data messages 40 and detects when the periodic data message 40 indicates an emergency brake application discrete signal has occurred 42. The onboard software 226 stores the time 44 that the emergency brake application event occurred. If the onboard software 226 stores either the trigger event 36 or the emergency brake time 44, the onboard system software 226 will check the time stamp of each event to see if the latest two events logged, from the trigger event 36 or emergency brake application 44, are in close proximity 46. If it is detected that the events occurred in close proximity 46, the onboard software 226 will trigger an emergency brake application with impact alert 48 and will request a digital video recorder download 50 covering the time of the event from the onboard DVR 52 and will request the data log file covering the time of the event 125 from the event data recorder 38. The onboard software 226 receives the downloaded video covering the time of the event 54 and the data log file covering the time of the event 127 and sends both to the back office 56/128.

[0038] Users will receive alerts indicating the actual force of the collision and if the collision resulted in a rollover or derailment. This, coupled with GPS location, video and immediate access to event recorder information, allows users to precisely relay the severity and scope of the incident to first responders as they are en route to an incident.

[0039] FIG. 2 depicts a flow diagram of a method application for fuel compensation using accelerometer-based pitch and roll. The WPU 202 (FIG. 6) software 226 (FIG. 6) pulls the raw 3-axis acceleration data (Ax, Ay, Az) 18 from the accelerometer 20 at a rate of 100 Hz. A low pass filter 22 is applied to the raw data (Ax, Ay, Az) 18, which results in filtered acceleration values (Afx, Afy, Afz) 24. The Board 214 (FIG. 6) axes of the filtered values (Afx, Afy, Afz) 24 are translated to asset axes (Af′x, Af′y, Af′z) 26. The asset's pitch 58 is the arc tangent of the asset's filtered x-axis and the asset's filtered z-axis:

$\arctan \left(\frac{x-\mathrm{axis}\mathrm{translated}\mathrm{filtered}\mathrm{acceleration}\mathrm{value}}{z-\mathrm{axis}\mathrm{translated}\mathrm{filtered}\mathrm{acceleration}\mathrm{value}}\right).$

[0040] The asset's roll 60 is the arc tangent of the asset's filtered y-axis and the asset's filtered z-axis:

$\arctan \left(\frac{y-\mathrm{axis}\mathrm{translated}\mathrm{filtered}\mathrm{acceleration}\mathrm{value}}{z-\mathrm{axis}\mathrm{translated}\mathrm{filtered}\mathrm{acceleration}\mathrm{value}}\right).$

For each model of asset the system is installed upon, the specific location of the fuel sensor mounting is captured. Specifically, the distance the sensor is mounted forward of the center of the fuel tank 62 is recorded. In addition, the distance the fuel sensor is mounted left of the center of the fuel tank 64 is also recorded.

[0041] The distance forward of center 62 is combined with the tangent of the asset's pitch 58 to obtain a first fuel distance adjustment. The distance left of center 64 is combined with the tangent of the asset's roll 60 to obtain a second fuel distance adjustment. The first and second fuel distance adjustments are combined to provide a single fuel distance adjustment 66. The onboard distance level sensor records the distance from the top of the tank to the fuel level present in the onboard fuel tank. The raw distance to the fuel 70 from the fuel sensor 68 is combined with the distance adjustment 66 to create an adjusted distance 72. The adjusted distance 72 is combined with a previously defined fuel tank geometric tank profile 74, which maps a distance to fuel value to a fuel volume 76. This results in a final fuel volume 78, which is adjusted as the asset travels through various terrains in which the pitch 58 and roll 60 are changing, compensating for the movement of the liquid within the tank of an operating mobile asset.

[0042] FIG. 3 depicts a flow diagram of a method application for potential rough operating condition detection using an accelerometer. The WPU 202 (FIG. 6) software 226 (FIG. 6) sends initialization commands to the firmware 224 (FIG. 6) to establish acceleration durations in each axis (Adx, Ady, Adz) 14 to be used for triggering events. These durations are stored onboard, in the device. The software 226 also sends initialization commands to the firmware 224 to establish acceleration thresholds in each axis (Atx, Aty, Atz) 16 to be used for triggering events. These durations are stored onboard, in the device. The microcontroller 222 (FIG. 6) pulls the raw 3-axis acceleration data (Ax, Ay, Az) 18 from the accelerometer 20 at a rate of 100 Hz. A low pass filter 22 is applied to the raw acceleration values 18, which results in filtered acceleration values (Afx, Afy, Afz) 24. The Board 214 (FIG. 6) axes of the filtered values 24 are translated to asset axes (Af′x, Af′y, Af′z) 26 and the Board 214 axes of the raw values 18 are translated to asset axes (A′x, A′y, A′z) 28. The filtered values of the asset axes (Af′x, Af′y, Af′z) 26 are added to the established thresholds for each axes (Atx, Aty, Atz) 16, and then this added threshold (Af′tx, Af′ty, Af′tz) 32 is continually compared 29 to the raw acceleration in the asset axes (A′x, A′y, A′z) 28. When a raw value 28 exceeds the threshold 32 in one or more axes, a timer is activated 30. When a raw value 28 no longer exceeds the threshold 32 in specific axis, the duration that the raw value 28 exceeded the threshold 32 is evaluated to determine if it exceeds the specified duration for that axis (Adx, Ady, Adz) 14. If the event duration was longer than the duration established for that axis (Adx, Ady, Adz) 14, a trigger event is stored 36, including specifics on which axis, duration of the event, and time of the trigger event.

[0043] In parallel with this monitoring, the onboard software 226 (FIG. 6) is monitoring asset speed via periodic messages from the onboard event data logger 38 (FIG. 1) and/or from an onboard GPS device 106 (FIG. 5, 6). The onboard software 226 monitors the asset speed 80 and detects when it exceeds a specified value 82. If both the speed 80 exceeds a specified value 82 and a trigger event stored 36 occur at the same time 84, the onboard system software 226 will check which axis the event was triggered in. If the event was triggered in the z-axis 86, the system will log a potential track issue alert 88. If the event was triggered in the x-or y- axis, the system will log an operator mishandling alert 90. If either a potential track issue alert 88 or an operator mishandling alert 90 occurs, the onboard software 226 will request a digital video recorder download 50 covering the time of the event from the onboard DVR 52. The onboard software 226 receives the downloaded video 54 and sends it to the back office 56.

[0044] Users can now use the normal operation of their mobile assets to precisely locate and alert, in real-time, areas where their assets are encountering rough operating environment, such as bad track/switch, rough seas, and poor roads. The user will receive an alert, a still or video image and the crucial operational black-box data immediately upon identification of a rough operating environment. Repair teams can respond to the exact location of the bad road or track. Marine routes can be adjusted to avoid bar currents or choppy waters. The effectiveness of any repairs or rerouting can be validated when the next mobile asset data recorder and transmitter system equipped asset traverses any previously flagged area.

[0045] FIG. 4 depicts a flow diagram of a method application for engine running detection using an accelerometer. The WPU 202 (FIG. 6) software 226 (FIG. 6) sends initialization commands to the firmware 224 (FIG. 6) to establish activity/inactivity durations in each axis (A1dx, A1dy, A1dz) 84 to be used for triggering events. These durations are stored onboard, in the device. The WPU 202 (FIG. 6) software 226 (FIG. 6) also sends initialization commands to the firmware 224 (FIG. 6) to establish activity/inactivity thresholds in each axis (A1x, A1ty, A1tz) 86 to be used for triggering events. These durations are stored onboard, in the device. The microcontroller 222 (FIG. 6) pulls the raw 3-axis acceleration data (Ax, Ay, Az) 18 from the accelerometer 20 at a rate of 100 Hz. A low pass filter 22 is applied to the raw acceleration values (Ax, Ay, Az) 18, which results in filtered acceleration values (Afx, Afy, Afz) 24. The Board 214 (FIG. 6) axes of the filtered values 24 are translated to asset axes (Af′z, Af′y, Af′z) 26 and the Board 214 axes of the raw values 18 are translated to asset axes (A′x, A′y, A′z) 28. The filtered values of the asset axes (Af′x, Af′y, Af′z) 26 are added to the established activity/inactivity thresholds for each axis (A1tx, A1ty, A1tz) 86 and then this added threshold (Af′1tx, Af′1ty, Af′1tz) 88 is continually compared to the raw acceleration in the asset axes (A′x, A′y, A′z) 28. When the raw value 28 exceeds the threshold 88 in one or more axes, a timer is activated 90. If the raw value 28 no longer exceeds the activity/inactivity threshold 88 in a specific axis, the duration that the raw value 28 exceeded the threshold 88 is evaluated to determine if it exceeds the specified duration for that axis (A1dx, A1dy, A1dz) 84. If the event duration was longer than the duration established for that axis (A1dx, A1dy, A1dz) 84, a trigger inactivity/activity event 34 is stored 92, including specifics on which axis, duration of the event, and time of the event trigger. The engine running status is updated 94 when activity/inactivity events are triggered.

[0046] FIG. 5 depicts a flow diagram of a method application for inertial navigation (dead reckoning). The microcontroller 222 (FIG. 6) pulls the raw 3-axis acceleration data (Ax, Ay, Az) 18 from the accelerometer 20 at a rate of 100 Hz. A low pass filter 22 is applied to the raw acceleration values (Ax, Ay, Az) 18, which results in filtered acceleration values (Afx, Afy, Afz) 24. The Board 214 (FIG. 6) axes of the filtered values 24 are translated to asset axes (Af′x, Af′y, Af′z) 26. The asset's pitch 58 is the arc tangent of the asset's filtered x-axis and the asset's filtered z-axis:

$\arctan \left(\frac{x-\mathrm{axis}\mathrm{translated}\mathrm{filtered}\mathrm{acceleration}\mathrm{value}}{z-\mathrm{axis}\mathrm{translated}\mathrm{filtered}\mathrm{acceleration}\mathrm{value}}\right).$

[0047] The asset's roll 60 is the arc tangent of the asset's filtered y-axis and the asset's filtered z-axis:

$\arctan \left(\frac{y-\mathrm{axis}\mathrm{translated}\mathrm{filtered}\mathrm{acceleration}\mathrm{value}}{z-\mathrm{axis}\mathrm{translated}\mathrm{filtered}\mathrm{acceleration}\mathrm{value}}\right).$

[0048] Acceleration in the asset's x-axis is integrated 96 to calculate the asset's speed 98:

∫ asset's acceleration_{x-axis translated filtered acceleration value}.

[0049] In parallel, the microcontroller 222 (FIG. 6) pulls 3-axis gauss data (Gx, Gy, Gz) 102 from the magnetometer 100 at 1 Hz. Using the magnetometer data 102 and the asset's pitch 58 and roll 60, a tilt compensated heading 104 is calculated. Also in parallel, the onboard GPS device 106 is providing location data updated at a 1 Hz frequency. The onboard software 226 determines if valid GPS data is available 108. If a GPS signal is available, the onboard software 226 will parse the data 110, into GPS speed 126, heading 128, latitude 114, and longitude 116 every second, and will store 118 the latitude 114 and longitude 116. If the GPS data is determined to not be available, the system 200 (FIG. 6) enters dead reckoning mode 112. In dead reckoning mode 112, the last known latitude 114 and longitude 116 are obtained from the GPS 106 and stored 118. Using the last known 118 latitude 114 and last longitude 116, along with the asset's speed 98, the wheel speed from the event recorder data 126, the tilt compensated heading 104 and the data 129 from the 3-axis gyroscope, a new position 120 is calculated. The new latitude 122 and the new longitude 124 positions are stored and used, and the process continues until valid GPS data is again available.

[0050] Users will receive precision departure and arrival alerts and logging in environments where GPS signals are blocked or partially blocked by overhangs and canopies. This system 200 (FIG. 6) allows users to define virtual ‘trip wires,’ even in areas where GPS devices are rendered useless due to RF signal loss or interference. The inertial navigation capabilities automate operator performance to a schedule matrix by alerting and logging the exact time an asset crosses a departure and arrival virtual ‘trip wire’ when a GPS signal cannot compute accurate location data.

[0051] While the present disclosure has been described in connection with certain embodiments, it is to be understood that the present disclosure is not to be limited to the disclosed embodiments but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the scope of the appended claims, which scope is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures as is permitted under the law.