Menu Driven System for Controlling Automated Assembly of Palletized Elements

This invention relates generally to a method and apparatus for process control of robots within a flexible assembly system (FAS), and more particularly to defining the geometrical position of workpieces held on a pallet in a flexible assembly system.

The following related application is hereby incorporated by reference for its teachings:

"System for Handling Defects Produced During the Automated Assembly of Palletized Elements," Galuga et al, U.S. Ser. No. 08/139,664, filed concurrently herewith.

A portion of the disclosure of this patent document contains material which is subject to copyright protection. The copyright owners have no objection to the facsimile reproduction, by anyone, of the patent document or the patent disclosure, as it appears in the Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.

Short-run production is increasingly utilized in today's competitive manufacturing environments. In order to meet such demands, many dedicated assembly lines are being replaced with flexible systems which enable rapid changeover of the line, thereby increasing the efficiency of production. Unfortunately, the industrial workcells, which typically employ programmable robots, have become extremely complex and the difficulty in modifying the operations of such workcells has resulted in rather inflexible assembly lines.

Flexible assembly is the ability of an automated or robot system to assemble different parts into different assembled units. For example, at the beginning of the workday a robot may be configured to place and secure circuit boards into the chassis of personal computers. After completing a desired quantity of assemblies the robot may be instructed to assemble components for a paper path in a xerographic copier. Although the materials and tools are different, the work processes share similarities that can be modeled such that a unified architecture can be developed to satisfy all types of flexible robot assembly applications.

The present invention is directed toward alleviating one of three types of data flow and control problems which arise within such a flexible assembly system, including: 1) encoding the location of workpieces used by robots in the workcell; 2) collecting and processing error data during production; and 3) providing flexibility in the control of operations within the system.

The first problem is resolved using a method and apparatus for defining the position of workpieces held on pallets used within a flexible assembly system. The pallet configuration capability of the present invention captures the geometrical configuration of workpieces on any type of pallet (i.e., a universal pallet) and translates the configuration into a data record which the robot understands.

The second problem is alleviated using a method which encodes process error data transmitted between the components of the flexible assembly system. This encoded or "encapsulated" error data, is subsequently used to identify substandard workpieces to a particular position on the pallet. The use of the encapsulated error data further enables efficient pallet routing decisions to be executed within the FAS cell. Ultimately, the proposed system improves the quality of the end items or assemblies produced therein.

The last problem in the flexible assembly system is resolved using a method which provides the user the choice of running a workcell either locally (at one of the robot stations), or at a remote source (a remotely located computer). Control capability from a remote system, a cell controller, is incorporated into the architecture, because of the advantages remote control provides. Remote control of cell operations correctly implies that multiple workcells containing multiple robots can be effectively controlled from a single location on the production floor. This in turn implies that fewer operators would be required to oversee a greater number of production operations, thereby further improving productivity.

The use of robots for the assembly of workpieces arranged in predefined locations is known, and control of the robots by a microcontroller which executes commands entered via a proprietary software language is also known. For example, Adept robots (Adept Technology, Incorporated, San Jose, Calif.) utilize the V+language to create Pascal-like data records to define system characteristics. Robotic systems have commonly been employed in "dedicated" or highly repetitive operations that remain constant over long runs on an assembly line. As a result, the need for flexibility in the system, including the ease of modifying the robot operations or other system characteristics, was not a priority in such systems. Hence, alteration of the operation of the systems usually requires knowledge of the proprietary robot language and system characteristics in order to effect the desired changes or, in other words, a person having a high level of technical training is needed to effect alterations in the robot's process.

Heretofore, many such "dedicated" assembly systems have been developed to fabricate and inspect mechanical and electrical assemblies, of which the following disclosures may be relevant:

The foregoing patents and publication are hereby incorporated by reference for their teachings, and relevant portions thereof may be briefly summarized as follows:

U.S. Pat. No. 4,521,807 discloses an optical inspection system for the inspection of circular workpieces which includes an electronic camera having an optical system for focussing an image of a circular workpiece on an electronically active image receiving surface, and means to move the workpiece into a predetermined position in the field of view of the optical system. Difficulties encountered when inspecting a circular workpiece are overcome by scanning an image receiving surface in directions extending radially across the image receiving surface to produce an output signal. A signal analyzer is arranged to monitor the output signal corresponding to each radial scan to detect the presence of any irregularity in the workpiece.

U.S. Pat. No. 4,541,011 discloses a system for determining and recording the position, shape, size and depth of a defect in a billet or similar workpiece. The workpiece is displayed on the screen of a video display device, and the positions of the defects displayed on the screen are recorded in response to being manually selected by an operator. In addition, the patent further discloses that position indications from sensors are applied to a processor which correlates these indications with the position of the defect on the screen to determine the position of the defect on the workpiece.

U.S. Pat. No. 4,787,143 teaches a method and apparatus for detecting and enabling correction of a failure in the mounting of electronic parts on substrates. Moreover, automatic detection of failure in the mounting, and classification between failed substrates and acceptable substrates, improves the operating efficiency of an electronic parts mounting line. The method adapted applies a code mark to each substrate on which electronic parts are to be mounted, detects a mounting failure, reads and stores the code mark and failure data, and separates the failed substrates from acceptable substrates based upon the code mark and stored failure data.

U.S. Pat. No. 4,916,286 teaches a method and apparatus for controlling work performed by an automatic work processing machine whereby a just finished work undergoes an imaging operation by the use of an imaging apparatus. The outputs from the imaging apparatus are converted into binary values to obtain a line image which is then analyzed to evaluate the finished work by comparing the binary values to reference values. If the finished work is not satisfactory, evidenced by a deviation between the binary values and the reference values, the deviation is supplied to a control system which automatically changes some of the working parameters to attain a work of satisfactory quality.

U.S. Pat. No. 4,929,845 describes a method and apparatus for the inspection of circuit boards to detect missing and misaligned active and passive surface-mounted components on the circuit board. First the board is illuminated with a top light to enhance the image of the passive components. Next the circuit board is displaced, relative to a camera linescan trained on the board, so that the camera captures the image of successive strips of surface area across the board, perpendicular to its direction of movement. The images generated are subsequently stored and thereafter processed to determine if any components are misaligned or missing.

U.S. Pat. No. 4,973,852 discloses an apparatus which enables a user to identify an empty location of component parts on a printed circuit board or other type of component assembly, where the presence, absence, orientation or polarity of components is identified. Identification is described as being accomplished by utilizing a pattern of marks, such as dots, which are located in such a manner so as to be totally or partially obscured when the component is placed in the correct location. The marks may subsequently be used in a number of ways, including providing the identification of any location not occupied by the intended component, and indicating the polarity or orientation of a component.

A Xerox Disclosure Journal publication entitled "Automatic Correction of Robot Displacement Errors", published in the September/October 1992 issue (Vol. 17, No. 5), hereby incorporated by reference, teaches a method for automatically correcting displacement errors in robotic systems for part locations accessed by robot arms on assembly trays. The difference between the true location of a part and the location moved to by the robot relative to a given frame defines a displacement error.

In the previously described automated assembly systems, the flexibility of the system is characterized by its ability to prepare a robot with the parts, tools, and process steps necessary to accomplish the required tasks with little delay or downtime. Hence, the requirements of a flexible robot assembly system include:

(1) the ability to move parts and assembly tools into and out of the robot's workspace in an efficient and timely manner;

(2) the ability to inform the robot which application program to use when assembling the parts presented to it; and

(3) the ability to inform and recover from exceptional conditions, or errors, which occur during an assembly process.

To meet these requirements, the flexible assembly system which incorporates the present invention includes an architecture which implements robot control in a manner which maximizes flexibility. The system also hides the details of the architecture and programming languages from the end-user responsible for characterizing or altering the geometric configurations of parts presented to the system. Lastly, the system includes the capability to efficiently handle process errors detected during assembly so as to reduce or eliminate defects.

In accordance with the present invention, there is provided a system for automatically performing assembly operations on workpieces located on a pallet. The system comprises: a user interface for specifying the location of the workpieces on the pallet; a workcell for performing operations to modify a workpiece present on the pallet; and a cell controller, in communication with said workcell, for controlling the operation of said workcell in response to a set of preprogrammed requirements.

In accordance with another aspect of the present invention, there is provided, in a workcell which includes a robot controller and a robot responsive to the robot controller for assembling workpieces, a method of controlling a system for assembling the workpieces located on a pallet presented to the workcell. The method comprises the steps of: defining the location of workpieces on the pallet; and communicating, from a cell controller having a set of preprogrammed requirements to the workcell, a plurality of commands for controlling the operation of the workcell, so as cause the workcell to assemble the workpieces presented on the pallet.

FIG. 1 is a schematic illustration of the principle components of a flexible assembly system which employs the present invention;

FIG. 2 is a flowchart depicting an example of the operations which might be executed by the various "objects" on a flexible assembly system;

FIG. 3 illustrates the physical layout of a single robot assembly cell in the flexible assembly system which incorporates the present invention;

FIG. 4 and 5 illustrate some alternative pallet geometries which could be used in the workcell of FIG. 3;

FIG. 6 is a representation of the menu screen generated by the present invention and displayed on the user interface shown in FIG. 1;

FIGS. 7A, 7B, 7C and 7D comprise a flowchart which illustrates the various processing steps carried out by the robot controller of FIG. 1 while a user is exercising the pallet configuration functions of the present invention;

FIG. 8 is a detailed illustration of the various frame locations taught by an operator during the execution of the new pallet configuration operations as represented on the menu screen of FIG. 6 and as detailed by the process steps of FIG. 7; and

FIGS. 9A and 9B show a general flowchart representing the integration of a flexible assembly system incorporating the present invention in a manufacturing process.

The present invention will be described in connection with a preferred embodiment, however, it will be understood that there is no intent to limit the invention to the embodiment described. On the contrary, the intent is to cover all alternatives, modifications, and equivalents as may be included within the spirit and scope of the invention as defined by the appended claims.

For a general understanding of the operation of the present invention, reference is made to the drawings. In the drawings, like reference numerals have been used throughout to designate identical elements. As illustrated in FIGS. 1 and 3, the Flexible Assembly System (FAS) which embodies the present invention comprises the following systems:

(1) conveyor systems to route parts on pallets to and from robot workstations;

(2) high speed data networks for controlling robots from a central intelligence controller; and

(3) piece-part inspection systems to allow for feedback on part locations and defects. Generally, application programs such as the one depicted by the flowchart of FIG. 2 define the steps a robot must take in order to assemble parts which are presented to it. These application programs utilize the systems listed above to complete the required assembly operation. Although the actual assembly process can be radically different between different programs, the application programs all share similar problems such as obtaining parts, releasing assembled units, communication to a centralized controller, error recovery, etc. The following description of a FAS embodiment includes an architecture that solves common control and data flow problems of such flexible assembly systems.

The flexible assembly system can be characterized as a set of objects. Generally, an object is an abstraction of a set of real world "things" that have the same characteristics. As an example, a robot is a real world thing. Robots have common characteristics regardless of their make or model. For instance, robot manipulators have speed constraints, maintain a run-time state, maintain a current location, have serial numbers, model and manufacturer names, etc. The specific values of these characteristics is what identifies a unique robot. Thinking of a robot as a simple object with characteristics, or attributes, makes it possible to define a system architecture that works for all such objects.

The FAS architecture defines several objects: a robot, a pallet, a pallet router, an assembly part, and a part inspector. An Adept robot and Texas Instrument robot are instances of the robot object. A xerographic paper path, an integrated circuit chip, and a car door are instances of assembly parts. Not only do object instances have similar types of attributes, they also have similar operations (methods) that can be performed on them. For example, all robots need to respond to a HALT command from the operator, and pallet objects should be able to maintain a state variable and provide state information to whomever or whatever requests such information. Using the familiar concept of data encapsulation, the only way to modify an object's attributes is through the object's methods. For example, to modify the speed attribute of a robot object the robot. SetSpeed() method must be invoked, and to determine the next available part on a pallet object the pallet. GetNextOperatingPosition() must to be invoked. The flexible assembly architecture provides application process flow and system object definitions (classes in object oriented terminology) that may be used by any assembly system that can be modeled by the defined objects. The process flow for a generalized assembly operation is illustrated in FIG. 2. Definitions of some of the objects included in the present embodiment are as follows:

ROBOT--a software driven manipulator arm that may be fitted with various tools on its "hand" to pick up, place down, and perform other types of manipulations on ASSEMBLY PARTs (see definition below).

Attributes:

Speed, CurrentOperatingPalletLocation, CurrentOperatinaPalletArea, NumberOfUnitsToBuild, NumberOfUnitsCompleted, NumberOfFailedUnits, CurrentState.

Methods (operations that the ROBOT objects must perform):

Initialize, ExecuteProgram, Halt, SetSpeed, Pause, Resume, StopAfterCurrentAssembly, GetState, GetCurrentLocation, SetNextPalletPosition.

PALLET--a container that holds ASSEMBLY PARTS, arranged in such a way as can be described to the ROBOT for manipulation.

Attributes:

NumberOfAreas, NumberOfTiers, FrameErrorCorrectionFactor, ArrayOfAreas, CurrentOperatingArea, (each area has the following attributes:) NumberOfParts, NameOfPart, NumberOfRows, NumberOfColumns, DistanceBetweenRows, DistanceBetweenColumns, TierOffset, CurrentState, CurrentOperatingRow, CurrentOperatingColumn, CurrentOperatingPartsAcquired, ReferenceLocation.

Methods:

ResetCurrentOperatingRow, ResetCurrentOperatingColumn, GetAreaAttribute, Get/Set:CurrentOperatingArea, GetAllOffSets, GetPartName, Initialize, GetCurrentOperatingPosition, GetNextOperatingPosition, GetNextPartLocation, GetRemainingPartQuantity, GetRemainingRows.

ROUTER--A module (comprised of software and/or hardware) which is responsible for the scheduling of ROBOT work and movement of pallets into and out of a ROBOT's local workspace.

Attributes:

NumberOfRobotsinSystem, NumberOfPalletPositionslnRobotWorkSpace, CurrentlyActiveRobots, ArrayOfPalletLocations.

Methods:

PurgePallet, DeliverPallet, ReRoutePallet, ScheduleWork, HandleError, CheckPalletAvailability.

ASSEMBLY PART--one or more pieces of material that can be combined with other pieces of material using a ROBOT in order to construct a single assembled unit.

Attributes:

PartName, PartManufacturer, PartSerialNumber, CurrentState (good, defective).

Methods:

SetPartState.

PART INSPECTOR--an external device that is used to inspect the presence or possibly the appearance of assembly parts on pallets.

Attributes:

CurrentState (on, off).

Methods:

InspectPart, ChangeState.

FIG. 1, an illustration of the principle components of the flexible assembly system, shows the communications link between cell controller 20 and robot controller 22. In the present embodiment, the cell controller and the robot controller function in concert to perform the operations of the ROUTER as described above. Cell controller 20 is preferably a microcomputer having at least one microprocessor, a storage memory, preferably a random access memory and/or permanent storage medium such as a hard disk. Robot controller 22 is preferably an Adept Robot Controller capable of hosting at least a second generation programming language, and allowing for integration with other equipment. Preferably the robot controller should also include memory and a high speed processing capability, so as to prevent a slowdown of the system while waiting for responses from the robot controller.

Cell controller 20 communicates to robot controller 22 via a BitBus Network, 24, (Intel.RTM. Corporation) or any similar network capable of sustaining a communication rate of about 375 KBaud. However, because individual robot systems are unable to communicate at such a high rate, particularly the Adept robot systems in the present embodiment, gateway 26 is used to receive the commands intended for the robot controller and convert them to a lower communication rate of about 19.2 KBaud. More specifically, gateway 26 is a communications device that receives the messages on the network 24 and converts them to be transferred, via RS-232 line 30, to robot controller 22.

Within robot controller 22 there are resident software programs which carry out the operations necessary to control and monitor a robot connected thereto, see reference numeral 40 in FIG. 1. As illustrated in FIG. 1, the robot controller includes both shared memory 32 and communications interface 34 through which both an application process and message polling process transfer command and status information. For example, the polling process polls the RS-232 serial channel for messages from the cell controller. The received messages, or commands, are passed to the application process and executed. Once completed, the application process responds and the message is again passed to the polling operation which in turn transmits it, via the gateway, back to the cell controller.

Communications Interface 34 provides reliable inter-process communication of celt controller commands and data for the robot controller 22. Communication amongst various application processes running on the robot controller is accomplished via shared memory 32 which contains global control variables. During operation of the robot controller, the polling process either reads from or writes to the global control variables, depending on the type of message or command received from the cell controller. Moreover, to prevent con flicts during access of the global variables contained in the shared memory, communications interface 34 arbitrates the access to the shared memory, thereby "hiding" the global variables from direct access by the processes running on the robot controller. Shared memory 32 is preferably a random access memory (RAM), but could also be any permanent storage device, such as a hard disk, which allows the random access of data stored thereon.

Turning now to FIG. 2, the flowchart depicts the normal operation of the FAS, where the methods defined for each object are utilized within a typical application program. For instance, router methods and robot methods running within the applications process 38 of FIG. 1 use a communication interface 34 to communicate with a remote cell controller 20. Furthermore, a pallet object is initialized with data established by a pallet configuration utility, and router methods utilize process error data communicated by the robots.

Beginning with step 90 in FIG. 2, an application program would instruct the robot object to execute the SetNextPalletPosition method. The SetNextPalletPosition operation would then identify the position 50 of the pallet holding parts to be assembled. Next, the router object is instructed to execute the method step CheckPalletAvailability, step 92, to determine if a pallet is located at the position returned by step 90. Subsequently, in steps 94 and 96, a number of calls are made to the pallet objects to sequentially determine the location of the next assembly part on the pallet (GetNextPartLocation), determine the number of parts that remain to be assembled, and the number of rows remaining on the pallet in the position returned by step 90. Once the preliminary steps are completed, an optional inspection step 98 may be performed to assure that one or more of the assembly components are satisfactory.

Having satisfactorily determined that the materials are available to complete an assembly, the User Defined Module, represented by step 100, is executed. This is the user programmable portion of the application program, where robot and pallet methods may be accessed to manipulate or produce the current assembly. Generally, the operations will either perform some operation on the assembly part or will move the part to be assembled with another part. All the operations to be executed in this step are controlled by applications level "calls" to the various objects, specifying methods or operations which the objects are to perform.

Once the User Defined Module is completed, test step 102 determines if either the remaining part quantity or a row variable have reached zero. The first indicating that no further parts are available to be processed, while the latter indicates that a pallet has been completed. A negative response to both will result in a restarting of the application process flow at step 90. An affirmative response will move the application program to test step 104, where a determination is made whether further parts are to be processed. If a sufficient number of assemblies have been produced, the step will produce an affirmative response and the application flow will stop, step 108. Otherwise, the application continues at step 106, where the router object is instructed to carry out the operations PurgeLastAccessedPallet and DeliverPallet. These two operations will serve to remove the completed pallet and to replace it with a new pallet. Although the application program flow diagram illustrates a sequential process, the reliance upon individual objects to accomplish the identified tasks increases the flexibility of the User Defined Module significantly over commonly available systems.

Turning now to FIG. 3, which illustrates the layout of a robot assembly cell in the FAS, conveyor 46 delivers parts or workpieces (not shown) on trays referred to as pallets, 48, which are in turn placed in one of the pallet locations indicated by reference numerals 50. Prior to using a tray, the robot must know the geometry of the tray. In other words, the robot must be instructed as to the configuration of the tray and the workpieces contained thereon. For example, as represented by the two alternative tray geometries in FIGS. 4 and 5, trays 48 must have critical parameters such as row spacing .DELTA.x, column spacing .DELTA.y, tier offset o.sub.t, row offset O.sub.r and other geometrical characteristics determined before the robot can locate the workpieces, 54, contained thereon. Previously, pallet geometries were specified to the robot by manually creating a data structure or file with an editor available on the Adept robot controller. However, the present invention incorporates a menu-based user-interface program which may be utilized to prompt the operator for the basic tray geometry and to calculate the required positional information using the information input by the operator.

The pallet configuration menu utility, assists operators wishing to define a set of pallet geometries and, as shown on a video display 60 in FIG. 6, has a series of options to produce user-interface (U/I) screen 62. The general operation of the pallet configuration program is represented by the flowchart illustrated in FIG. 7, in conjunction with the software listing contained in Appendix A. Initially, the software displays U/I screen 62, block 110, which prompts the user for a selection of the desired operation. Upon a selection by the operator, block 112, the designated operation will be initiated by one of the seven test blocks 114a-g, which test the operator's selection.

Upon selecting the first option, "Configure New Pallets", the user is prompted to fill in various characteristics of the pallet. Specifically, the operator is required to fill in information including, but not limited to, the following: pallet location; number of areas on the pallet; least number of parts in an area; number of rows in each area; number of columns in each area; number of tiers in each area; tier offset; part name in each area; total number of parts in each area; .DELTA.x (row offset); .DELTA.y (column offset); area frame error; and status of robot pallet position in the assembly. Once entered, the characteristics are used to calculate the nominal locations of workpieces. After entering the appropriate information at block 116, the operator is then queried as to whether it is necessary to teach a frame, block 118.

Teaching a frame, block 120, involves placing the robot in a condition which allows a manual control pendant to control the position of the robot. Using the control pendant, the operator is instructed to first move the robot to contact the registration corner location of each pallet, illustrated in FIG. 3 by reference numeral 56. Subsequently, the operator is instructed to move the robot to the farthest point along the x-axis, or in other words, the corner of the pallet in the x-direction, and similarly the farthest point along the y-axis. Once the positions are manually taught, they are returned from the robot and stored in the data structure which defines the pallet. Next the pallet frame error is calculated at block 122, as described by the inventors in the Xerox Disclosure Journal publication entitled "Automatic Correction of Robot Displacement Errors", published in September/October 1992, Vol. 17, No. 5, and incorporated herein by reference.

As illustrated in FIG. 8, shaded locations A, B, and C represent the three frame locations taught by the operator in the previous step. The horizontal line, reference numeral 80, represents the desired path of the robot. However, due to the inherent inaccuracy within the robot cell arising, for example, from both horizontal and vertical elastic deflections in structural members of the robot, or from teaching of the frame in a "free" mode when the gears and linkages are relaxed, positioning error about the frame is inherent. Thus, by teaching the robot frame locations A, B and C, the actual path which the robot will follow falls along a line that is often angled or skewed with respect to the desired path, for example, line 82 in FIG. 8. As an example, a command which directs the robot to move five columns in the y-direction could result in the robot being positioned at B, rather than at the exact center of the location of the sixth part location, represented by reference numeral 84.

At step 122, the operator has the option of entering the maximum frame positioning error manually, or allowing the robot to automatically calculate it. To calculate the frame positioning error automatically, the program moves the manipulator about the pallet simulating operation during production. When the robot gets to the furthest points on the pallet it stops and records its position. If the robot is not exactly on the taught reference point, where it is supposed to be, it subtracts this difference and labels it epsilon. Then, the next time the robot runs, it automatically corrects its path by a computed incremental epsilon to get closer to the original target positions of parts along its path. Whatever the method employed, the result is that the frame positioning error is added to the pallet's area structure. As yet another alternative, the operator can avoid teaching all three points, and manually enter coordinates for three points, two points, or just one point. When the teaching steps are completed, control is returned back to the upper level at block 110 where the main menu will again be displayed.

Upon selection of the "Reconfigure Pallets" option at test block 114b, the applications process running in the robot controller first builds a secondary menu screen at block 130 that displays on U/I screen 62 four possible functions; modify pallet structure, delete pallet structure, delete area structure, and reteach frame. If the modify pallet option is selected, as determined by test block 132a, the pallet configuration characteristics, similar to those described above with respect to defining a new pallet, are displayed and the operator may change some or all of the parameter values. As represented by block 134, the values are then saved in memory. The values may be permanently stored in the data structure which defines the altered pallet upon selection of the sixth option of the main menu, displayed on U/I screen 62, "Store Configuration File" and represented by test block 114f. The remaining three selections are deletion operations that, as reflected by blocks 136, 137 and 138, can respectively delete either an entire pallet structure contained in memory or simply an area that has been defined on a pallet, or allow the reteaching of frame reference locations for a specified pallet area.

Turning to the next possible selection from the main menu, the "View Configuration" option, test block 114c, displays the details of the pallet structure contained in the memory of the robot controller. More specifically, this is a listing of the parameters previously entered by the operator and/or calculated by the robot controller prior to the last time they were saved in the memory 32. This option may be used to confirm or check the configuration of the pallet.

The next functional option displayed on the user-interface screen is the "Help" option which, as selected via test block 114d and as reflected by blocks 150, 152 and test block 154, is a looping process. The Help option displays a secondary menu of help options and, upon selection by the operator, the desired help information. Upon selection of the "Draw Pallet" option, test block 114e, from the main menu, the software actually draws a picture of whichever pallet was selected onto U/I screen 62. In many instances, the drawing may be used by the programmer to verify that the right data has been input. Ideally, the representation of the pallet on the screen would look like the pallet that was being modeled. If none of the functional choices are selected by the operator, the "Quit" option may be selected via test block 114e, to cause the application process to exit from the pallet configuration software. Furthermore, any invalid selections will be ignored by the software process.

Having described the basic operations of the individual workcells, and the interaction between the robot and pallet, as controlled by the robot controller, attention may be turned to the manufacturing process associated with the flexible assembly system, and more specifically to the use of error data generated by one or more robot controllers during the assembly operations. In the FAS, a number of workcells, similar to that depicted in FIG. 3, may be connected in series or in parallel to perform assembly operations on workpieces provided to the workcells on pallets. Having previously defined the locations of parts and/or workpieces on each pallet type used by the workcell, a pallet identifier and workpiece location then serves as an identifier for each of the workpieces contained thereon. Accordingly, any errors detected during assembly or inspection of the assembled workpieces can be referenced, using the previously described pallet configuration data structure, at any time based upon the pallet location. This information is communicated from the application program, which is controlling the assembly of the workpieces on the robot controller, to the cell controller and then to other manufacturing databases so as to accurately track and control the disposition of defective workpieces.

In one embodiment, at least three alternative systems may be employed to achieve part and defect detection. The first two systems implemented were physically attached to the robot manipulator, while the third system was a manual inspection or rework station. More specifically, the first system was a commonly known vision system using a camera to verify part presence or absence. The second implementation was a force sensor that indicated a problem with the assembly process when the force, or load exerted by the robot did not fall within defined limits. For instance, either too much force or too little force was applied to accomplish the operation. The third defect detection method was a manual inspection/rework station where an operator provided the defect feedback as a result of visual or other analysis of the assembly. Upon detection of an error or defect, the error information that is returned to the cell controller includes the pallet location, an error identifier, an area identifier, row, and column. The workcell, in communication with a conveyor barcode reader (not shown), adds one more identification element, a pallet ID. Thus, the "system" knows where the error occurred, what the error was, which pallet contains the workpiece with the error, and exactly which component is affected.

The error information, referred to as process error data, may be generated by a number of occurrences, including, but not limited to: missing parts, misaligned pallets, excess force applied by the robot during assembly, or a missing tool for the robot. Similarly, additional process error data could be generated through the use of external devices which are commonly available, such as vision systems and electrical signature analysis systems, each providing negative feedback when a fault is detected in the process or upon inspection of an assembled workpiece.

The use of process error data will be described with respect to the assembly process steps illustrated in FIGS. 9A and 9B. The steps executed within region 208 represent those steps carried out locally by the robot under the control of a robot controller, which in turn is monitored and controlled by a cell controller. As represented by step 200, there may be one or more prior assembly steps conducted on some of the workpieces supplied to workcell 48 of FIG. 3. Subsequently, as step 202 and 204 illustrate, assembly operations are carried out by the local robot and tested to determine whether they have been completed successfully or whether an error or defect was detected. As described in detail below, the defect detection means may be any tool, device or process suitable for generating an indication of an error in the assembly process, preferably a numerically encoded indication. Once the workpieces are assembled, the pallet is then passed to a subsequent stage of the process, represented by step 206, where further processing may occur, or where the workpiece may be installed in the final product.

As shown in detail in FIGS. 9A and 9B, when an error is detected by error checking step 204, which is likely to be an integral part of assembly operation 202, the first operation conducted by the application process executing on the robot controller is to interrogate the process error data, step 220. This step essentially determines where the error has been reported from, for example the robot or an inspection device object, and what type of error was detected. The present embodiment represents the error numerically. If the error is attributable to the robot, the assigned error number will correspond with the robot manufacturer's hardware error numbering scheme. If the error is not attributable to the robot, for instance when a part is either missing or defective, or when the end of arm tooling is missing, a predefined, application specific or library type number is assigned. For example, application specific numbers would be used to indicate a force sensor detected error indicating possible part damage due to excessive force being applied, while a library type error (common to all applications) would indicate that a tool was not attached, a part was missing, etc. Having determined the specifics of the error, the process error data is then reported to the cell controller via the communication channel previously described with respect to FIG. 1. As indicated by the steps depicted in region 210 of FIG. 9, the cell controller then acts to respond to the process error data received from the robot controller.

More specifically, the robot controller first tests to determine if the error reported is correctable, step 222, and if so takes further action to determine if the error is one which may be automatically corrected, as determined by step 224 The robot determines the following:

1) whether the error is correctable (e.g., whether to try to pick up a part a second time);

2) whether the error is something which it has no control over (e.g., a missing part, which it would simply know to skip over and continue to the next part); or

3) whether is needs operator assistance.

When the robot needs help, it requests it by sending the appropriate error number directly to the cell controller. This decision process may be accomplished using the "limited intelligence" within the robot. In addition the cell controller might also be programmed to send an operator to investigate when it receives a predefined number of correctable errors from a particular robot, even though the robot has not signaled for help. In such an embodiment, the cell controller would be tracking the errors. The robot (slave) would simply "push" the information over to the cell controller (master), thereby accomplishing the robot's primary responsibilities of simply building and communicating. Assuming that an error cannot be corrected automatically, step 225 would determine whether the error was correctable at the present station. If not, the workpiece would be tagged as defective, step 232, and corrected at a subsequent station. Otherwise, when the error is correctable at the present station, step 226 would signal the operator at the workcell to intervene and/or manually adjust the robot, workpieces or pallets to enable subsequent assembly to continue. If automatic correction were possible, step 228 would depict the application program, running on the robot controller, generating the commands which direct the robot to carry out the adjustment and continue assembling the workpieces. In either situation, the workpieces for which an error was detected would be marked or "tagged" by the robot controller, and then the information subsequently transmitted to the cell controller. The cell controller, in turn records the workpiece location and pallet identification in a storage memory in order to assure that a subsequent inspection of the workpiece closely scrutinizes whether it was properly assembled.

If the process error data indicated that the error was not correctable, by a negative response in test step 222, the robot controller would then determine, via step 234, whether the error was fatal. Generally, a fatal error is one from which the robot controller is unable to recover without intervention, such as a robot crash. In this case, the assembly operations of the robot are stopped, step 236, and the robot controller returns a command to the cell controller to signal the workcell operator to intervene arid rectify the reported problem. After the operator has resolved the problem, he/she inputs directly to the robot controller whether a defective workpiece has resulted from the error, step 230. This information is then passed to the cell controller and if the answer is yes, the cell controller tags the workpiece by storing its location and pallet identification information into memory, step 232. All defects which may result in a substandard assembly, whether they are created due to a malfunction in the process, or due to substandard pieces in supply, are tagged and processed by the cell controller in the same way. The "prognosis" of all workpieces and assemblies within the system is known at the cell controller at all times. Once the defective workpiece is tagged, the cell controller determines, at step 250, whether the pallet has been completed. If not, processing continues at step 202 where the assembly operation is accomplished for the next workpiece on the pallet. If assembly of all the workpieces on the pallet was completed, further processing will be accomplished by the cell controller as described hereafter.

Subsequent to handling the process error data, including tagging defective workpieces if necessary, the cell controller may make further decisions based upon the process error which was reported. For example, the cell controller would make decisions with respect to the future use of the workpieces contained on the pallet, for example step 240, either accepting the pallet, preferably when defect free, or rejecting the pallet containing defective assembled workpieces. If rejected, the cell controller may further determine at step 242 whether the rejected pallet may be sent to a rework station, step 244, where the defect may be manually corrected, or to scrapping operation 246 where the entire pallet may be scrapped.

In the situation where the pallet is rejected and sent for rework, represented by step 244, the tagged workpieces, those having been identified as exhibiting a defect, may be graphically identified for the rework station operator based upon the information stored in memory by the cell controller. The cell controller can send a pallet to a manual rework station, bundled with the pallet's process error data record, so that a local terminal at the rework station may be used to display the defect assemblies. Specifically, having previously identified the characteristics of the pallet, as well as the location of the defective workpieces, the rework station can generate a display screen which represents the pallet, including workpiece locations, the exact location of the defective workpieces thereon, and detailed information as to the nature of each defect. An operator at the rework station can then either remove, replace, or rework the defective workpiece, depending upon the defect identified, thereby enabling the pallet to proceed for subsequent processing.

The features enabled by the pallet characterization and process error data enable the cell controller to cooperate with the application programs running on the robot controller to not only increase the flexibility of the assembly system described, but to significantly reduce the amount of scrap generated. Due to the almost instantaneous feedback of process error data to the cell controller, there is an imperceptible time lag during which defective workpieces are being assembled. Furthermore, because of the ability to quickly detect workpiece assembly errors, the need for a final inspection process can be eliminated as well.

In recapitulation, the present invention is a method and apparatus for easily describing the critical characteristics of the pallet or tray used to supply workpieces to an automated assembly station or workcell.

It is, therefore, apparent that there has been provided, in accordance with the present invention, a method and apparatus for improving the flexibility of an automated assembly system, while increasing efficiency and reducing the scrapping of defective parts. While this invention has been described in conjunction with preferred embodiments thereof, it is evident that many alternatives, modifications, and variations will be apparent to those skilled in the art. Accordingly, it is intended to embrace all such alternatives, modifications and variations that fall within the spirit and broad scope of the appended claims. ##SPC1##