The present invention relates to robots and more particularly to a method and apparatus for automated robot motion programming.
Programmable robots are commonly used for a variety of repetitive industrial applications. As appreciated, a robot only performs tasks and motions that are preprogrammed. Programming of robot motions can be a complicated and time-consuming process. Methods of reducing programming time include programming robot motions by lead through teaching in real setup and offline programming in a computer simulated setup. Lead through teaching is disclosed generally in U.S. Pat. No. 4,408,286 (“the '286 patent”).
The typical lead through teaching methods include the steps of moving the robot through a set of desired motions, sensing specific points during the movement of the robot, recording the specific points in a microprocessor, and utilizing the recorded points to create movement commands. The robot repeats the desired motions according to the created movement commands. An operator programming a robot utilizing a lead through teaching method is responsible for guiding the robot and for maintaining the desired position and orientation of the robot in three or more dimensions. A drawback to conventional lead through teaching methods is that an operator must constantly guide the robot through motions burdened with the requirement to accurately guide the robot through the desired motion while never allowing a collision with an object in the workspace and never allowing the robot to apply excessive pressure to the workpiece.
As shown in
Improvements in the prior art for the lead through teaching method include the use of various types of teaching handles (not shown) that aid the operator in guiding the robot through the desired motions. Teaching handles improve the operator's mechanical advantage over the robot, however the operator is still left to manipulate and accurately control the robot through the desired motions while maintaining control in three or more dimensions. One example of such a handle is shown and described in U.S. patent application Ser. No. 11/051,383 filed on Feb. 4, 2005 and assigned to the same assignee as the present invention. Another example of such a handle is shown and described in U.S. Pat. No. 6,385,508. A handle is described but not shown in U.S. Pat. No. 6,285,920 which also describes joy-stick jogging of the robot.
The offline programming method requires that a computer model is available for the workpiece and an accurate calibration equipment and procedure is available to convert the program data in the simulated setup into the program data in the robot setup. As a result, this method cannot be utilized if the computer model is not available or the calibration accuracy does not suffice to meet the process requirement.
Consequently, there is a need for a method and apparatus that allows the operator to program robot motions with minimum effort in the real setup so that the process requirement and motion path are presented together and the operator can provide in-situ and timely judgment for any particulars in programming the robot. Further, it is desirable that a spatial relationship between the robot and the workpiece is constantly maintained without any burden on the operator. Further, it is also desirable that the robot and the workpiece never collide even for applications where close contact is necessary. Further, it is desirable to free the operator from having to guide the robot in three dimensions for all the program points in a complex path during programming operations in an effort to reduce the complexity of programming desired motions of the robot. Further, simplification of the teaching operation allows for more efficient and automated teaching of robot motions, making frequent programming of programmable robots for small batch jobs economically feasible.
As is described in more detail below, the present invention allows the operator to first teach the robot the gross spatial relationship between the robot fixture and the workpiece, such relationship referred to hereinafter as guiding points, and then allow the robot under control of the robot controller to follow the path that the controller has determined from the guiding points and use force control to update the guiding points and also to learn one or more points intermediate adjacent guiding points to thereby generate a program for the path the robot will follow when the robot is used to perform the desired operation on the workpiece. That desired operation may for example, and without limitation, be deburring, polishing or milling.
The off-line teaching of guiding points is described by Frederick Proctor Proceedings Of The International Manufacturing Technology Forum, Chicago, Ill., Jun. 4-5, 1992 “Control Sensors For Advanced Manufacturing”. Proctor, however, does not teach using force control to learn one or more points intermediate adjacent guiding points. Published PCT application WO 94/20262 teaches using the guiding points and force control during the machining of a workpiece but as is the case with Proctor does not teach using force control to learn one or more points intermediate adjacent guiding points before the robot is used to perform the desired operation on the workpiece.
A method of programming a robot, said robot having a force sensor and a real or tangible tool fixture for cooperating with a workpiece having a known contour and a controller, said method comprising:
providing a plurality of gross spatial relationships between said tool fixture and said workpiece contour;
defining a first path for movement of said tool fixture relative to said workpiece contour, said first path including said gross spatial relationships;
moving said tool fixture using said first path as a guide;
determining whether said moving tool fixture is in contact with said workpiece contour using said force sensor;
when said moving tool fixture is not in contact with said workpiece contour, using said controller to change the direction of said moving tool fixture so as to bring said tool fixture in contact with said workpiece contour;
gathering spatial relationships between said tool fixture and said workpiece contour at points on the workpiece contour, intermediate the gross spatial relationships, where said tool fixture contacts said workpiece contour; and
using said gathered spatial relationships to define a second path for movement of said tool fixture relative to said workpiece.
A computer program product for programming a robot, said robot having a force sensor and a real or tangible tool fixture for cooperating with a workpiece having a known contour and a controller, said computer program product comprising:
computer usable program code configured to record a plurality of gross spatial relationships between said tool fixture and said workpiece contour;
computer usable program code configured to define from said plurality of gross spatial relationships between said tool fixture and said workpiece contour a first path for movement of said tool fixture relative to said workpiece contour, said first path including said gross spatial relationships;
computer usable program code configured to move said tool fixture using said first path as a guide;
computer usable program code configured to determine whether said moving tool fixture is in contact with said workpiece contour using said force sensor;
computer usable program code configured to use said controller to change, when said moving tool fixture is not in contact with said workpiece contour, the direction of said moving tool fixture so as to bring said tool fixture in contact with said workpiece contour;
computer usable program code configured to gather spatial relationships between said tool fixture and said workpiece contour at points on the workpiece contour, intermediate the gross spatial relationships, where said tool fixture contacts said workpiece contour; and
computer usable program code configured to use said recorded spatial relationships to define a second path for movement of said tool fixture relative to said workpiece.
A system for programming a robot, said robot having a force sensor and a real or tangible tool fixture and a controller for cooperating with a workpiece having a known contour, said system comprising:
a computing device having therein program code usable by said computing device, said program code configured to:
gather a plurality of gross spatial relationships between said tool fixture and said workpiece contour;
define from said plurality of gross spatial relationships between said tool fixture and said workpiece contour a first path for movement of said tool fixture relative to said workpiece contour, said first path including said gross spatial relationships;
move said tool fixture using said first path as a guide;
determine whether said moving tool fixture is in contact with said workpiece contour using said force sensor;
change using said controller, when said moving tool fixture is not in contact with said workpiece contour, the direction of said moving tool fixture so as to bring said tool fixture in contact with said workpiece contour;
gather spatial relationships between said tool fixture and said workpiece contour at points on the workpiece contour, intermediate the gross spatial relationships, where said tool fixture contacts said workpiece contour; and
use said recorded spatial relationships to define a second path for movement of said tool fixture relative to said workpiece.
Referring to the Figures, wherein like numerals indicate like or corresponding parts throughout the several views, a method of and apparatus for programming a robot is disclosed. As is shown in
The software that performs the method of the present invention is resident in the controller 11 and can be loaded into the controller 11 from a computer readable medium, including but not limited to a CD-ROM or a flash drive, or by other well known means such as, for example, by connecting controller 11 to an intranet or the Internet, and downloading the software from the same site where controller 11 is located or from another site that is remote from the site where controller 11 is located. As is well known to those of ordinary skill in the art, the software that performs the method of the present invention may also be resident on a computing device (not shown) other than controller 11 that interfaces with robot 10. The software can be loaded into that computing device by any one of a number of well known techniques, including but not limited to, those described above.
The automated path learning of the present invention may be used in connection with many different processes. One example of such a process is a deburring application, shown in
As shown in
The guiding points 151-154 can be taught by any of a number of methods, including the conventional remote joy-stick jogging or keyed jogging. A preferred method is lead through teaching. Lead-through robot programming is disclosed generally in the '286 patent. This programming method is different from the teach pendant method described above in which the robot is moved through a programming path of motion by use of a robot teach pedant. Rather lead-through teaching uses a force sensor 20 of
Next, a robot program based is generated in the robot controller 11 based on the linear segment of the guiding points 151-154, resulting in a program path 150. As is shown in
As is described in more detail below, the method of the present invention prevents this from occurring as it uses the compliant motion mode, that is, force control, to allow the tool fixture 12 under the control of controller 11 to use the program path 150 to orient the tool fixture 12 relative to the workpiece contour 141 so that the tool fixture follows contour 141 to teach one or more points on the contour that are intermediate adjacent guiding points 151-154. In addition the force control keeps the tool fixture 12 from striking the workpiece 14 as it would in the absence of force control if it tried to follow the portion of path 150 between guiding points 151 and 152 or brings the tool fixture 12 adjacent to the workpiece 14 as it follows the path 150 between guiding points 152 and 153.
Referring to
Further, it can be specified in the controller 11 that a constant contact force in the Yp direction (e.g., 20 N) is maintained. Because of this constraint, if the program path 150 is into the actual workpiece contour as shown for point 155, the tool tip will yield along the Yp axis until it reaches the equilibrium, of 20N, resulting in a new point 255 which is physically on the workpiece contour. On the other hand, if the program path 150 is away from the workpiece 14, as in case of point 156, the controller would bring the tool tip closer to the work-piece 14 until the equilibrium is reached of 20N, resulting in a new point 256. Therefore, it should be appreciated that but for the compliant force mode the tool fixture 12 would follow path 150. If the tool fixture 12 followed path 150, the tool fixture 12 would strike the workpiece 14 when the tool fixture 12 tries to move along that part of 150 between points 151 and 152 and be too far away from the contour 141 when the tool fixture 12 moves along that part of path 150 between points 152 and 153. The use of the compliant force mode ensures that this does not happen and that the tool fixture 12 follows the contour 141 to learn the intermediate points such as for example the intermediate points 255 and 256 shown in
As described above, in the teaching of the intermediate points on the contour 141 the robot 10 holding the tool fixture 12 is moving in the compliant motion mode along the workpiece contour 141. During that teaching, the actual robot position and orientation are continuously recorded. As described above, the tool tip would always be in continuous contact with the workpiece 14, resulting in a recorded spatial relationship that is the exact replicate between the tool fixture 12 and the workpiece 14, and a robot program generated based on the recorded path can be directly used to carry out the actual process. When the robot 10 is executing the actual process, it is not necessary for the robot controller 11 to engage any force control behavior, unless such control would benefit the process.
It should be appreciated that new points 255 and 256 described above are examples of points on the contour of workpiece 14 that are intermediate adjacent guiding points. It should be further appreciated that the intermediate points 255 and 256 are not taught to robot 10 as described above in connection with
Sometimes it is desirable but not necessary to have a post processing step in the form of a post processing algorithm that operates on the continuously recorded spatial relationship, which can be conveniently recorded with a constant distance (e.g., 2 mm) between each adjacent point. As can be appreciated by one skilled in the art, in a contour where the curvature changes slowly, points with a larger distance in between the points are acceptable for the process while in the contour where the curvature changes sharply, dense points are still necessary. After the post processing step, a more efficient program with less program data can be generated for production use.
The post processing algorithm is a further improvement of the continuously recorded path. In practice, the continuously recorded path is time sampled which can give rise to many thousands of points. For example, if the time sampling occurs every 4 ms and the robot is moving at 4 mm/s, then 1000 points/s or 1000 points/4 mm are recorded. These recorded points are the “intermediate points” described herein. This large number of points gives a great accuracy to the spatial relationship between the robot tool fixture 12 relative to the workpiece 14, however, in cases where the contour of the workpiece does not change that dramatically, it is a waste of resources. The post processing algorithm processes this large volume of points intelligently by removing unnecessary intermediate points so that the final path has both enough fine details and a minimum number of points.
One example of such a post processing algorithm is described below with reference to
In this algorithm, a path point buffer 60 of
The algorithm checks if two or more long line segments are formed by the unmerged points 4 to 14 according to a Path Accuracy Parameter specified by the user of the algorithm. This determination of the formation of two or more long line segments is achieved by verifying the distance of each unmerged point (point 4 to point 14) to the line 62 connecting the first unmerged point (point 4) and the last unmerged point (point 14). If there is one distance greater than the Path Accuracy Parameter, merging is then performed as shown in the lower part of
Post processing can also be used to correct the learned positions and orientations. This often happens for demanding applications such as arc welding where the posture of the tool is required to maintain fixed spatial relations with the path.
For most processes, such as a deburring process with a milling tool, it is usually required that the tool maintain a certain angle with the work surface. This requirement is first achieved while teaching the guiding points when the operator can determine the required angle at each taught guiding point. This is where the lead through programming method is desirable for intuitiveness and convenience. While the robot is in compliant continuous motion where the robot uses compliant force control to finalize the guiding points and teach one or more and possibly as described above thousands of intermediate points, this desired angle is still preserved since the orientation of the robot is still in closed-loop position control. When the desirable angle between two guiding points are different, all existing industrial robot controllers will provide an interpolated orientation between the two points, which is preserved in the step of compliant path learning. As is described above the set of compliant path learning is where the robot uses compliant force control to finalize the guiding points and teach one or more and possibly thousands of intermediate points, resulting in a continuous and smooth transition between the two distinctive guiding points.
To facilitate the above programming process, an artificially tangible tool with the same guiding dimensions as the real process tool is usually desirable. For example, in the deburring process with an end milling tool, moving the tool with a sharp cutting edge along the workpiece surface can create undesirable friction and damage to the part's surface. However, an artificially tangible tool with a cylindrical shape that has the same dimension as the real tool eliminates the problem and greatly enhances the programming experience.
Referring to
The method of the present invention is unique in that it uses the contact behavior of the robot under compliant force control so that such movement of the robot 10 follows the exact process path to be programmed. During this procedure, the operator is only involved with the first step of teaching the gross movement of the robot 10, while the bulk of the steps are automated with the robot controller 11.
It is easy to realize that the steps shown in
As can be appreciated, the present invention can provide tremendous savings in time and cost for the teaching and programming of a robot, especially for applications and processes where tool fixture and workpiece contact is required, for example, a deburring process. With little modification, the same programming principle can be readily applied to other processes, such as a milling or cutting process where alignment of tool orientation with the cutting plane is of ultimate importance. It is then appropriate in that case to activate compliant force control in the ZT, RxT, RyT direction/orientation to allow surface alignment, as shown in
In general, the present invention is aimed at programming the target of a trajectory using force control while teaching only a few and simple guiding points. The prior art solution is for general curvature tracking. This works for surface programming, but is a slow method that requires a lot of operator work to define where on the surfaces the tracking is needed and how to handle discontinuities and surface roughness.
The method of the present invention can also utilize a CAD (Computer Aided Design) method for offline programming of suitable points for the programming of the object coordinate system, definition of guiding points (which means that manual teaching of the guiding points is not needed) along the surfaces or edges that the trajectory is defined relative to, and the definition of force directions along the trajectory for fine tuning.
It should be appreciated that the present invention allows an operator to program robot motions with minimum effort in the real setup so that the process requirement and motion path are presented together and the operator can provide in-situ and timely judgment for any particulars in programming the robot. It should further be appreciated that the present invention allows that a spatial relationship between the robot and the workpiece is constantly maintained without imposing a burden on the operator to maintain that relationship.
It should also be further appreciated that the present invention prevents the robot and the workpiece from colliding even for applications where close contact is necessary. It should also be appreciated that the present invention frees the operator from having to guide the robot in three dimensions for all the program points in a complex path during programming operations in an effort to reduce the complexity of programming desired motions of the robot.
The simplification of the teaching operation provided by the present invention allows for more efficient and automated teaching of robot motions, making frequent programming of programmable robots for small batch jobs economically feasible.
It is to be understood that the description of the foregoing exemplary embodiment(s) is (are) intended to be only illustrative, rather than exhaustive, of the present invention. Those of ordinary skill will be able to make certain additions, deletions, and/or modifications to the embodiment(s) of the disclosed subject matter without departing from the spirit of the invention or its scope, as defined by the appended claims.
This application claims the priority of U.S. provisional patent application Ser. No. 60/656,435 filed on Feb. 24, 2005, entitled “Method For Automated Robot Path Learning” the contents of which are relied upon and incorporated herein by reference in their entirety, and the benefit of priority under 35 U.S.C. 119(e) is hereby claimed.
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/US2006/004677 | 2/10/2006 | WO | 00 | 8/21/2007 |
Number | Date | Country | |
---|---|---|---|
60656435 | Feb 2005 | US |