Workpiece self-adaptive positioning compensation method for multi-robot cooperative grinding and polishing

Abstract

The invention provides a workpiece self-adaptive positioning compensation method for multi-robot cooperative grinding and polishing, and relates to the technical field of intelligent machining and grinding and polishing. The method comprises the following steps: firstly, calibrating eyes on the hand and outside the hand for a carrying measurement robot, a plurality of grinding robots and scanning equipment in the grinding and polishing system to obtain a calibration result matrix of the eyes on the hand and a calibration result matrix of the eyes outside the hand; calibrating the base coordinate system of the workpiece under the polishing robot to obtain a base coordinate system matrix of the workpiece under the coordinate system of the polishing robot; in the actual machining process, actual workpiece point cloud data and the poses of the corresponding carrying and measuring robots are obtained, a workpiece base coordinate system matrix calibrated through vision under a plurality of grinding robot coordinate systems is calculated, and workpiece positioning is achieved. According to the method, the positioning precision and positioning efficiency of the system can be improved, and the stability and consistency of machining quality are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent processing and grinding and polishing, and in particular to a workpiece adaptive positioning compensation method for multi-robot collaborative grinding and polishing. Background Art

[0002] At present, the robot positioning function used in the automated polishing of the surfaces of complex structural parts such as aircraft engine components is mainly carried out using the traditional teaching and pointing method. This method has low efficiency, poor repeatability, and a low degree of automation, which seriously affects the rhythm, efficiency, and output in automated production.

[0003] Since the accuracy of the teaching-to-point method is determined by the operator's proficiency and operating skills, it has its own uncertainty. In addition, each workpiece in mass production must be positioned repeatedly and independently, which consumes a lot of manpower, material resources, time, etc. Therefore, research on automated calibration technology is essential.

[0004] In addition, there is still research progress on workpiece positioning for single-robot automated processing systems, but there is still a lack of collaborative processing positioning methods for multi-robot systems. Single-robot automated processing systems perform well when positioning workpieces that are small in size, light in weight, and easy to clamp, but it is difficult to simultaneously meet the positioning and processing requirements for workpieces that are large in size, heavy in weight, and difficult to clamp. Therefore, there is an urgent need for a workpiece positioning method for multi-robot collaborative grinding and polishing. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a workpiece adaptive positioning compensation method for multi-robot collaborative grinding and polishing in view of the deficiencies of the above-mentioned prior art, so as to achieve the purpose of multi-robot joint positioning of workpieces by combining theoretical calculation analysis with automatic control.

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a workpiece adaptive positioning compensation method for multi-robot collaborative grinding and polishing, based on the scanning device connected to the handling and measuring robot in the grinding and polishing system to complete the collection of the workpiece point cloud for positioning compensation calculation, specifically including the following steps:

[0007] Step 1: perform eye-on-hand and eye-out-hand calibration on the handling and measuring robot and multiple polishing robots and scanning devices in the polishing system to obtain eye-on-hand calibration result matrix and eye-out-hand calibration result matrix;

[0008] Step 2: Calibrate the base coordinate system of the workpiece under the grinding robot to obtain the base coordinate system matrix of the workpiece under the grinding robot coordinate system;

[0009] Step 3: Based on the multiple calibration result matrices obtained in step 1 and the base coordinate system matrix of the workpiece in the polishing robot coordinate system obtained by calibration in step 2, the actual workpiece point cloud data and the corresponding handling and measurement robot posture are acquired, and the base coordinate system matrix of the workpiece calibrated by vision in multiple polishing robot coordinate systems is calculated to achieve workpiece positioning.

[0010] Preferably, the step 3 comprises:

[0011] Step 3.1: Based on the multiple calibration result matrices obtained in step 1 and the base coordinate system matrix of the workpiece in the grinding robot coordinate system obtained by calibration in step 2, the position transfer matrix from the handling and measuring robot to the grinding robot during calibration is calculated;

[0012] In the actual processing process, the relative position relationship between the handling and measuring robot and the multiple grinding robots remains unchanged. Therefore, at any time, the position transfer moment from the handling and measuring robot to the grinding robot is equal to the position transfer matrix from the handling and measuring robot to the grinding robot during calibration.

[0013] Step 3.2: Obtain the workpiece point cloud data and the corresponding handling and measuring robot posture during the actual processing, and convert the workpiece point cloud data from the scanning device coordinate system to the polishing robot coordinate system;

[0014] Step 3.2.1: Keep the workpiece in the actual processing process consistent with the clamping method during calibration, and the worktables in the grinding and polishing system are at the same initial angle; ensure that multiple grinding robots are moved to a safe position so that the handling and measuring robot will not interfere with the grinding robot during the movement process;

[0015] Step 3.2.2: The handling and measuring robot moves to a certain position p, and the scanning device collects the point cloud data of the workpiece at this position, ensuring that the point cloud data can cover a certain plane or curved surface of the workpiece; wherein each point position of the point cloud data is V = (x, y, z), then there is a corresponding point cloud P = {V};

[0016] Step 3.2.3: Repeat step 3.2.2 until N point cloud data are obtained. The corresponding transport measurement robot position p n , n = 1, ..., N, and record the position matrix of the handling and measuring robot at the corresponding position; ensure that the point cloud at these N positions can basically cover all the spatial information in the processing flow channels of several workpieces;

[0017] Step 3.2.4: Summarize the actual scanning process obtained in step 3.2.3, where the handling measurement robot is located at point p n The pose matrix and the workpiece point cloud in the corresponding scanning device coordinate system The position transfer matrix from the polishing robot to the scanning device in the actual scanning process is calculated; the workpiece point cloud in the polishing robot coordinate system in the actual scanning process is further calculated;

[0018] Step 3.3: Obtain the workpiece base coordinate system matrix using visual calibration under the grinding robot to achieve workpiece positioning;

[0019] Step 3.3.1: Filter and crop the N workpiece point clouds in the grinding robot coordinate system obtained during the actual scanning process and then merge them to obtain the point cloud of the local workpiece in the grinding robot coordinate system;

[0020] Step 3.3.2: Obtain the digital model point cloud of the workpiece, and calculate the calibrated digital model point cloud of the workpiece in the grinding robot coordinate system in combination with the workpiece base coordinate system calibrated in the grinding robot coordinate system;

[0021] Step 3.3.3: Apply the ICP algorithm to calculate the transfer matrix from the point cloud of the local workpiece in the grinding robot coordinate system to the digital model point cloud of the calibrated workpiece in the grinding robot coordinate system;

[0022] Step 3.3.4: Based on the transfer matrix of the point cloud of the local workpiece in the grinding robot coordinate system offset to the digital model point cloud of the calibrated workpiece in the grinding robot coordinate system and the workpiece base coordinate system matrix calibrated by the digital model in the grinding robot coordinate system, the workpiece base coordinate system matrix calibrated by vision in the grinding robot coordinate system is calculated; the grinding robot uses the workpiece base coordinate system matrix calibrated by vision to realize the positioning of the workpiece.

[0023] Preferably, the method further performs error compensation on the workpiece base coordinate system matrix obtained by visual calibration in step 3 to improve positioning accuracy so as to achieve a workpiece positioning result that meets the requirements, including:

[0024] 1) When the base coordinate system of the workpiece is calibrated in the polishing robot coordinate system in step 2, the workpiece is ensured to be in the initial position, and the actual workpiece point cloud data and the corresponding handling and measuring robot posture are obtained;

[0025] 2) Performing the operations of step 3.2 to step 3.3 on the actual workpiece point cloud data obtained in 1) and the corresponding handling and measuring robot posture, obtaining the transfer matrix of the local workpiece point cloud offset under the grinding robot at the calibrated position to the digital model point cloud of the calibrated workpiece under the grinding robot; and then calculating the comprehensive error between the workpiece base coordinate system matrix calibrated by visual calibration under the grinding robot at the calibrated position and the workpiece base coordinate system matrix calibrated by digital model under the grinding robot;

[0026] 3) Use the workpiece base coordinate system matrix calculated using the visual calibration under the grinding robot obtained by the comprehensive error correction step 3.3.

[0027] Preferably, the method further introduces turntable control to perform approximate compensation of the workpiece base coordinate system, including:

[0028] (1) Based on the comprehensive error between the workpiece base coordinate system matrix calibrated by vision in the grinding robot coordinate system at the calibrated position and the workpiece base coordinate system matrix calibrated by digital model in the grinding robot coordinate system, and the transfer matrix of the local workpiece point cloud in the grinding robot coordinate system offset to the digital model point cloud of the calibrated workpiece in the grinding robot coordinate system during the actual processing, the correction matrix in the grinding robot coordinate system is calculated;

[0029] (2) converting the calculated correction matrix in the grinding robot coordinate system into Cartesian correction coordinates that can be recognized by the robot;

[0030] (3) The turntable in the polishing system is rotated for compensation according to the Cartesian correction coordinates in the polishing robot coordinate system to process the workpiece, that is, the compensation rotation angle of the turntable is calculated.

[0031] The beneficial effects of adopting the above technical solution are: the workpiece adaptive positioning compensation method for multi-robot collaborative grinding and polishing provided by the present invention can improve the system positioning accuracy and positioning efficiency, ensure the stability and consistency of processing quality, and adapt to the overall positioning calibration of the coordinated processing process between multiple robots, which is advanced and can be used as a reference. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A schematic diagram of the structure of a grinding and polishing system provided in an embodiment of the present invention;

[0033] Figure 2 A schematic diagram of the structure of a grinding workbench provided in an embodiment of the present invention;

[0034] Figure 3 A schematic diagram of the structure of a polishing robot provided by an embodiment of the present invention;

[0035] Figure 4 A schematic diagram of the tool rack structure provided by an embodiment of the present invention;

[0036] Figure 5 A process flow chart of a grinding and polishing system provided in an embodiment of the present invention.

[0037] In the figure: 1. Handling and measuring robot; 2. Loading and unloading table; 3. Turning table; 4. Grinding workbench; 5. Grinding robot; 6. Tool rack; 7. Turntable; 8. Three-jaw chuck; 9. Workpiece; 10. Quick change; 11. Grinding tool; 12. Spindle; 13. Belt sander; 14. Grinding wheel. DETAILED DESCRIPTION

[0038] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0039] In this embodiment, a workpiece adaptive positioning compensation method for multi-robot collaborative grinding and polishing is implemented based on the grinding and polishing system;

[0040] Grinding and polishing system Figure 1 As shown, it includes a loading and unloading measurement system, a grinding workbench 4, multiple grinding robots, a tool rack 6 and an electrical control system;

[0041] The loading and unloading measurement system consists of a handling and measuring robot 1, i.e., a 300kg six-axis industrial robot of Kuka, a loading and unloading table 2, and a flipping table 3, which completes the loading and unloading and flipping of workpieces on a polishing table 4. At the same time, the handling and measuring robot completes the collection of workpiece point clouds through a rigid connection scanning device for positioning calculation;

[0042] like Figure 2 As shown, the grinding workbench 4 is a rotating mechanism, which is composed of a turntable 7, a three-jaw chuck 8, and a workpiece 9. For a workpiece 9 of a cylindrical, conical or other structure, the three-jaw chuck 8 is used to position the workpiece from the inside.

[0043] like Figure 3 As shown, the grinding robot 5 is a KUKA 70kg six-axis industrial robot. The end of the robot flange is connected to the grinding tool 11 through a quick change 10, which can realize rapid switching of different tools and grinding processing of different areas of the workpiece.

[0044] like Figure 4 As shown, the tool rack 6 includes a spindle 12, a belt sander 13, and a grinding wheel 14 which are connected in sequence, and can complete automated processing of different areas.

[0045] In this embodiment, the grinding and polishing system includes grinding robots R1 and R2 carrying grinding tools, and a handling and measuring robot R3 integrating handling and measuring functions. Each robot performs handling, measurement and positioning, grinding and polishing on the workpiece on the grinding workbench, and collaborates to complete the grinding and polishing of the workpiece; the handling and measuring robot R3 completes loading by clamping the workpiece from the loading and unloading table to the grinding workbench through the tooling, completes flipping by clamping the workpiece from the grinding workbench to the flipping table through the tooling, and completes unloading by clamping the workpiece from the flipping table to the loading and unloading table through the tooling; at the same time, the handling and measuring robot R3 completes the acquisition of the workpiece point cloud through the scanner C rigidly connected to the tooling for positioning calculation; the grinding workbench is a rotating platform, on which is provided a three-jaw chuck and a turntable for fixing and clamping the workpiece to be ground; the grinding robots R1 and R2 are industrial robots, and the end flange of the robot is connected to the grinding tool for grinding the workpiece.

[0046] The method is as follows Figure 5 As shown, the following steps are included:

[0047] Step 1: In the preparation stage, the handling measurement robot R3 and scanner C in the polishing system are calibrated with eyes in the hand to obtain the eye-in-hand calibration result matrix, that is, the hand-eye calibration matrix. The hand-eye calibration matrix represents the positional relationship between the flange of robot R3 and the scanner; the eye-out-of-hand calibration of robots R1, R2 and scanner C is performed to obtain the eye-out-of-hand calibration result matrix The hand-eye calibration matrix represents the positional relationship between the robots R1 and R2 and the scanner respectively.

[0048] In this embodiment, a calibration ball with an accuracy test report is prepared, the test radius of the calibration ball is obtained, and a tooling connector that can fix the calibration ball to the end of the robot flange and a connector that can place the calibration ball on a horizontal plane are prepared.

[0049] Step 1.1: Calibrate the eyes on the hand of the handling measurement robot R3 and scanner C. The process is as follows:

[0050] Step 1.1.1: Fix the calibration ball on a horizontal plane in a certain posture to ensure that it is aligned with the base coordinate system of robot R3 during the subsequent acquisition of the workpiece point cloud. The relative position relationship does not change and will not be affected by occlusion, interference and other conditions.

[0051] Step 1.1.2: Move the robot R3 so that the calibration ball is within the field of view of the scanner C to collect the workpiece point cloud data P C , and record the robot R3 in its base coordinate system and tool coordinate system The coordinate value of in, The robot R3 is in its base coordinate system The three-dimensional coordinate values ​​under The robot R3 is in its base coordinate system The three-dimensional coordinate values ​​of .

[0052] Generally, the subsequent The default coordinate value is R m The base coordinate system of the robot and tool coordinate system Next (m=1,2,3).

[0053] In addition, define the R3 robot flange and scanner C to perform eye-on-hand calibration result matrix for:

[0054]

[0055] Step 1.1.3: Repeat step 1.1.2 until four sets of robot R3 poses are obtained. And the point cloud collected by scanner C in the corresponding position And ensure that the following requirements are met:

[0056] ① The four groups of robot R3 postures The quantities vary, and The quantities are the same, that is

[0057]

[0058] ② The four groups of robot R3 postures The weight difference between the two weights should be as large as possible and not less than 5mm, that is,

[0059]

[0060] ③ Workpiece point cloud collected by scanner C It should be ensured that: each point cloud contains half a complete and independent spherical point cloud without interference, that is, it can be finally cut into an independent hemispherical point cloud; the spherical point cloud is uniform and smooth without large noise data.

[0061] Step 1.1.4: Repeat step 1.1.2 until six sets of robot R3 poses are obtained. And the point cloud collected by scanner C in the corresponding position And ensure that the following requirements are met:

[0062] ① The six groups of robot R3 postures The quantities vary, i.e.

[0063]

[0064] ② These six groups of robot postures The weight difference between the two weights should be as large as possible and not less than 5 mm; the six groups of robot postures The difference between the two components should be as large as possible and not less than 5°, that is,

[0065]

[0066] ③Point cloud collected by scanner C It should be ensured that: each point cloud contains half a complete and independent spherical point cloud without interference, that is, it can be finally cut into an independent hemispherical point cloud; the spherical point cloud is uniform and smooth without large noise data.

[0067] Step 1.1.5: Point cloud collected by scanner C Filter and crop. If the number of points in the point cloud exceeds the set value, additional downsampling is performed to obtain a post-processed point cloud containing only the calibration ball point cloud.

[0068] Step 1.1.6: Using four sets of robot R3 poses and post-processing point clouds Calculate the hand-eye calibration matrix Quantity; and use Components and six groups of robot R3 poses and post-processing point clouds Calculate the hand-eye calibration matrix Component, that is, the hand-eye calibration matrix

[0069] Step 1.1.7: Use the obtained hand-eye calibration matrix and and point cloud The 10 hemispherical point clouds in the scanner C coordinate system are calculated and merged into the complete spherical point cloud in the robot R3 coordinate system, that is,

[0070]

[0071] in, It is the complete spherical point cloud in the robot R3 coordinate system;

[0072] Step 1.1.8: Complete sphere point cloud Do spherical fitting and get the fitting radius RMS Error According to the measurement radius of the calibration sphere in the given calibration sphere detection report Get the fitting radius error for:

[0073]

[0074] In order to eliminate accidental errors, multiple fittings are performed to obtain the average value, and the evaluation unit is ensured to be mm, and the final average fitting radius error is obtained. The average root mean square error Qualified calibration results should meet the following requirements:

[0075]

[0076] If not, repeat steps 1.1.1 to 1.1.7 until the requirements are met.

[0077] Step 1.2: Calibrate the eyes of the polishing robot R1 and scanner C outside the hand. The process is as follows:

[0078] Step 1.2.1: Fix the calibration ball at the end of the flange of the grinding robot R1 in a certain posture to ensure that it is consistent with the tool coordinate system of the grinding robot R1 during the subsequent point cloud collection process. The relative position relationship does not change and will not be affected by occlusion, interference, etc. The mobile robot R3 makes the scanner C in a position convenient for scanning the calibration ball, and records the position of the R3 robot in its base coordinate system at this time. and tool coordinate system The position matrix In addition, the eye-out-of-hand calibration result matrix of the R1 robot and scanner C is defined for:

[0079]

[0080] Step 1.2.2: Move the polishing robot R1 so that the calibration ball is within the field of view of the scanner C and collect the point cloud data P C’ , and record the grinding robot R1 in its base coordinate system and tool coordinate system The coordinate value of

[0081] Step 1.2.3: Repeat step 1.2.2 until four sets of polishing robot R1 poses are obtained. And the point cloud collected by scanner C in the corresponding position And ensure that the following requirements are met:

[0082] ① These four groups of grinding robot postures The quantities vary, and The quantities are the same, that is

[0083]

[0084] ② These four groups of grinding robot postures The weight difference between the two weights should be as large as possible and not less than 5mm, that is,

[0085]

[0086] ③Point cloud collected by scanner C It should be ensured that: each point cloud contains half a complete and independent spherical point cloud without interference, that is, it can be finally cut into an independent hemispherical point cloud; the spherical point cloud is uniform and smooth without large noise data.

[0087] Step 1.2.4: Repeat step 1.2.3 until six sets of polishing robot poses are obtained. And the point cloud collected by scanner C in the corresponding position And ensure that the following requirements are met:

[0088] ① These six groups of robot postures The quantities vary, i.e.

[0089]

[0090] ② The four groups of robot postures The weight difference between the two weights should be as large as possible and not less than 5 mm; the four groups of robot postures The difference between the two components should be as large as possible and not less than 5°, that is,

[0091]

[0092] ③Point cloud collected by scanner C It should be ensured that: each point cloud contains half a complete and independent spherical point cloud without interference, that is, it can be finally cut into an independent hemispherical point cloud; the spherical point cloud is uniform and smooth without large noise data.

[0093] Step 1.2.5: Point cloud collected by scanner C Filter and crop. If the number of points in the point cloud exceeds the set value, additional downsampling is performed to obtain a post-processed point cloud containing only the calibration ball point cloud.

[0094] Step 1.2.6: Exploitation and point cloud Calculate the hand-eye calibration matrix Quantity; and use Quantity and and point cloud Calculate the hand-eye calibration matrix Component, that is, the hand-eye calibration matrix

[0095] Step 1.2.7: Use the obtained hand-eye calibration matrix to and point cloud The 10 hemispherical point clouds in the scanner C coordinate system are calculated and merged into the complete spherical point cloud in the robot R1 coordinate system, that is,

[0096]

[0097] in, It is the complete spherical point cloud in the robot R1 coordinate system;

[0098] Step 1.2.8: Complete spherical point cloud in the robot R1 coordinate system Do spherical fitting and get the fitting radius RMS Error According to the measurement radius of the calibration sphere in the given calibration sphere detection report Get the fitting radius error for:

[0099]

[0100] In order to eliminate accidental errors, multiple fittings are performed to obtain the average value, and the evaluation unit is ensured to be mm, and the final average fitting radius error is obtained. The mean root mean square error Qualified calibration results should meet the following requirements:

[0101]

[0102] If it is not satisfied, repeat steps 1.2.1 to 1.2.7 until the requirements are met. The final calibration result matrix of the R1 robot and scanner C with the eye outside the hand is:

[0103] Step 1.3: Calibrate the eyes of the polishing robot R2 and scanner C outside the hand. The process is the same as the calibration process of the polishing robot R1 and scanner C outside the hand in step 1.2. Among them, the position of robot R3 remains unchanged from step 1.2.1. The calibration result matrix of the polishing robot R2 and scanner C with the eye outside the hand is:

[0104] Step 2: Calibrate the base coordinate system of the workpiece under robots R1 and R2 to obtain the base coordinate system matrix of the workpiece under robots R1 and R2

[0105] Step 2.1: Calibrate the base coordinate system of the workpiece under the grinding robot R1. The calibration process is carried out according to the standard process specified by the robot used. The base coordinate system matrix of the workpiece under the grinding robot R1 is obtained by calibration:

[0106] Step 2.2: Calibrate the base coordinate system of the workpiece under the grinding robot R2. The calibration process is carried out according to the standard process specified by the robot used. The base coordinate system matrix of the workpiece under the grinding robot R2 is obtained by calibration:

[0107] Step 3: Perform the basic application stage. Based on the multiple calibration result matrices obtained from the calibration of the grinding and polishing system in the preparation stage of step 1, obtain the actual workpiece point cloud data and the corresponding robot R3 posture, and calculate the workpiece base coordinate system matrix using visual calibration under robot R1 and robot R2; if the grinding and polishing system does not require high accuracy for visual calibration, or the accuracy of the scanner and hand-eye calibration is high enough, you only need to use the operating process of the "basic application stage" to obtain workpiece positioning results that meet the requirements.

[0108] Step 3.1: Summarize the data of the grinding and polishing system calibration phase: It is known that the calibration result matrix of robot R3 and scanner C for eye-on-hand is It is known that the calibration result matrix of robot R1 and scanner C with eyes outside the hand is It is known that the calibration result matrix of robot R2 and scanner C with eyes outside the hand is It is known that when robots R1 and R2 perform eye-in-hand external calibration with scanner C, the pose matrices of robot R3 are It is known that the workpiece base coordinate system matrix calibrated by digital model under robot R1 is The workpiece base coordinate system matrix calibrated by digital model under robot R2 is:

[0109] Step 3.2: Based on the data summarized in step 3.1, calculate the position transfer matrix from robot R3 to robot R1 during calibration for:

[0110]

[0111] Similarly, during calibration, the position transfer matrix from robot R3 to robot R2 is for:

[0112]

[0113] Obviously, the relative position relationship between robots R1, R2, and R3 remains unchanged. Therefore, at any time, the position transfer moment from the transport and measurement robot to the polishing robot is equal to the position transfer matrix from the transport and measurement robot to the polishing robot during calibration, that is:

[0114]

[0115] in, is the position transfer moment from robot R3 to robot R1 at any time, is the position transfer moment from robot R3 to robot R2 at any time;

[0116] Step 3.3: Acquire the actual workpiece point cloud data and the corresponding robot R3 posture, convert the workpiece point cloud data from the scanner coordinate system to the robot coordinate system, obtain the workpiece base coordinate system matrix using visual calibration under robots R1 and R2, and realize the positioning of the workpiece. The process is as follows:

[0117] Step 3.3.1: Keep the workpiece in actual processing in the same clamping mode as that during calibration, and the worktables are at the same initial angle. Ensure that both robots R1 and R2 are moved to a safe position so that robot R3 will never interfere with robot R1 or robot R2 during movement.

[0118] Step 3.3.2: Move the robot R3 to a certain position p, and the scanner C collects the point cloud data at this position to ensure that the point cloud data can cover a certain plane or curved surface of the workpiece; wherein, each point of the point cloud data is V = (x, y, z), and there is a corresponding point cloud P = {V};

[0119] Step 3.3.3: Repeat step 3.3.2 until N point cloud data are obtained. The corresponding robot R3 position p n , n = 1, ..., N, and record the robot R3 pose matrix at the corresponding position Ensure that the point clouds at these N positions can basically cover all the spatial information in the processing flow channels of several workpieces;

[0120] Step 3.3.4: Summarize the actual scanning process obtained in step 3.3.3, where robot R3 is located at point p n The pose matrix The workpiece point cloud corresponding to the scanner C coordinate system Calculate the position transfer matrix from robot R1 to scanner C during the actual scanning process for:

[0121]

[0122] Similarly, the position transfer matrix from robot R2 to scanner C during the actual scanning process is for:

[0123]

[0124] Further calculation shows that the workpiece point cloud under robot R1 during the actual scanning process is:

[0125]

[0126] Similarly, the workpiece point cloud under robot R2 during the actual scanning process is:

[0127]

[0128] Step 3.3.5: Filter and crop the N workpiece point clouds obtained during the actual scanning process under robot R1 and merge them to obtain the point cloud of the local workpiece under robot R1:

[0129]

[0130] Similarly, the point cloud of the local workpiece under robot R2 is

[0131]

[0132] in, Represents the result of filtering and cropping the original point cloud V. The cropping method is to retain only the valid matching point cloud and delete the noise and external interference.

[0133] Step 3.3.6: Obtain the workpiece digital model point cloud Combined with the workpiece base coordinate system calibrated under robot R1 The calculated digital model point cloud of the calibrated workpiece under robot R1 is:

[0134]

[0135] Similarly, the digital model point cloud of the workpiece calibrated by robot R2 is:

[0136]

[0137] Step 3.3.7: Apply the ICP algorithm to calculate the transfer matrix from the point cloud offset of the local workpiece under robot R1 to the digital model point cloud of the calibrated workpiece under robot R1:

[0138]

[0139] Similarly, the transfer matrix of the point cloud of the local workpiece under robot R2 to the digital point cloud of the calibrated workpiece under robot R2 is

[0140]

[0141] Step 3.3.8: The final calculated workpiece base coordinate system matrix using visual calibration under robot R1 is

[0142]

[0143] Similarly, the workpiece base coordinate system matrix using visual calibration under robot R2 is

[0144]

[0145] Robot R1 and robot R2 use the workpiece base coordinate system matrix calibrated by vision Realize the positioning of workpiece.

[0146] In this embodiment, if the polishing system requires high accuracy for visual calibration, or the accuracy of scanner and hand-eye calibration is not high, it is necessary to perform the operation process of "Extended Application Phase 1" to compensate for the error of the workpiece base coordinate system matrix obtained by visual calibration in step 2, improve the positioning accuracy, and achieve a workpiece positioning result that meets the requirements. This item is an optional function after completing steps 1-3.

[0147] 1) In the preparation stage, expand step 1.6, i.e., calibrate the base coordinate system of the workpiece under robots R1 and R2, ensure that the workpiece is in the initial position, and perform step 3.3 of the basic application stage, i.e., obtain the actual workpiece point cloud data and the corresponding R3 robot posture;

[0148] 2) Perform steps 3.4 to 3.7 of the basic application stage on the results in 1) to obtain the transfer matrix of the point cloud of the local workpiece under the robot R1 at the calibrated position to the digital model point cloud of the calibrated workpiece under the robot R1 and the transfer matrix of the point cloud of the local workpiece under the robot R2 at the calibrated position to the digital model point cloud of the calibrated workpiece under the robot R2; the comprehensive error of the workpiece base coordinate system matrix calibrated by visual calibration under the robot R1 at the calibrated position and the workpiece base coordinate system matrix calibrated by digital model under the robot R1 for:

[0149]

[0150] Similarly, the comprehensive error between the workpiece base coordinate system matrix calibrated by vision under robot R2 at the calibrated position and the workpiece base coordinate system matrix calibrated by digital model under robot R2 is for:

[0151]

[0152] The subscript calib indicates that the data is only valid for the workpiece at the calibrated position;

[0153] 3) In actual application, the default value in step 3.2 is After completing the entire process of grinding and polishing system calibration to step 3.7, modify step 3.8 to calculate the workpiece base coordinate system matrix calibrated by vision under robot R1 for:

[0154]

[0155] Similarly, the workpiece base coordinate system matrix of robot R2 using visual calibration is corrected and calculated for:

[0156]

[0157] In this embodiment, when the offset of the workpiece is too large, in addition to considering the influence of the change of the base coordinate system of the workpiece, it is also necessary to consider the interference between the tool and the workpiece. When the tool movement path is more complicated or the internal margin of the workpiece is small, it is possible to consider abandoning the correction amount of part of the matrix and introducing the turntable control to perform approximate compensation of the workpiece base coordinate system. This item is an optional function after completing the calibration of the grinding and polishing system, the basic application stage and the extended application stage 1, which is the extended application stage 2.

[0158] (1) Summarize the comprehensive error between the workpiece base coordinate system matrix calibrated by vision under robot R1 at the calibrated position and the workpiece base coordinate system matrix calibrated by digital model under robot R1 This is in contrast to step 3.7 of the basic application phase in the actual application process, i.e., the transfer matrix of the point cloud of the local workpiece under robot R1 to the digital model point cloud of the calibrated workpiece under R1. Calculate the correction matrix of robot R1 for:

[0159]

[0160] Similarly, the correction matrix of robot R2 is calculated for:

[0161]

[0162] (2) The calculated correction matrix of robot R1 is converted into Cartesian coordinates that can be recognized by the robot:

[0163]

[0164] Similarly, the calculated correction matrix of robot R2 is converted into Cartesian coordinates recognizable by the robot:

[0165]

[0166] Among them, a, b, and c are the rotation components around the z, y, and x directions respectively.

[0167] (3) The turntable is rotated and processed according to the correction coordinates of robots R1 and R2, that is, the compensation rotation angle rotation_angel of the turntable is calculated:

[0168]

[0169] Among them, is(up)=1 means that the workpiece is on the front side during processing, and is(up)=0 means that the workpiece is on the back side during processing.

[0170] In this embodiment, based on the workpiece adaptive positioning compensation method for multi-robot collaborative grinding and polishing of the present invention, the process flow of the grinding and polishing system is as follows: Figure 5 As shown, specifically:

[0171] S1: Complete the hand-eye calibration of robots R1, R2, R3 and scanner c, and use the workpiece digital model to complete the base coordinate system calibration of robots R1 and R2.

[0172] S2: When the workpiece is at the calibration position, the mobile robot R3 collects several point clouds and saves the corresponding positions. Using the calibration results of S1, the comprehensive error matrix of R1 and R2 is calculated.

[0173] S3: When the workpiece is in production, that is, it is clamped for the second time, the R3 robot is moved according to the position in S2 to collect point clouds. Using the results of S1 and S2, the compensation angle between the corrected base coordinate system and the turntable in the robot R1 and R2 coordinate systems is calculated.

[0174] S4: If there is no interference, the corrected base coordinate system under the robot R1 and R2 coordinate systems can be passed to the robot program, and the robot program will modify the saved base coordinate system. If there is interference, the compensation angle will be passed to the robot program, and the robot program will trigger the turntable to rotate to complete the compensation correction, and then the grinding process will be carried out.

[0175] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope defined by the claims of the present invention.

Claims

1. A workpiece adaptive positioning compensation method for multi-robot collaborative grinding and polishing, characterized by: The scanning device connected to the handling and measuring robot in the grinding and polishing system completes the acquisition of the workpiece point cloud for positioning compensation calculation, which specifically includes the following steps: Step 1: perform eye-on-hand and eye-out-hand calibration on the handling and measuring robot and multiple polishing robots and scanning devices in the polishing system to obtain the eye-on-hand calibration result matrix and the eye-out-hand calibration result matrix; Step 2: Calibrate the base coordinate system of the workpiece under the grinding robot to obtain the base coordinate system matrix of the workpiece under the grinding robot coordinate system; Step 3: Based on the multiple calibration result matrices obtained in step 1 and the base coordinate system matrix of the workpiece in the polishing robot coordinate system obtained by calibration in step 2, the actual workpiece point cloud data and the corresponding handling and measurement robot posture are acquired, and the base coordinate system matrix of the workpiece calibrated by vision in multiple polishing robot coordinate systems is calculated to achieve workpiece positioning.

2. The method for adaptive positioning compensation of workpieces for multi-robot collaborative grinding and polishing according to claim 1, characterized in that: The step 3 comprises: Step 3.1: Based on the multiple calibration result matrices obtained in step 1 and the base coordinate system matrix of the workpiece in the grinding robot coordinate system obtained by calibration in step 2, the position transfer matrix from the handling and measuring robot to the grinding robot during calibration is calculated; Step 3.2: Obtain the workpiece point cloud data and the corresponding handling and measuring robot posture during the actual processing, and convert the workpiece point cloud data from the scanning device coordinate system to the polishing robot coordinate system; Step 3.3: Obtain the workpiece base coordinate system matrix using visual calibration under the grinding robot to achieve workpiece positioning.

3. The method for adaptive positioning compensation of workpieces for multi-robot collaborative grinding and polishing according to claim 2, characterized in that: During the actual processing, the relative position relationship between the handling and measuring robot and the multiple polishing robots remains unchanged. Therefore, at any time, the position transfer moment from the handling and measuring robot to the polishing robot is equal to the position transfer matrix from the handling and measuring robot to the polishing robot during calibration.

4. The method for adaptive positioning compensation of workpieces for multi-robot collaborative grinding and polishing according to claim 2, characterized in that: The step 3.2 comprises: Step 3.2.1: Keep the workpiece in the actual processing process consistent with the clamping method during calibration, and the worktables in the grinding and polishing system are at the same initial angle; ensure that multiple grinding robots are moved to a safe position so that the handling and measuring robot will not interfere with the grinding robot during the movement process; Step 3.2.2: The handling and measuring robot moves to a certain position p, and the scanning device collects the point cloud data of the workpiece at this position, ensuring that the point cloud data can cover a certain plane or curved surface of the workpiece; wherein each point position of the point cloud data is V = (x, y, z), then there is a corresponding point cloud P = {V}; Step 3.2.3: Repeat step 3.2.2 until N point cloud data P are obtained. n C The corresponding transport measurement robot position p n , n = 1, ..., N, and record the position matrix of the handling and measuring robot at the corresponding position; ensure that the point cloud at these N positions can basically cover all the spatial information in the processing flow channels of several workpieces; Step 3.2.4: Summarize the actual scanning process obtained in step 3.2.3, where the handling measurement robot is located at point p n The pose matrix and the workpiece point cloud in the corresponding scanning device coordinate system The position transfer matrix from the polishing robot to the scanning device in the actual scanning process is calculated; and the workpiece point cloud in the polishing robot coordinate system in the actual scanning process is further calculated.

5. The method for adaptive positioning compensation of workpieces for multi-robot collaborative grinding and polishing according to claim 2, characterized in that: The step 3.3 comprises: Step 3.3.1: Filter and crop the N workpiece point clouds in the grinding robot coordinate system obtained during the actual scanning process and then merge them to obtain the point cloud of the local workpiece in the grinding robot coordinate system; Step 3.3.2: Obtain the digital model point cloud of the workpiece, and calculate the calibrated digital model point cloud of the workpiece in the grinding robot coordinate system in combination with the workpiece base coordinate system calibrated in the grinding robot coordinate system; Step 3.3.3: Apply the ICP algorithm to calculate the transfer matrix from the point cloud of the local workpiece in the grinding robot coordinate system to the digital model point cloud of the calibrated workpiece in the grinding robot coordinate system; Step 3.3.4: Based on the transfer matrix of the point cloud of the local workpiece in the grinding robot coordinate system offset to the digital model point cloud of the calibrated workpiece in the grinding robot coordinate system and the workpiece base coordinate system matrix calibrated by the digital model in the grinding robot coordinate system, the workpiece base coordinate system matrix calibrated by vision in the grinding robot coordinate system is calculated; the grinding robot uses the workpiece base coordinate system matrix calibrated by vision to realize the positioning of the workpiece.

6. The method for adaptive positioning compensation of workpieces for multi-robot collaborative grinding and polishing according to claim 2, characterized in that: The method further performs error compensation on the workpiece base coordinate system matrix obtained by visual calibration in step 3 to improve positioning accuracy so as to achieve a workpiece positioning result that meets the requirements, including: 1) When the base coordinate system of the workpiece is calibrated in the polishing robot coordinate system in step 2, the workpiece is ensured to be in the initial position, and the actual workpiece point cloud data and the corresponding handling and measuring robot posture are obtained; 2) Performing the operations of step 3.2 to step 3.3 on the actual workpiece point cloud data obtained in 1) and the corresponding handling and measuring robot posture, obtaining the transfer matrix of the local workpiece point cloud offset under the grinding robot at the calibrated position to the digital model point cloud of the calibrated workpiece under the grinding robot; and then calculating the comprehensive error between the workpiece base coordinate system matrix calibrated by visual calibration under the grinding robot at the calibrated position and the workpiece base coordinate system matrix calibrated by digital model under the grinding robot; 3) Use the workpiece base coordinate system matrix calculated using the visual calibration under the grinding robot obtained by the comprehensive error correction step 3.

3.

7. The method for adaptive positioning compensation of workpieces for multi-robot collaborative grinding and polishing according to claim 6, characterized in that: The method also introduces turntable control to perform approximate compensation of the workpiece base coordinate system, including: (1) Based on the comprehensive error between the workpiece base coordinate system matrix calibrated by vision in the grinding robot coordinate system at the calibrated position and the workpiece base coordinate system matrix calibrated by digital model in the grinding robot coordinate system, and the transfer matrix of the local workpiece point cloud in the grinding robot coordinate system offset to the digital model point cloud of the calibrated workpiece in the grinding robot coordinate system during the actual processing, the correction matrix in the grinding robot coordinate system is calculated; (2) converting the calculated correction matrix in the grinding robot coordinate system into Cartesian correction coordinates that can be recognized by the robot; (3) The turntable in the polishing system is rotated for compensation according to the Cartesian correction coordinates in the polishing robot coordinate system to process the workpiece, that is, the compensation rotation angle of the turntable is calculated.

Images