Station arrangement optimization design method for multi-station mechanical arm calibration device

Abstract

The invention discloses a station arrangement optimization design method for a multi-station mechanical arm calibration device, and the method comprises the following steps: 1, designing a plurality of station arrangement schemes, and selecting one of the station arrangement schemes as a to-be-tested scheme; 2, mounting a measuring rod and placing a 3D measuring device in an off-line simulation environment; 3, corresponding mechanical arm joint angles are solved based on mechanical arm inverse kinematics, then the mechanical arm joint angles are screened through configuration screening, and finally a large number of feasible and reasonable mechanical arm joint angles are obtained; 4, the 3D measuring device is placed on other stations in the scheme to be measured, the step 3 is repeated, and finally all feasible and reasonable mechanical arm joint angles are obtained; 5, selecting an optimal joint angle measurement combination based on the observability indexes, obtaining the observability index of the optimal joint angle measurement combination, and recording the observability index as an optimal observability index; and 6, selecting other station arrangement schemes as to-be-tested schemes, repeating the steps 2-5, and determining an optimal station arrangement scheme by taking the optimal observability index as an evaluation index.

Description

Technical Field

[0001] The present invention relates to the field of robots, and in particular to a station layout optimization design method for a multi-station robot arm calibration device. Background Art

[0002] Chinese invention 202110278848.0 proposes a robot calibration device based on multi-station measurement, comprising a calibrated device (1), a calibration device (2), and a multi-station template (3), wherein the multi-station template (3) provides multiple calibration measurement positions for expanding the calibration measurement range of the calibration device (2) on the calibrated device (1), the calibrated device (1) is connected to a robot arm and the calibration device (2) is connected to the multi-station template (3), and the multi-station template (3) is provided with M positioning elements (2 0) and N clamping elements (21), M is greater than 1, N is greater than 1, the positioning element (20) is used to position the calibration device (2), the clamping element (21) is used to clamp the calibration device (2), and the calibration device can quickly calibrate the robot at any time in the industrial site. The calibration device can measure at multiple stations through a multi-station template, and a larger measurement range can be obtained. At the same time, the relative position between each station is a known information, which can be used for robot kinematic parameter identification to improve calibration accuracy. This method uses a multi-station fixture to realize multi-station clamping of a 3D measuring device. This design of fixing multiple stations on the same substrate enables the device to have excellent accuracy retention and portability, so that the relative position between multiple stations can be maintained for a long time, and the fixture can theoretically be quickly deployed near any robot arm for on-site calibration. However, the layout of each station of the multi-station fixture currently relies entirely on the designer's design experience and lacks theoretical support. Especially for large robotic arms, since the length of the measuring rod is limited, in order to ensure the calibration accuracy, it is necessary to give full play to the advantages of multi-station measurement. This means that the layout of multiple stations is crucial to the final calibration effect. Therefore, a more scientific station layout optimization design method is urgently needed. Summary of the invention

[0003] In order to solve the above-mentioned technical problems, the purpose of the present invention is to provide a station layout optimization design method for a multi-station robot arm calibration device, which can complete the design of the station layout scheme more scientifically, thereby giving full play to the advantages of multi-station measurement.

[0004] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solutions:

[0005] A station layout optimization design method for a multi-station robotic arm calibration device comprises the following steps:

[0006] (1) Designing multiple workstation layout schemes, wherein the workstation layout schemes include the number of workstations, the XYZ coordinates and the attitude angle of each workstation, and selecting one of the workstation layout schemes as the scheme to be tested;

[0007] (2) In an offline simulation environment, the measuring rod is installed at the end of the robot arm, and the 3D measuring device is placed at a station in the solution to be tested;

[0008] (3) maintaining the point constraint of the center of the measuring ball at the end of the measuring rod, uniformly distributing the attitude angle of the measuring rod in the measurement space of the 3D measuring device, solving the corresponding joint angles of the manipulator based on the inverse kinematics of the manipulator, and then screening them through configuration screening, and finally obtaining a large number of feasible and reasonable joint angles of the manipulator, wherein the configuration screening includes simulation collision detection;

[0009] (4) Place the 3D measurement device on other workstations in the test solution and repeat step (3) to finally obtain all feasible and reasonable robot arm joint angles;

[0010] (5) N groups of robot arm joint angles used in calibration measurement are selected from all feasible and reasonable robot arm joint angles. These N groups of robot arm joint angles are recorded as optimal measurement joint angle combinations. The optimal measurement joint angle combination is selected based on the observability index, and the observability index of the optimal measurement joint angle combination is obtained, which is recorded as the optimal observability index;

[0011] (6) Select other workstation layout schemes as the schemes to be tested, repeat steps (2) to (5), obtain the optimal observability index corresponding to different workstation layout schemes, and determine the optimal workstation layout scheme using the optimal observability index as the evaluation index.

[0012] Preferably, each workstation arrangement includes at least three workstations, and each workstation is arranged around the robot arm.

[0013] Preferably, grid coordinate points are selected at equal intervals in the working space of the robot arm, and all workstation layout plans require that the workstations be placed on corresponding grid coordinate points.

[0014] Preferably, a rotation method is used to implement the design of multiple station layout schemes in step (1) and the determination of the optimal station layout scheme in step (6), and the specific process is as follows:

[0015] a1. Select an initial multi-station solution;

[0016] b1. Optimize one of the workstations as the workstation to be optimized, keep the other workstations fixed, move the workstation to be optimized to a new position near its original position, complete the design of a new workstation layout scheme, execute steps (2) to (5), obtain the optimal observability index corresponding to the workstation layout scheme, and repeat the search near the original position of the workstation to be optimized until the optimal observability index corresponding to the multi-workstation scheme is optimal, thereby completing a position transformation;

[0017] c1. Repeat step b1 with the other stations as stations to be optimized, thus completing a round of position rotation;

[0018] d1. Repeat steps b1-c1 continuously and perform multiple rounds of position rotation. When the optimal observability index no longer increases or the increase is less than the threshold ε1, exit the loop and output the multi-position solution that obtains the optimal optimal observability index during the rotation process to complete the iterative optimization.

[0019] Preferably, in step b1, in order to achieve 2D position optimization of the workstation, a step-by-step search method is used to perform local optimization, and the steps are as follows:

[0020] a2. Take the grid coordinate point of one of the workstations as the center point, search all the grid points in the local range near the center point, and obtain the optimal observability index corresponding to all the grid points;

[0021] b2. If the optimal observability index of all grid points in the local range no longer increases relative to the optimal observability index of the center point, or the increase is less than the threshold ε2, then the loop is exited, and the grid point with the best optimal observability index in the local range is used as the optimized workstation position, otherwise the grid point with the best optimal observability index in the local range is used as the center point of the next step, recorded as the third grid point;

[0022] c2. If the number of searches reaches the set upper limit N, the loop is exited and the third grid point is used as the optimized workstation position; otherwise, all grid points within the local range of the third grid point are searched, and the optimal observability index corresponding to all grid points is obtained. The number of searches is increased by 1, and the process jumps to step b2 for solution.

[0023] Preferably, in order to determine the height parameters of the workstation, a plurality of candidate values ​​are first prepared for the height parameters of the workstation, and then in step b1, in the process of repeated search near the original position of the workstation to be optimized, the height parameters of the workstation to be optimized are set to different candidate values ​​in turn until the optimal observability index corresponding to the multi-station solution is optimal, thereby completing a position transformation.

[0024] Preferably, for various special-shaped workspaces, when there are obstacles in the working environment of the robot arm and the workstation cannot be arranged, it is necessary to first mark the grid points in the obstacle area in the workspace, and then in the process of workstation rotation optimization, in each step of the local search, that is, step a2 and step c2 are added with an obstacle judgment. The workstation to be optimized can only be moved to the grid without obstacles. Only then can a new workstation layout scheme be designed, and then steps (2)-(5) are executed to obtain the optimal observability index corresponding to the workstation layout scheme.

[0025] The present invention has the following beneficial effects due to the adoption of the above technical solution:

[0026] 1. The present invention is completely based on the optimal observable index for iterative optimization, is no longer limited by the designer's experience, has a theoretical basis support, and is easier to find the global optimum;

[0027] 2. The present invention adopts a design method based on iterative optimization, which is a heuristic algorithm, and its initial value is selected based on engineering experience, which often has global information, is relatively close to the global optimum, and has high calculation efficiency;

[0028] 3. It can meet the optimization needs of different heights, different opening directions, and special-shaped workspaces, greatly improving the diversity of multi-station layout solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a structural schematic diagram of the present invention;

[0030] Figure 2 It is a schematic diagram of 2D position optimization of the workstation of the present invention;

[0031] Figure 3 It is a schematic diagram of optimizing the height parameters and opening orientation parameters of the workstation of the present invention;

[0032] Figure 4 It is a schematic diagram of optimizing the workstation arrangement scheme of the present invention in the presence of obstacles. DETAILED DESCRIPTION

[0033] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0034] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0035] Embodiment 1,

[0036] like Figure 1 As shown, a station layout optimization design method for a multi-station robotic arm calibration device includes the following steps:

[0037] (1) Designing multiple workstation layout schemes, wherein the workstation layout schemes include the number of workstations, the XYZ coordinates and the attitude angle of each workstation, and selecting one of the workstation layout schemes as the scheme to be tested;

[0038] (2) In an offline simulation environment, a measuring rod 13 is mounted at the end of a robotic arm 11, and a 3D measuring device is placed 8 at a station in a solution to be tested;

[0039] (3) maintaining the point constraint of the center 12 of the measuring ball at the end of the measuring rod, uniformly distributing the attitude angle of the measuring rod in the measuring space of the 3D measuring device, solving the corresponding joint angles of the manipulator based on the inverse kinematics of the manipulator, and then screening them through configuration screening, and finally obtaining a large number of feasible and reasonable joint angles of the manipulator, wherein the configuration screening includes simulation collision detection;

[0040] (4) Place the 3D measurement device on other workstations in the test solution and repeat step (3) to finally obtain all feasible and reasonable robot arm joint angles;

[0041] (5) N groups of robot arm joint angles used in calibration measurement are selected from all feasible and reasonable robot arm joint angles. These N groups of robot arm joint angles are recorded as optimal measurement joint angle combinations. The optimal measurement joint angle combination is selected based on the observability index, and the observability index of the optimal measurement joint angle combination is obtained, which is recorded as the optimal observability index;

[0042] (6) Select other workstation layout schemes as the schemes to be tested, repeat steps (2) to (5), obtain the optimal observability index corresponding to different workstation layout schemes, and determine the optimal workstation layout scheme using the optimal observability index as the evaluation index.

[0043] The measurement space of the 3D measuring device refers to defining the intersection of the axes of three orthogonally placed displacement sensors as O1, using point O1 as three reference surfaces parallel to the measurement planes of the three sensor probes, and the three reference surfaces dividing the spherical space with the measurement point as the center into eight equal parts, and taking one-eighth of the open area facing outwards of the three displacement sensors as the measurement space.

[0044] The specific method of measuring the attitude angle of the ball bar in a uniformly distributed 3D measurement space of the displacement measuring device is as follows: the coordinates on the spherical surface are defined by the longitude and latitude in geography according to a certain division value, and the direction of the line connecting the intersection of each longitude and latitude line and the center of the sphere is the z-axis direction of the measuring spherical coordinate system of the end of the robot arm in the measured posture, and the direction of the x-axis is obtained by rotating an initial x-axis around the z-axis by an angle γ, and the angle γ is divided into -180° to 180° by a division value d γ By traversing and determining the z-axis and the x-axis, the entire end measurement spherical coordinate system can be determined;

[0045] Alternatively, instead of using longitude and latitude to define the coordinate points on the sphere, the points can be evenly distributed on the sphere, and the direction of the line connecting the evenly distributed points and the center of the sphere is defined as the z-axis direction of the spherical coordinate system measured by the end of the manipulator. The direction of the x-axis is obtained by rotating an initial x-axis around the z-axis by an angle γ, which is divided into a scale value d from -180° to 180°. γ By traversing and determining the z-axis and the x-axis, the entire end measurement spherical coordinate system can be determined.

[0046] The simulation collision detection includes static collision detection. The static collision detection is that after the robot arm moves to a certain joint angle, it will collide with itself or collide with the outside world when it is statically placed. In this case, the joint angle will be eliminated. Through the collision detection judgment in the simulation, the self-collision refers to the collision between the joint links of the robot arm, and the collision with the outside world refers to the collision between the robot arm and the ground, base, and measuring device.

[0047] The observability index can characterize the degree of pathological condition of the Jacobian matrix when the kinematic parameters are identified based on the optimization method. Specifically, it is represented by the optimal observability index. There are multiple evaluation methods for the observability index. The larger the value of the optimal observability index, the better the flexibility of the robot arm under this set of joint angles.

[0048] The measurement joint angle selection adopts the DETMAX algorithm, and the specific steps are as follows:

[0049] a. Randomly select m groups of data from all feasible and reasonable joint angle data, and calculate the observation index values ​​corresponding to these m groups of data;

[0050] b. Find a set of data from the remaining data so that the observed index value corresponding to the (m+1) set of data after adding the set of data to the originally selected m sets of data is the largest;

[0051] c. Traverse and select a group of data from the (m+1) groups of data so that the observed index value of the m groups of data after deleting the group of data is the largest;

[0052] d. Continue looping in this way until the group of data deleted is exactly the group of data added, thereby obtaining the m groups of data with the largest observable index value and realizing the selection of the optimal measurement joint angle combination.

[0053] like Figure 1 As shown, each workstation arrangement includes at least three workstations, and each workstation is arranged around the robot arm.

[0054] Grid coordinate points are selected at equal intervals in the workspace of the robot arm, and all workstation layout plans require that the workstations be placed on the corresponding grid coordinate points.

[0055] The rotation method is used to realize the design of multiple workstation layout schemes in step (1) and the determination of the optimal workstation layout scheme in step (6). The specific process is as follows:

[0056] a1. Select an initial multi-station solution;

[0057] b1. Optimize one of the workstations as the workstation to be optimized, keep the other workstations fixed, move the workstation to be optimized to a new position near its original position, complete the design of a new workstation layout scheme, and then execute steps (2) to (5) to obtain the optimal observability index corresponding to the workstation layout scheme, and repeat the search near the original position of the workstation to be optimized until the optimal observability index corresponding to the multi-workstation scheme is optimal, thereby completing a position transformation;

[0058] c1. Repeat step b1 with the other stations as stations to be optimized, thus completing a round of position rotation;

[0059] d1. Repeat steps b1-c1 continuously and perform multiple rounds of position rotation. When the optimal observability index no longer increases or the increase is less than the threshold ε1, exit the loop and output the multi-position solution that obtains the optimal optimal observability index during the rotation process to complete the iterative optimization.

[0060] The setting of the threshold ε1 is very critical. If it is set too small, it will be difficult to jump out of the loop, which will greatly reduce the efficiency of the solution. In addition, the optimal observability index solved by the DETMAX algorithm itself has a certain fluctuation range as the initial candidate group changes, so it is not meaningful to pursue a small improvement. However, if it is set too large, the loop may be exited early due to the fluctuation of the optimal observability index, making it impossible to find the global optimal solution. After parameter adjustment testing, this embodiment sets ε1 to 0.007.

[0061] Since each position change only involves the position optimization of one workstation, the number of optimization parameters is greatly reduced, so that most of the existing mature optimization algorithms (such as search-based optimization algorithms, gradient-based optimization algorithms, and heuristic optimization algorithms, etc.) can be applied and solve the optimization problem well. Since the solution domain has been gridded in this embodiment, we use a step-by-step search method to perform local optimization from the perspective of simplification. The process is as follows: Figure 2 As shown, Figure 2 A local range of 3×3 is used, and other ranges can be selected according to the situation.

[0062] In step b1, in order to achieve the 2D position optimization of the workstation, a step-by-step search method is used to perform local optimization, and the steps are as follows:

[0063] a2. Take the grid coordinate point of one of the workstations (such as Figure 2 The yellow grid in the figure is used as the center point, and all grid points in the local range near the center point are searched, and the optimal observability index corresponding to all grid points is obtained (such as Figure 2 )

[0064] b2. If the optimal observability index of all grid points in the local range no longer increases relative to the optimal observability index of the center point, or the increase is less than the threshold ε2, then the loop is exited, and the grid point with the best optimal observability index in the local range is taken as the optimized workstation position, recorded as the second grid (such as Figure 2 Otherwise, the grid point with the best local observability index is taken as the center point of the next step and recorded as the third grid point (such as Figure 2 )

[0065] c2. If the number of searches reaches the upper limit N, the loop is exited and the third grid point (such as Figure 2 The orange grid in (a) is used as the optimized workstation position; otherwise, all grid points in the local range of the third grid point are searched, and the optimal observability index corresponding to all grid points is obtained, the number of searches is increased by 1, and the process jumps to step b2 for solution.

[0066] Among them, ε2 is the same as ε1, and its size needs to take into account both solution efficiency and solution accuracy. Moreover, compared with a single round of position rotation, the result of a single local search is more affected by the random fluctuation of the optimal observability index, so ε2 needs to be smaller. After parameter adjustment testing, this embodiment sets ε2 to 0.002. In addition, the upper limit N of the number of searches also needs to be set with caution. If N is set too small, each search will only search a smaller local range, which will inevitably reduce the probability of searching for the local optimum, and may even cause the optimal observability index of a single round of position rotation to be insufficiently improved and exit the loop early, thereby missing the global optimum. But on the other hand, this may prevent each step of position change from falling into the local optimum, and instead increase the probability of the algorithm searching for the global optimum. More importantly, reducing N can greatly reduce the amount of iterative calculations, thereby improving the efficiency of the overall algorithm.

[0067] Example 2

[0068] The same parts of this embodiment as those of embodiment 1 are not described in detail, and the differences are as follows:

[0069] In order to determine the height parameters of the workstation, a plurality of candidate values ​​are first prepared for the height parameters of the workstation. Then, in step b1, in the process of repeated search near the original position of the workstation to be optimized, the height parameters of the workstation to be optimized are set to different candidate values ​​in turn until the optimal observability index corresponding to the multi-workstation solution is optimal, thereby completing a position transformation.

[0070] For the scenario where the open measurement area (opening) of the measuring device is tilted to one side, in order to determine the opening orientation angle of the workstation, a plurality of candidate values ​​are first prepared for the opening orientation angle of the workstation. Then, in step b1, during the repeated search near the original position of the workstation to be optimized, the opening orientation angle parameters of the workstation to be optimized are set to different candidate values ​​in turn until the optimal observability index corresponding to the multi-station solution is optimal, thereby completing a position transformation.

[0071] like Figure 3 As shown in the figure, taking the height of the workstation as an example, before we optimize the 2D position of a workstation, we first fix other parameters unchanged, so that the 2D position parameters of the position and the workstation orientation parameters remain unchanged, calculate all parameters within the value range of the workstation height parameter, find the parameter with the largest optimal observability index as the initial value of the local optimal parameter, obtain the first height parameter, and then use the step-by-step search method (solve Figure 3 ) realizes the 2D position optimization of the workstation (i.e., the process of steps a2 to b2 in Example 1), and searches for the best among the three parameters on the left, right, and middle of the first height parameter to find the parameter with the largest optimal observability index as the local optimal parameter for the next 2D position local search ( Figure 3The orange parameters in the first row of the figure are shown in Figure 1. As each search proceeds, the local optimal parameters gradually change until the final optimal 2D position is determined. The station height parameter with the largest optimal observability index is output as the final optimization result ( Figure 3 ).

[0072] It should be noted that for the installation height, according to the results of small batch experiments, we found that if the installation height adopts conventional boundary conditions, it is easy to fall into the local optimum, that is, the installation height is always at a boundary value (such as the lowest value -150mm or the highest value 450mm), and will not jump from one boundary value to another. This may be because when the multi-station layout plan includes stations with the lowest and highest installation heights at the same time, it is very beneficial to improve the optimal observability index. In the process of station rotation, once the installation height of several stations falls into a boundary value, the remaining stations will immediately fall into another boundary value accordingly. This "high-high-low-low" configuration is difficult to change in the subsequent iteration process, which is very unfavorable for finding the global optimum.

[0073] Therefore, we directly connect the upper and lower boundary values ​​of the workstation height to form a closed loop, that is, the left parameter of the lowest value -150mm is the highest value 450mm, so that when searching locally, it is still possible to change this "high-high-low" configuration, for example, from "high-high-low" to "low-high-low". This solves the problem well.

[0074] like Figure 3 As shown in the second row of figures in FIG. 1 , as for the workstation orientation angle, since this embodiment only considers the rotation angle of the 3D measuring device around the plumb line, its parameter range itself is a closed loop (such as Figure 3 As shown in the second row of figures, -180° is equivalent to 180°), so there is no need to consider this problem and local optimization can be directly performed.

[0075] Example 3

[0076] The same parts of this embodiment as those of embodiment 1 are not described in detail, and the differences are as follows:

[0077] For various special-shaped workspaces, when there are obstacles in the working environment of the robot arm and it is impossible to arrange workstations, it is necessary to first mark the grid points in the obstacle area in the workspace. Then, during the process of workstation rotation optimization, in each step of local search, that is, step a2 and step c2 are added with an obstacle judgment. The workstation to be optimized can only be moved to the grid without obstacles. Only then can a new workstation layout scheme be designed. Then, steps (2)-(5) are executed to obtain the optimal observability index corresponding to the workstation layout scheme.

[0078] like Figure 4As shown, because the robot arm (such as Figure 4 The working environment is different, and there may be obstacles in the working environment. You only need to mark the grid points in the workspace in the obstacle area (such as Figure 4 The pink area in the figure) and based on engineering experience, the feasible area (such as Figure 4 Select a specified number of initial stations (such as Figure 4 Then, in the process of station rotation optimization, an obstacle judgment is added at each local search step, and each station will be gradually iteratively moved to the optimal position in the feasible area (such as Figure 4 ).

[0079] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. Ordinary technicians in the field can change, modify, replace, modify, delete some features, add features or re-combine features to form a technical solution in the above embodiments without departing from the principles and purpose of the present invention. Any simple modifications, equivalent changes and modifications made to the above embodiments based on the innovative principles of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A station layout optimization design method for a multi-station robotic arm calibration device, characterized in that: The steps include: (1) Designing multiple workstation layout schemes, wherein the workstation layout schemes include the number of workstations, the XYZ coordinates and the attitude angle of each workstation, and selecting one of the workstation layout schemes as the scheme to be tested; (2) In an offline simulation environment, the measuring rod is installed at the end of the robot arm, and the 3D measuring device is placed at a station in the solution to be tested; (3) maintaining the point constraint of the center of the measuring ball at the end of the measuring rod, uniformly distributing the attitude angle of the measuring rod in the measurement space of the 3D measuring device, solving the corresponding joint angles of the manipulator based on the inverse kinematics of the manipulator, and then screening them through configuration screening, and finally obtaining a large number of feasible and reasonable joint angles of the manipulator, wherein the configuration screening includes simulation collision detection; (4) Place the 3D measurement device on other workstations in the test solution and repeat step (3) to finally obtain all feasible and reasonable robot arm joint angles; (5) N groups of robot arm joint angles used in calibration measurement are selected from all feasible and reasonable robot arm joint angles. These N groups of robot arm joint angles are recorded as optimal measurement joint angle combinations. The optimal measurement joint angle combination is selected based on the observability index, and the observability index of the optimal measurement joint angle combination is obtained, which is recorded as the optimal observability index; (6) Select other workstation layout schemes as the schemes to be tested, repeat steps (2) to (5), obtain the optimal observability index corresponding to different workstation layout schemes, and determine the optimal workstation layout scheme using the optimal observability index as the evaluation index.

2. The station layout optimization design method of a multi-station robot arm calibration device according to claim 1 is characterized in that: Each workstation arrangement scheme includes at least three workstations, and each workstation is arranged around the robot arm.

3. The station layout optimization design method of a multi-station robot arm calibration device according to claim 1 is characterized in that: Grid coordinate points are selected at equal intervals in the workspace of the robot arm, and all workstation layout plans require that the workstations be placed on the corresponding grid coordinate points.

4. The station layout optimization design method of a multi-station robot arm calibration device according to claim 3 is characterized in that: The rotation method is used to realize the design of multiple workstation layout schemes in step (1) and the determination of the optimal workstation layout scheme in step (6). The specific process is as follows: a1. Select an initial multi-station solution; b1. Optimize one of the workstations as the workstation to be optimized, keep the other workstations fixed, move the workstation to be optimized to a new position near its original position, complete the design of a new workstation layout scheme, execute steps (2) to (5), obtain the optimal observability index corresponding to the workstation layout scheme, and repeat the search near the original position of the workstation to be optimized until the optimal observability index corresponding to the multi-workstation scheme is optimal, thereby completing a position transformation; c1. Repeat step b1 with the other stations as stations to be optimized, thus completing a round of position rotation; d1. Repeat steps b1-c1 continuously and perform multiple rounds of position rotation. When the optimal observability index no longer increases or the increase is less than the threshold ε1, exit the loop and output the multi-position solution that obtains the optimal optimal observability index during the rotation process to complete the iterative optimization.

5. The station layout optimization design method of a multi-station robot arm calibration device according to claim 4 is characterized in that: In step b1, in order to achieve the 2D position optimization of the workstation, a step-by-step search method is used to perform local optimization, and the steps are as follows: a2. Take the grid coordinate point of one of the workstations as the center point, search all the grid points in the local range near the center point, and obtain the optimal observability index corresponding to all the grid points; b2. If the optimal observability index of all grid points in the local range no longer increases relative to the optimal observability index of the center point, or the increase is less than the threshold ε2, then the loop is exited, and the grid point with the best optimal observability index in the local range is used as the optimized workstation position, otherwise the grid point with the best optimal observability index in the local range is used as the center point of the next step, recorded as the third grid point; c2. If the number of searches reaches the set upper limit N, the loop is exited and the third grid point is used as the optimized workstation position; otherwise, all grid points within the local range of the third grid point are searched, and the optimal observability index corresponding to all grid points is obtained. The number of searches is increased by 1, and the process jumps to step b2 for solution.

6. The station layout optimization design method of a multi-station robot arm calibration device according to claim 4 is characterized in that: In order to determine the height parameters of the workstation, a plurality of candidate values ​​are first prepared for the height parameters of the workstation. Then, in step b1, in the process of repeated search near the original position of the workstation to be optimized, the height parameters of the workstation to be optimized are set to different candidate values ​​in turn until the optimal observability index corresponding to the multi-workstation solution is optimal, thereby completing a position transformation.

7. The station layout optimization design method of a multi-station robot arm calibration device according to claim 5 is characterized in that: For various special-shaped workspaces, when there are obstacles in the working environment of the robot arm and it is impossible to arrange workstations, it is necessary to first mark the grid points in the obstacle area in the workspace. Then, during the process of workstation rotation optimization, in each step of local search, that is, step a2 and step c2 are added with an obstacle judgment. The workstation to be optimized can only be moved to the grid without obstacles. Only then can a new workstation layout scheme be designed. Then, steps (2)-(5) are executed to obtain the optimal observability index corresponding to the workstation layout scheme.

Images