A method for determining the coordinate transformation relationship between a robot and a measuring instrument

By dividing the quadrants in the robot's work space and selecting scattered common points, the problem of low coordinate conversion accuracy caused by random selection of common points is solved, and efficient and accurate coordinate conversion relationship calculation is achieved.

CN114754672BActive Publication Date: 2024-11-05Chinese People's Liberation Army Cyberspace Force Information Engineering University
View PDF 15 Cites 0 Cited by

Patent Information

Application Number
CN202210095265.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-01-26
Publication Date
2024-11-05
Estimated Expiration
2042-01-26

AI Technical Summary

Technical Problem

In the prior art, random selection of common points leads to low accuracy in the conversion parameters of the coordinate system of the robot and the measuring instrument, and complex operation, requiring a large amount of data and time.

Method used

By dividing the quadrants in the robot's working space, calculating the average value of the teaching points as the first center of gravity point, and selecting points with far distances as common points to ensure the dispersion of common points in the base coordinate system, and using these common points to measure The coordinate conversion relationship is calculated with the coordinates in the base coordinate system.

Benefits of technology

Improve the accuracy and stability of the coordinate conversion relationship, reduce the calculation time, and achieve fast and accurate coordinate conversion.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114754672B_ABST
    Figure CN114754672B_ABST
Patent Text Reader

Abstract

The present invention relates to a method for determining the coordinate conversion relationship between a robot and a measuring instrument, and belongs to the field of digital measurement. A measuring instrument is arranged in the robot workspace, and the coordinate system of the robot mounting base is defined as the base coordinate system, and the coordinate system of the measuring instrument is defined as the measuring coordinate system. Taking any two axes in the base coordinate as a reference, the robot working space is divided into four quadrants, and the average value of the coordinates of each teaching point is calculated, and the coordinate point corresponding to the average value is recorded as the first center of gravity point, and the teaching point far from the first center of gravity point is selected as a common point, but the coordinate values ​​of two adjacent common points in another axis are different. The conversion relationship between the measuring coordinate system and the base coordinate system can be calculated using the coordinates of the four common points in the measuring coordinate system and the base coordinate system respectively. The present invention only needs to select four common points with reasonable spatial distribution in the teaching point to accurately calculate the coordinate conversion relationship, and has strong stability, high precision and high efficiency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a method for determining the coordinate conversion relationship between a robot and a measuring instrument, and belongs to the field of digital measurement. Background Art

[0002] In recent years, industrial robots, as the core equipment of modern manufacturing industry, have become the core force to promote the development of manufacturing industry with their strong stability and high reliability. The development and application of robot technology is of great significance to the national economic construction and national defense construction. The repeatability of industrial robots is very high, but the determination of positioning accuracy is poor. The absolute positioning accuracy before calibration is usually 2-3mm, which is difficult to meet the needs of high-end industries such as precision parts processing, spacecraft assembly, and weapons and equipment production. Calibration of the kinematic parameters of industrial robots is the mainstream method to improve their absolute accuracy. The calibration of robot kinematic parameters can be divided into four stages: modeling, measurement, parameter identification and error compensation. In the stage of measuring the end position error of the robot, a laser tracker is often used to obtain the end position error of the robot. Before error measurement, the problem of unifying the measurement coordinate system of the laser tracker and the base coordinate system of the industrial robot must be solved.

[0003] In order to unify the coordinate system of the laser tracker and the base coordinate system of the industrial robot, the existing technology uses a method that combines the robot's single-axis motion with the reading of the teaching pendant, and uses the geometric relationship of the robot's motion trajectory to solve the conversion parameters. However, this method is complicated to operate and requires the collection of a large amount of data to ensure the accuracy of the solution. The whole process takes a long time. Therefore, the existing technology uses a method based on distance constraints to calibrate the tool coordinate system, and then solves the coordinate conversion parameters through the least squares iteration method based on the Rodriguez matrix. This method can significantly improve the efficiency and quality of solving the conversion parameters. In addition, the least squares fitting algorithm based on singular value decomposition can also be used to solve the conversion parameters.

[0004] No matter which of the above methods is used, the problem of selecting common points is involved. At present, the random selection method is often used to select common points. The random selection method is prone to the situation that the selected common points are densely distributed in the spatial geometry, resulting in some areas of the entire robot working space having common points and some areas having no common points, which makes the solution parameters unstable and unreliable, and the accuracy of the coordinate system transformation parameters solved is low. Summary of the invention

[0005] The purpose of the present invention is to provide a method for determining the coordinate transformation relationship between a robot and a measuring instrument, so as to solve the problem of low accuracy of coordinate system transformation parameters solved by randomly selecting common points.

[0006] To achieve the above objectives, the technical solutions and corresponding beneficial effects of the present invention include:

[0007] A method for determining a coordinate transformation relationship between a robot and a measuring instrument of the present invention comprises the following steps:

[0008] 1) Control the robot to teach n positions in the robot working space corresponding to n teaching points, where n is teaching;

[0009] 2) Define the coordinate system of the robot installation base as the base coordinate system, and divide the robot working space into 4 quadrants based on any two axes of the base coordinate system; calculate the average value of the coordinates of n teaching points, and the coordinate point corresponding to the average value is the first center of gravity; select a teaching point far from the first center of gravity in each quadrant as a common point, respectively as the first common point, the second common point, the third common point and the fourth common point, and the coordinate value of another axis of two common points in adjacent quadrants in the base coordinate system is different;

[0010] 3) Define the coordinate system of the measuring instrument as the measuring coordinate system, obtain the coordinates of the common point in the measuring coordinate system through the measuring instrument, and calculate the coordinate transformation relationship between the robot base coordinate system and the measuring coordinate system according to the coordinates of the selected common point in the measuring coordinate system and the coordinates of the common point in the base coordinate system.

[0011] The beneficial effect of the above technical solution is: in order to realize the coordinate conversion relationship determination method of a robot and a measuring instrument of the present invention, a measuring instrument is set at a set position in the robot workspace, the robot is installed on a base, the coordinate system corresponding to the robot installation base is the base coordinate system, and the coordinate system corresponding to the measuring instrument is the measurement coordinate system. When implementing this method, the robot working space is divided into 4 quadrants based on any two axes in the base coordinate system, and each teaching point is distributed in the 4 quadrants, and then the average value of the coordinates of each teaching point is calculated, and the coordinate point corresponding to the average value in the 4 quadrants of the robot working space is recorded as the first center of gravity point, and the teaching point far from the first center of gravity point in the 4 quadrants is selected as a common point, but the coordinate values ​​of two adjacent common points in another axis are different, so as to ensure the dispersion of the selected 4 common points in the robot working space. Then, the conversion relationship between the measurement coordinate system and the base coordinate system can be calculated by using the coordinates of the 4 common points in the measurement coordinate system and the base coordinate system, that is, the coordinate conversion relationship between the robot and the measuring instrument is obtained.

[0012] The present invention takes into account the influence of the spatial geometric distribution of common points on the accuracy of solving the transformation relationship. By using the present invention, it is only necessary to select 4 common points with a reasonable spatial distribution in the teaching points in the robot's working space, so as to quickly and accurately calculate the coordinate transformation relationship with strong stability, high accuracy and high efficiency.

[0013] Furthermore, the common points selected in step 3) also include a fifth common point, and the fifth common point is selected by the following method: two common points in adjacent quadrants are selected from the first common point, the second common point, the third common point and the fourth common point, and the average value of the coordinates corresponding to the two common points is calculated, and the coordinate point corresponding to the average value is the second center of gravity point. Among all teaching points except the first common point, the second common point, the third common point and the fourth common point, the teaching point closest to the second center of gravity point is selected as the fifth common point.

[0014] The beneficial effect of the above technical solution is: in addition to the four common points selected from each teaching point, a fifth common point is also selected, and considering the dispersion of the distribution of all common points, the selected fifth common point is a teaching point close to the center position of two of the common points, and then the five common points are used to participate in the calculation to obtain the coordinate transformation parameters, thereby improving the accuracy of the coordinate transformation relationship calculation.

[0015] Furthermore, the common points selected in step 3) also include a sixth common point, and the sixth common point is selected by the following method: selecting the remaining two common points from the first common point, the second common point, the third common point and the fourth common point, and calculating the average value of the coordinates corresponding to the remaining two common points, the coordinate point corresponding to the average value is the third center of gravity point, and among all teaching points except the first common point, the second common point, the third common point, the fourth common point and the fifth common point, the teaching point closest to the third center of gravity point is selected as the sixth common point.

[0016] The beneficial effects of the above technical solution are: in addition, 6 common points are selected from each teaching point, and considering the dispersion of the distribution of all common points, the sixth common point is selected as the teaching point close to the center position of the remaining two common points, and then the 6 common points are used to participate in the calculation to obtain the coordinate transformation parameters. Without increasing the workload too much, the spatial set strength of the common points can be improved, thereby improving the solution accuracy of the coordinate transformation calculation.

[0017] Further, in step 2), any two axes of the base coordinate system are the X-axis and the Y-axis of the base coordinate system, and the other axis is the Z-axis of the base coordinate system.

[0018] Further, in step 2), the coordinates of the n teaching points are the coordinates of the n teaching points in the measurement coordinate system.

[0019] The beneficial effect of the above technical solution is that the coordinates of the teaching point in the measuring coordinate system can be directly obtained by the measuring instrument. Therefore, the present invention uses the coordinates of the teaching point in the measuring coordinate system in the calculation of various distances and center of gravity points in the selection of common points, ensuring the rapidity of the calculation of the entire process.

[0020] Furthermore, a flange is provided at the end of the robot, and a target ball is installed at the end of the robot flange. The coordinate system of the robot flange end is defined as the flange end coordinate system, the coordinate system of the target ball is defined as the tool coordinate system, and the center of the target ball is defined as the origin of the tool coordinate system; the coordinates of the common point in step 3) in the base coordinate system are obtained by the following method:

[0021] ①According to the robot forward kinematics, the coordinate transformation relationship between the flange end coordinate system and the base coordinate system is obtained;

[0022] ② Obtain the teaching pendant readings of n teaching points, select n1 teaching pendant readings from the n teaching pendant readings, the teaching pendant readings represent the coordinate transformation relationship between the flange end coordinate system and the base coordinate system, where n1 ≥ 3; then obtain the coordinates of the n1 teaching points in the measurement coordinate system, and use the following formula to calculate the position of the target sphere center in the flange end coordinate system:

[0023]

[0024] in, is the teaching pendant reading of two teaching points located at positions j and k among the n1 teaching points. are the coordinates of the two teaching points in the measurement coordinate system, is the coordinate of the target sphere center in the flange end coordinate system;

[0025] ③ According to the teaching pendant readings of n teaching points and the position of the target sphere center in the flange end coordinate system, determine the coordinates of each teaching point in the base coordinate system.

[0026] The beneficial effect of the above technical solution is that the present invention takes into account the actual structural position relationship of the robot, and uses the connection and spatial relationship between the target ball, flange, and robot base to accurately calculate the position of each teaching point in the base coordinate system.

[0027] Furthermore, in step ②, the position of the target sphere center in the flange end coordinate system is calculated using the least squares method.

[0028] Furthermore, n1=3.

[0029] The beneficial effect of the above technical solution is: selecting 3 teaching points from all teaching points to obtain the teaching pendant readings for calculation can greatly reduce the workload.

[0030] Furthermore, the measuring instrument is a laser tracker.

[0031] The beneficial effects of the above technical solution are: the target position measured by the laser tracker is directly the position in the measurement coordinate system corresponding to the laser tracker, and the measurement accuracy of the laser tracker is relatively high, so the laser tracker is selected as the measuring instrument. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the test system model adopted by the present invention;

[0033] Figure 2 It is a schematic diagram of the base coordinate system of the industrial robot of the present invention;

[0034] Figure 3 It is a schematic diagram of the quadrant division of the present invention;

[0035] Figure 4 It is the teaching point distribution diagram of the present invention;

[0036] Figure 5 It is the quadrant distribution combination point difference distribution diagram of the present invention;

[0037] Figure 6 It is a random combination of one point difference distribution diagram of the present invention;

[0038] Figure 7 This is a random combination of two points with poor distribution. DETAILED DESCRIPTION

[0039] The present invention is further described in detail below with reference to the accompanying drawings and embodiments.

[0040] Example:

[0041] The present invention provides a method for determining the coordinate transformation relationship between a robot and a measuring instrument. The hardware used in the method is as follows: Figure 1 As shown, it mainly includes a laser tracker and its supporting measuring equipment, an industrial robot and auxiliary tooling. The method involves multiple spatial rectangular coordinate systems. The coordinate system defined on the laser tracker is the laser tracker measurement coordinate system, represented by {M}; the coordinate system defined on the installation base of the industrial robot is the robot base coordinate system, represented by {B}, and the coordinate system always remains stable during the movement of the robot; the coordinate system defined on the flange end of the industrial robot is the flange end coordinate system, represented by {E}, and the coordinate system is the reference datum of the tool coordinate system; the coordinate system defined on the tool installed on the flange end of the robot is the tool coordinate system, represented by {T}, and the origin of the coordinate system is the tool center point (TCP), and the TCP is fixed relative to the flange end coordinate system {E}, and the robot uses the TCP as a reference point to plan the motion trajectory. In the present invention, the target ball center P is defined as the tool center point, and the target ball is fixedly connected to the flange end of the industrial robot through the target ball base.

[0042] After determining the various coordinate systems of the robot, a method for determining the coordinate conversion relationship between a robot and a measuring instrument of the present invention can be implemented. The overall idea of ​​the method is: the robot is controlled to carry a target ball in the robot workspace to teach several different points as teaching points, and several common points are selected from the teaching points, and the common points are used to calculate the coordinate conversion parameters of the measurement coordinate system and the base coordinate system. The following is a detailed introduction.

[0043] Step 1: Control the robot to carry the target ball and teach several different points in the robot workspace as teaching points.

[0044] Step 2: Select 6 common points for determining the coordinate transformation parameters between the measurement coordinate system and the base coordinate system.

[0045] The accuracy of solving the coordinate system transformation parameters is closely related to the selection of common points. The main factors affecting the accuracy of solving the transformation parameters are the accuracy, quantity and spatial distribution of the common points.

[0046] Among them, the accuracy of coordinate transformation mainly depends on the measurement accuracy of the common points involved in the calculation of coordinate transformation parameters, that is, the accuracy of the three-dimensional coordinates of the target ball center P after each movement of the industrial robot obtained by the robot teaching pendant and the laser tracker respectively. The model of the laser tracker used in the present invention is Leica AT901-B. The nominal measurement accuracy of three-dimensional point coordinates is ±(15μm+6μm / m). A RRR1.5 precision target ball with a diameter of 38.1mm is used as a reflective prism. With the target ball, the point position accuracy is better than ±75μm within a working range of 10m. The robot used is an ABB IRB-4600 industrial robot with a payload of 40kg and a nominal repeatability accuracy of ±0.06mm. It is strictly calibrated before leaving the factory, and its absolute positioning accuracy is about ±1mm. During the entire operation, the laser tracker is placed about 1.5 meters outside the working space of the industrial robot. Figure 1 As shown in the figure. The point position accuracy measured by the laser tracker is better than ±75μm, which is much higher than the absolute positioning accuracy of the industrial robot. Therefore, the quality of the coordinate transformation result mainly depends on the absolute positioning accuracy of the industrial robot.

[0047] The number of common points is also one of the factors that affect the coordinate system conversion parameters. To solve the conversion parameters of two spatial rectangular coordinate systems, at least three common points are required. When the number of common points is too small, the position error of the common points will have a greater impact on the conversion result, which can easily cause the solution result to be unreliable. The number of common points is generally 3 to 6, and should not exceed 6, because when the number of common points is too large, the calculation workload will be increased, and the accuracy of solving the conversion parameters will not be significantly improved.

[0048] In engineering practice, the spatial distribution of common points is relatively complex. It is possible that the spatial distribution of common points is relatively dense, and the distance between common points and conversion points is relatively far. Generally, the accuracy of coordinate system conversion parameter solution is positively correlated with the spatial distribution geometric strength of the common points involved in the calculation. When the accuracy and number of common points are the same, when the spatial distribution of common points is relatively dense, it is easy to cause the pathological nature of the normal equation in the conversion parameter solution process. The more discrete the distribution of common points, the better the spatial geometric strength and the higher the solution accuracy. Conversely, the lower the solution accuracy.

[0049] The method of selecting common points based on quadrant distribution adopted by the present invention is to divide the robot workspace into four quadrants along the X-axis and Y-axis of the robot base coordinate system {B}, such as Figure 2 , Figure 3 As shown. The industrial robot is installed on the load-bearing surface of the automatic guided vehicle (AGV), and the AGV can carry the industrial robot to achieve displacement in space. During the experiment, the AGV deployed four hydraulic legs to increase the friction between the platform and the ground to ensure the stability of the platform.

[0050] The method for selecting common points based on quadrant distribution mainly considers the impact of the spatial geometric distribution of common points on the conversion results. According to the distribution of common points in different quadrants in the actual working space of the industrial robot, a group of common points with relatively discrete distribution are selected to participate in the calculation of coordinate transformation parameters to improve the accuracy of solving the transformation parameters. In order to improve the accuracy of solving the coordinate transformation parameters, the present invention takes into account the difficulty of operation and data processing, and selects 6 common points as the starting data for solving. The controller of the robot used in the present invention has a certain degree of movement restriction on the robot's working space, and controls the industrial robot to teach 22 positions in the safe area of ​​its working space, such as Figure 4 As shown in the figure, if the robot controller has no motion restrictions on the robot's working space, 24 positions can be taught symmetrically and evenly. Six points are selected from these 22 teaching points as common points for solving coordinate transformation parameters, and the remaining teaching points that are not involved in the calculation of transformation parameters are used as transformation points to verify the accuracy of coordinate transformation.

[0051] The method for selecting common points based on quadrant distribution adopted by the present invention comprises the following steps:

[0052] a) Centroid all teaching points. Since the measurement accuracy of the laser tracker is significantly higher than the absolute positioning accuracy of the robot, the coordinates of the teaching points measured by the laser tracker are selected for centroiding. That is, the average value of the X, Y, and Z coordinate components of all teaching points in the laser tracker measurement coordinate system is obtained, and the first centroid point is recorded as G0.

[0053] b) Find the distance from the teaching point to the first center of gravity G0. Calculate the distance between each teaching point and the first center of gravity G0 obtained after centroiding.

[0054] c) Select a point farthest from the first centroid G0 in each quadrant I, II, III, and IV as the common point. When selecting the common point, the common points in adjacent quadrants should be at different heights on the Z axis. It should be noted that, in theory, it is best to select the point farthest from the first centroid G0 in each quadrant as the common point. However, in order to meet the condition that "the common points in adjacent quadrants have different heights on the Z axis", it is possible to select the point in a quadrant that is the second farthest from the first centroid G0 as the common point. Combined with Figure 4 As shown, the common point selected in quadrant I is P17, the common point selected in quadrant II is P13, the common point selected in quadrant III is P7, and the common point selected in quadrant IV is P3.

[0055] d) Take the average value of the coordinates of the two common points selected in quadrants Ⅰ and Ⅱ and record it as the second center of gravity point G1. That is, calculate the average value of the coordinates of P17 and P13, and the point corresponding to the calculated average value is the second center of gravity point G1. Select the point closest to the second center of gravity point G1 from the remaining teaching points except P17, P13, P7 and P13, which is P15.

[0056] e) Take the average value of the coordinates of the two common points selected in quadrants III and IV and record it as the third center of gravity point G2. That is, calculate the average value of the coordinates of P7 and P3, and the point corresponding to the calculated average value is the third center of gravity point G2. Select the point closest to the third center of gravity point G2 from the remaining teaching points except P17, P13, P7, P13 and P15, which is P5.

[0057] The common points selected by the common point selection method based on quadrant distribution of the present invention are P3, P5, P7, P13, P15, and P17, and this group of common points is recorded as a quadrant distribution combination. The conversion parameters of the laser tracker measurement coordinate system and the base coordinate system are calculated using the quadrant distribution combination.

[0058] Step 3: Use the coordinates of the selected 6 common points in the measurement coordinate system and the base coordinate system to calculate the coordinate transformation relationship between the robot base coordinate system and the measurement coordinate system.

[0059] Among them, the robot cannot directly obtain the coordinates of each common point in the base coordinate system. What can be directly obtained is the conversion relationship between the flange end coordinate system and the base coordinate system. At this time, a series of coordinate conversion and processing are required to obtain it. The specific conversion process is as follows.

[0060] During the calculation process, the relative position relationship between the laser tracker and the industrial robot base remains stable, that is, the relative position of {M} and {B} remains stable. The robot is controlled to carry the target ball to teach several different points in the robot workspace as teaching points, and the teaching pendant readings at each point (the teaching pendant readings represent the coordinate conversion relationship between the flange end and the base coordinate system) and the laser tracker measurement readings are recorded.

[0061] From the existing technologies such as the structure of industrial robots and the definition of the coordinate systems of each link, it can be known that the relationship between the flange end coordinate system {E} and the base coordinate system {B} of the industrial robot can be expressed by the homogeneous transformation matrix express:

[0062]

[0063] in, is the rotation matrix of the flange end coordinate system {E} relative to the base coordinate system {B}, which consists of three direction vectors and They respectively represent the direction cosines of the three principal vectors of the flange end coordinate system {E} relative to the base coordinate system {B}, and are used to describe the posture of the flange end coordinate system {E} in the base coordinate system {B}. is the translation matrix of the flange end coordinate system {E} relative to the base coordinate system {B}, and its value represents the position of the origin of the flange end coordinate system in the base coordinate system. The above values ​​can be obtained by the robot forward kinematics. Based on the above calculation, the coordinate position relationship between the flange end coordinate system {E} and the base coordinate system {B} is obtained.

[0064] The tool installed at the end of the robot flange is installed on the flange. Therefore, the tool coordinate system {T} corresponding to the tool installed at the end of the flange and the flange end coordinate system {E} corresponding to the flange end can also be transformed using the homogeneous transformation matrix express.

[0065] The coordinate position relationship between the flange end coordinate system {E} and the base coordinate system {B} is calculated, and the coordinate position relationship between the tool coordinate system {T} and the flange end coordinate system {E} is also calculated. Therefore, the coordinate position relationship between the tool coordinate system {T} and the base coordinate system {B} can be calculated using the coordinate transformation principle.

[0066] Specifically, the relationship between the flange end coordinate system {E} and the base coordinate system {B} is expressed by the homogeneous transformation matrix The transformation relationship between the tool coordinate system {T} and the flange end coordinate system {E} is expressed by the homogeneous transformation matrix The transformation relationship between the tool coordinate system {T} and the robot base coordinate system {B} is expressed as express.

[0067] According to the coordinate transformation principle, the above coordinate system satisfies:

[0068]

[0069] Right now:

[0070]

[0071] Expand to:

[0072]

[0073]

[0074] After calculating the coordinate position relationship between the tool coordinate system {T} and the base coordinate system {B}, firstly use the TCP calibration method based on distance constraint to obtain TCP, that is, the position of the target ball center P in the flange end coordinate system {E} (that is, obtain the position of the target ball center in the tool coordinate system {T} in the flange end coordinate system {E}), and then use the coordinate position relationship between the flange end coordinate system {E} and the base coordinate system {B} to obtain the position of TCP in the robot base coordinate {B}. The present invention takes into account the influence of the spatial geometric distribution of common points on the accuracy of solving the conversion parameters, adopts the common point selection method based on quadrant distribution to select common points, and then obtains the conversion parameters, uses the parameters not involved in the coordinate conversion to calculate the remaining teaching points for coordinate conversion, and verifies the quality of the coordinate conversion results.

[0075] However, before coordinate transformation, TCP needs to be calibrated, that is, the target sphere center position is calibrated. Because TCP is stable relative to the flange end coordinate system {E}, TCP calibration is to obtain the position of the target sphere center P in the flange end coordinate system {E}.

[0076] According to the coordinate transformation principle, the coordinates of the target ball center P in the robot base coordinate system {B} can be expressed as:

[0077]

[0078] in, It can be obtained from the industrial robot teaching pendant; i is the number of positions reached by TCP in the robot workspace. Take any two different positions j and k of TCP in the robot workspace, and the positions of the target ball center in the base coordinate system {B} at these two positions are expressed as The coordinate deviation in the base coordinate system {B} is:

[0079]

[0080] For any two points in space, their coordinate values ​​may be different in different coordinate systems of the same scale, but the distance between the two points is the same in different coordinate systems of the same scale. Therefore, at positions j and k, the distance between the target sphere centers P is the same in the robot base coordinate system {B} and the laser tracker measurement coordinate system {M}. That is:

[0081]

[0082] Will Substitution have to:

[0083]

[0084] It can be calculated from the parameters read by the robot teaching pendant; It can be directly measured by a laser tracker; is the unknown quantity to be determined. It is obtained by the following method:

[0085]

[0086] Further expansion leads to the nonlinear equation:

[0087]

[0088] The nonlinear equation is linearized by first-order Taylor series expansion, and the error equation is obtained as follows:

[0089]

[0090] in,

[0091]

[0092]

[0093]

[0094]

[0095] From the above, we can see that the unknown quantity There are three unknowns (x, y, z), so at least three equations need to be established to solve. Assuming that the robot teaches r points in its workspace, r(r-1) / 2 equations can be established, so three teaching points are required to establish three equations. In order to improve the accuracy of solving the unknowns, more than three teaching points are usually selected and solved using the least squares principle. The unknowns (x, y, z) are obtained. At this point, the TCP position calibration is completed. As another implementation method, other methods such as maximum likelihood estimation can also be used to solve unknowns (x, y, z).

[0096] The calibration of TCP position is to calculate the position of the target ball center P in the flange end coordinate system {E}. After determining the position of the target ball center P in the flange end coordinate system {E}, the position of the target ball center P in the base coordinate system {B} can be calculated by using the coordinate position relationship between the flange end coordinate system {E} and the base coordinate system {B}. Each teaching point is a different point taught by the industrial robot in its working space, that is, it is to control the industrial robot to mark the target ball center P at different positions in its working space. Therefore, after determining the position of the target ball center P in the base coordinate system {B}, the position of each teaching point in the base coordinate system {B} can be determined. At the same time, the tracking measuring instrument can also be used to obtain the position of each teaching point in the laser tracker measurement coordinate system. Therefore, the same teaching point has a position relationship in the base coordinate {B} and also has a position relationship in the laser tracker measurement coordinate system. The teaching point can be used to calculate the conversion parameters between the laser tracker measurement coordinate system and the base coordinate system.

[0097] The essence of coordinate transformation is to obtain the relative posture relationship between the laser tracker measurement coordinate system and the robot base coordinate system. The present invention ensures that the relative posture of the laser tracker and the robot base remains stable during the entire test process. Before the test, the laser tracker is used to test the stability of the AGV platform under different operating conditions of the industrial robot, and its stability meets the requirements of the coordinate transformation test. During the test, the position of the target sphere center P in the robot flange end coordinate system {E} is first calculated by the TCP calibration principle; then the coordinate system is transformed through three groups of common points, and the least squares method is used to solve the transformation parameters. Finally, the teaching points that are not involved in the calculation of the transformation parameters are used for transformation, and the accuracy of the transformation results is analyzed.

[0098] In order to verify the accuracy of the present invention, two groups of teaching points with different spatial distribution from the above common points are randomly selected as common point random combination one P3, P5, P7, P9, P10, P19) and random combination two (P1, P13, P15, P17, P20, P21) for comparison. After solving the conversion parameters using each common point combination, in order to show its accuracy, the remaining teaching points (P2, P4, P6, P8, P11, P12, P14, P16, P18, P22) that are not involved in the calculation of the conversion parameters are used as conversion points for coordinate conversion to verify the accuracy of the conversion result.

[0099] The purpose of using TCP calibration is to obtain the position of the target ball center in the industrial robot flange end coordinate system {E}. When solving the TCP position, at least three teaching points in the robot workspace need to be measured. In order to further improve the TCP solution accuracy and the efficiency of solving the conversion parameters, the common points for solving the conversion parameters are used as the data for TCP solution, and the TCP position is solved using the TCP calibration method in step 1). The laser tracker measurement coordinates of the three sets of common points and the corresponding robot teaching pendant readings are shown in Tables 1, 2, and 3, respectively. In the table, A, B, and C represent the rotation angles of the flange end coordinate system around the Z axis, Y axis, and X axis of the robot base coordinate system, respectively.

[0100] Table 1 Coordinates of laser tracker measurements and teaching pendant readings at each point in the quadrant distribution combination

[0101]

[0102] Table 2 Coordinates measured by laser tracker and readings of teaching pendant at each point in random combination 1

[0103]

[0104]

[0105] Table 3 Coordinates measured by laser tracker and readings of teaching pendant at each point in random combination 2

[0106]

[0107] The TCP is solved based on the least squares principle. The position of the TCP in the robot flange end coordinate system {E} solved by each point in the quadrant distribution combination is (0.94, -0.53, 38.60), the position of the TCP in the robot flange end coordinate system {E} solved by each point in the random combination 1 is (-0.56, -1.80, 39.87), and the position of the TCP in the robot flange end coordinate system {E} solved by each point in the random combination 2 is (0.97, -0.35, 38.10), the unit is mm.

[0108] Based on TCP calibration, The coordinates of TCP (target ball center P) in the robot base coordinate system {B} can be further obtained. The calculated common point quadrant distribution combinations correspond to the coordinates of TCP at each teaching point in the robot base coordinate system {B} in the experiment as shown in Table 4.

[0109] Table 4 Coordinates of each point in the quadrant distribution combination in the base coordinate system

[0110]

[0111]

[0112] The coordinate system is transformed using the least squares principle, and the homogeneous transformation matrix of the robot base coordinate system relative to the laser tracker coordinate system is obtained by the corresponding experiment of the common point quadrant distribution combination. for:

[0113]

[0114] Depend on The coordinates of the TCP at each teaching point in the robot base coordinate system {B} in the corresponding experiment calculated from the random combination of common points are shown in Table 5.

[0115] Table 5 Coordinates of points in random combination 1 in the base coordinate system

[0116]

[0117] The homogeneous transformation matrix of the robot base coordinate system relative to the laser tracker coordinate system is obtained by a random combination of common points and a corresponding experiment. for:

[0118]

[0119] Depend on The calculated coordinates of TCP (target sphere center P) at each teaching point in the common point random combination two correspondence experiments in the robot base coordinate system {B} are shown in Table 6.

[0120] Table 6 Coordinates of each point in the base coordinate system in random combination 2

[0121]

[0122]

[0123] The homogeneous transformation matrix of the robot base coordinate system relative to the laser tracker coordinate system is obtained by solving the common point random combination two correspondence experiments for:

[0124]

[0125] The present invention uses the point position error after conversion of the conversion point and the average value of the absolute value of the conversion difference in the X, Y and Z directions as indicators to measure the feasibility of the common point selection method based on quadrant distribution.

[0126] Take the conversion experiment corresponding to the common point quadrant distribution combination as an example. First, in the TCP calibration result Based on the solution of , we can further find the coordinates of the transformation point in the base coordinate system {B}. Secondly, we use the transformation matrix The coordinates of the conversion point in the base coordinate system {B} are converted to the coordinates in the tracker coordinate system. Finally, the coordinates directly measured by the laser tracker are subtracted from the coordinates obtained by the coordinate conversion to obtain the point position difference after the conversion. In the conversion experiment corresponding to the common point quadrant distribution combination, the coordinates of the conversion point in the base coordinate system, the converted coordinates, and the coordinates measured by the laser tracker are shown in Table 7:

[0127] Table 7 Common point quadrant distribution combinations corresponding to the conversion point coordinates in the experiment

[0128]

[0129]

[0130] Similarly, randomly combine a solution at a common point And the common point random combination two solutions On this basis, the coordinates of the conversion point in the base coordinate system, the coordinates after conversion, and the coordinates measured by the laser tracker are obtained as shown in Table 8 and Table 9 respectively.

[0131] Table 8 Coordinates of transformation points in the random combination of common points and corresponding experiments

[0132]

[0133] Table 9 Coordinates of transformation points in the random combination of common points and the corresponding experiment

[0134]

[0135]

[0136] The coordinates of the conversion point obtained after coordinate conversion in the laser tracker measurement coordinate system are subtracted from the coordinates of the conversion point directly measured by the laser tracker to obtain the conversion point differences of the three groups of common point combination corresponding experiments. The differences in the X, Y, and Z directions and the average values ​​of the absolute values ​​of the differences in each direction in the three groups of common point combination corresponding experiments are shown in Table 10. The distribution of point differences is shown in Figure 5 , Figure 6 , Figure 7 shown.

[0137] Table 10 The conversion points of the common point combination experiments are poor

[0138]

[0139] The conversion accuracy of the X, Y and Z coordinate components of the conversion points of the three groups of common point combination experiments and the point position accuracy of the coordinate conversion are further statistically analyzed, and the statistical results are shown in Table 11.

[0140] Table 11 Coordinate transformation accuracy statistics

[0141]

[0142]

[0143] As shown in Tables 10 and 11, the average values ​​of the absolute values ​​of the conversion errors at the conversion points in the experiments corresponding to the common point quadrant distribution combinations are 0.45mm, 0.49mm, and 0.42mm, respectively. Compared with the average values ​​of the absolute values ​​of the conversion errors at the conversion points in the X, Y, and Z directions of 0.53mm, 0.54mm, and 0.80mm in the experiments corresponding to the random combination 1 and the average values ​​of the absolute values ​​of the conversion errors at the conversion points in the X, Y, and Z directions of 0.82mm, 0.67mm, and 0.68mm in the experiments corresponding to the random combination 2, the conversion accuracy is better, and the values ​​in the three directions of X, Y, and Z are more balanced. Figure 5 , Figure 6 , Figure 7 It can also be seen that the poor distribution of conversion points of the common point quadrant distribution combination is more concentrated than that of random combination 1 and random combination 2, that is, the stability of coordinate transformation is stronger.

[0144] In the corresponding experiment of the common point quadrant distribution combination, the RMS errors of each conversion point on the X, Y and Z coordinate components are 0.52mm, 0.58mm, and 0.50mm, respectively, and the comprehensive RMS error is 0.93mm. In the corresponding experiment of the common point random combination 1, the RMS errors of each conversion point on the X, Y and Z coordinate components are 0.60mm, 0.70mm, and 0.92mm, respectively, and the comprehensive RMS error is 1.30mm. In the corresponding experiment of the common point random combination 2, the RMS errors of each conversion point on the X, Y and Z coordinate components are 1.02mm, 0.92mm, and 1.79mm, respectively, and the comprehensive RMS error is 2.25mm. The comprehensive RMS error of the conversion point in the corresponding experiment of the random combination 1 is about 1.4 times that of the corresponding experimental result of the quadrant distribution combination, and the comprehensive RMS error of the conversion point in the corresponding experiment of the random combination 2 is about 2 times that of the corresponding experimental result of the quadrant distribution combination. The errors of each conversion point in each coordinate component and the position error in the corresponding experiment of the common point quadrant distribution combination are smaller than the conversion errors in the corresponding experiments of random combination one and random combination two, which also shows that its conversion quality is better.

[0145] In summary, the quality of the conversion results of the common point quadrant distribution combination corresponding to the experiment is better than that of the random combination 1 and random combination 2. The coordinate transformation parameters solved by the common point quadrant distribution combination are more accurate and more stable.

[0146] In this embodiment, the robot workspace is divided into four quadrants using the X-axis and Y-axis in the base coordinate system. As other implementations, the robot workspace can also be divided into four quadrants using the X-axis and Z-axis, or the Y-axis and Z-axis. In the corresponding step c) of step 2, when selecting common points, the common points in adjacent quadrants must have different heights on the Y-axis or X-axis.

[0147] In this embodiment, in order to balance the rapidity and accuracy of coordinate conversion, 6 common points are selected when selecting common points in step 3, and the coordinate conversion relationship between the measurement coordinate system and the base coordinate system is determined by using the selected 6 common points. As other implementation methods, when selecting common points, only steps a) to c) are performed to select 4 common points, and steps d) to e) are no longer performed, and the coordinate conversion relationship between the measurement coordinate system and the base coordinate system is determined by using these 4 common points to improve calculation efficiency. Of course, it is also possible to only perform steps a) to d) to select 5 common points, and no longer perform step e), and use these 5 common points to determine the coordinate conversion relationship between the measurement coordinate system and the base coordinate system.

[0148] In this embodiment, the measuring instrument is a laser tracker with high measuring accuracy. As another embodiment, an industrial measuring camera can also be selected.

Claims

1. A method for determining the coordinate transformation relationship between a robot and a measuring instrument, characterized in that: The steps include: 1) Control the robot to teach n positions in the robot working space corresponding to n teaching points, where n≥6; 2) Define the coordinate system of the robot installation base as the base coordinate system, and divide the robot working space into 4 quadrants based on any two axes of the base coordinate system; calculate the average value of the coordinates of n teaching points, and the coordinate point corresponding to the average value is the first center of gravity; in each quadrant, select a teaching point that is far away from the first center of gravity as a common point as the first common point, the second common point, the third common point and the fourth common point, and the two common points in adjacent quadrants have different coordinate values ​​of another axis in the base coordinate system; the selected common points also include a fifth common point, and the fifth common point is selected by the following method: select two common points in adjacent quadrants from the first common point, the second common point, the third common point and the fourth common point, and calculate the average value of the coordinates corresponding to the two common points, and the coordinate point corresponding to the average value is the second center of gravity, Among all the teaching points except the first common point, the second common point, the third common point and the fourth common point, the teaching point closest to the second centroid is selected as the fifth common point; 3) Define the coordinate system of the measuring instrument as the measuring coordinate system, obtain the coordinates of the common point in the measuring coordinate system through the measuring instrument, and calculate the coordinate transformation relationship between the robot base coordinate system and the measuring coordinate system according to the coordinates of the selected common point in the measuring coordinate system and the coordinates of the common point in the base coordinate system.

2. The method for determining the coordinate transformation relationship between a robot and a measuring instrument according to claim 1, characterized in that: The common points selected in step 2) also include a sixth common point, and the sixth common point is selected by the following method: select the remaining two common points from the first common point, the second common point, the third common point and the fourth common point, and calculate the average value of the coordinates corresponding to the remaining two common points, the coordinate point corresponding to the average value is the third center of gravity point, and among all teaching points except the first common point, the second common point, the third common point, the fourth common point and the fifth common point, the teaching point closest to the third center of gravity point is selected as the sixth common point.

3. The method for determining the coordinate transformation relationship between a robot and a measuring instrument according to claim 1, characterized in that: In step 2), any two axes of the base coordinate system are the X-axis and the Y-axis of the base coordinate system, and the other axis is the Z-axis of the base coordinate system.

4. The method for determining the coordinate transformation relationship between a robot and a measuring instrument according to claim 1, characterized in that: In step 2), the coordinates of the n teaching points are the coordinates of the n teaching points in the measurement coordinate system.

5. The method for determining the coordinate transformation relationship between a robot and a measuring instrument according to claim 1, characterized in that: A flange is provided at the end of the robot, and a target ball is installed at the end of the robot flange. The coordinate system of the robot flange end is defined as the flange end coordinate system, the coordinate system of the target ball is defined as the tool coordinate system, and the center of the target ball is defined as the origin of the tool coordinate system; the coordinates of the common point in step 3) in the base coordinate system are obtained by the following method: ①According to the robot forward kinematics, the coordinate transformation relationship between the flange end coordinate system and the base coordinate system is obtained; ② Obtain the teaching pendant readings of n teaching points, select n1 teaching pendant readings from the n teaching pendant readings, the teaching pendant readings represent the coordinate transformation relationship between the flange end coordinate system and the base coordinate system, where n1 ≥ 3; then obtain the coordinates of the n1 teaching points in the measurement coordinate system, and use the following formula to calculate the position of the target sphere center in the flange end coordinate system: in, is the teaching pendant reading of two teaching points located at positions j and k respectively among the n1 teaching points, are the coordinates of the two teaching points in the measurement coordinate system, is the coordinate of the target sphere center in the flange end coordinate system; ③ According to the teaching pendant readings of n teaching points and the position of the target sphere center in the flange end coordinate system, determine the coordinates of each teaching point in the base coordinate system.

6. The method for determining the coordinate transformation relationship between a robot and a measuring instrument according to claim 5, characterized in that: In step ②, the position of the target sphere center in the flange end coordinate system is calculated using the least squares method.

7. The method for determining the coordinate transformation relationship between a robot and a measuring instrument according to claim 5, characterized in that: The n1=3.

8. The method for determining the coordinate transformation relationship between a robot and a measuring instrument according to any one of claims 1 to 7, characterized in that: The measuring instrument is a laser tracker.

Citation Information

Patent Citations

  • Multi-scale measurement field integrated detecting method used for airplane digital assembly

    CN104154859A

  • Spacecraft product final assembly precision measuring method based on workpiece coordinate system

    CN105651166A

  • Surveying coordinate system registration method

    CN107860346A

  • Method for solving coordinates of arbitrary point by using single-picture photography

    CN108344399A

  • Image-based distance measurement method and device

    CN109752709A