Robot kinematics parameter error calibration method based on mark point real-time tracking

Abstract

The invention discloses a robot kinematics parameter error calibration method based on mark point real-time tracking, and the method comprises the steps: tracking a preset mark point in real time through a camera measurement unit, dynamically monitoring the actual position and posture change of a robot in a motion process, and calculating the error of kinematics parameters in real time. Kinematics parameters of the robot at all time points are determined in real time, dynamic compensation is conducted on the robot according to real-time data, and the positioning precision and motion stability of the robot in a dynamic task are improved; according to the method, the cost of adopted equipment is far lower than that of traditional robot calibration equipment such as a laser tracker, meanwhile, the problem that light is blocked when the laser tracker conducts measurement is solved, and specific values of kinematics parameters of the robot at all time points can be determined in real time; and the robot can be dynamically compensated according to the real-time data, so that the positioning precision and the motion stability of the robot in a dynamic task are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the field of precision measurement, and particularly to a method for calibrating the kinematic parameter errors of a robot based on real-time tracking of marker points. Background Art

[0002] As one of the core devices of intelligent manufacturing, industrial robots have become a key driving force for its development due to their multiple advantages such as high efficiency, precision, reliability, and safety. However, there are errors in the processing and production of robots or the positioning accuracy of robots gradually decreases during the working process with the wear of joint drive systems such as gears, which directly affects the positioning accuracy of robots.

[0003] Robot calibration, as a method that can significantly improve the positioning accuracy of robots, has been widely adopted and applied by many researchers due to its advantages such as simple operation and low cost.

[0004] Most researchers use a laser tracker to measure the end of the robot for calibration. However, the laser tracker needs to maintain the light tracking between the target ball, the light is easily blocked and it is difficult to determine the spatial coordinates of the robot moving outside the measurement area. Moreover, the laser tracker can only measure the position of one object at a time, which affects the measurement efficiency.

[0005] At the same time, most calibration methods mainly calibrate the kinematic parameters by directly tracking the position of the end of the robot. The core lies in solving the optimal fitting solution between the actual position and the theoretical position of all sampling points through calibration algorithms such as Levenberg-Marquard. However, it is difficult to meet the requirements of real-time monitoring and precise compensation of kinematic parameters during the dynamic movement of the robot.

[0006] With the wide application of robot technology in complex dynamic tasks, such as in scenarios of intelligent manufacturing, automated assembly, and high-precision operation, the need for real-time monitoring and compensation of robot kinematic parameter errors is becoming increasingly urgent.

[0007] In order to break through the limitations of the existing calibration methods, it is urgent to propose a method for calibrating the kinematic parameter errors of a robot based on real-time tracking of marker points.

[0008] Based on the above deficiencies in the existing technology, the present invention is proposed. Summary of the Invention

[0009] Aiming at the above deficiencies in the existing technology, the purpose of the present invention is to provide a method for calibrating the kinematic parameter errors of a robot based on real-time tracking of marker points to solve the requirements of real-time monitoring and precise compensation of kinematic parameters during the movement of the robot.

[0010] To achieve the invention object of the present invention, the technical solution provided by the present invention is as follows:

[0011] A method for calibrating the kinematic parameter error of a robot based on real-time tracking of marker points, which uses a camera measurement unit to track preset marker points in real time, dynamically monitors the actual position and attitude changes of the robot during movement, calculates the kinematic parameter errors in real time, determines the kinematic parameters of the robot at each time point in real time, and dynamically compensates the robot according to the real-time data to improve the positioning accuracy and motion stability of the robot in dynamic tasks.

[0012] The specific steps are as follows:

[0013] Step S1, paste marker points on the robot and at fixed positions in the robot's working space.

[0014] Paste marker points on the joints of the robot to ensure that there are at least 3 non-collinear marker points to avoid the inability to uniquely determine rigid body motion; Step S2, place the camera positions so that the fields of view of multiple cameras can cover the positions of the marker points pasted in Step S1.

[0015] The camera measurement unit consists of a triangular bracket and a camera. Set the position of the camera measurement unit so that the fields of view of multiple cameras can cover the positions of the marker points pasted on the robot and in the robot's working space; when the robot moves within a large range of arbitrary spaces in the feasible region, the camera measurement unit can identify and track each marker point.

[0016] Step S3, the robot starts working, driving the scanning measurement system to scan and measure an object.

[0017] The robot drives the precise scanning measurement system to move according to a preset program. Under the drive of the robot, the scanning measurement system scans and measures the object along a set trajectory; the line structured light of the scanning measurement system is accurately projected onto the surface of the object, and the camera quickly captures the information of the object to be measured. The data is transmitted and stored in real time, providing a data basis for subsequent detection.

[0018] Step S4, the camera measurement unit tracks the marker points in real time and measures the marker points pasted on the robot and in the robot's working space.

[0019] As the robot scans and measures the object, the cameras of the camera measurement unit record the dynamic changes of each marker point to achieve the purpose of tracking; at the same time, the camera measurement unit also measures the marker points pasted in the robot's working space.

[0020] Step S5, calculate the actual kinematic parameters of each joint of the robot according to the camera measurement of the marker points.

[0021] Step S6, record the kinematic parameters of the robot joints and the actual kinematic parameters measured and calculated by the camera measurement unit, and compensate them;

[0022] Record the kinematic parameters of the robot and the actual kinematic parameters measured and calculated by the camera measurement unit. Compare the parameter errors between the actually measured kinematic parameters and the actual kinematic parameters recorded by the robot to obtain the actual kinematic parameter errors. Substitute the actual kinematic parameter errors into the calculation to realize the compensated object point cloud coordinates;

[0023] Through the compensation of the robot kinematic parameters, real-time compensation of the robot kinematic parameter errors at this measurement position is achieved to improve the robot positioning accuracy and complete the calibration task;

[0024] Step S7, determine whether the object scanning measurement in step S3 is completed. If it is, proceed to the next step S8 to complete the measurement; if not, return to step S3, and the robot continues to work, driving the scanning measurement system to scan and measure the object;

[0025] Step S8, complete the measurement, and realize robot calibration and measurement.

[0026] The preferred technical solution provided by the present invention is:

[0027] In the said step S1, the marked points pasted at the fixed positions of the robot are specifically marked points pasted on the joints of the robot, ensuring that there are at least 3 non-collinear marked points to avoid the non-unique determination of rigid body motion; the center of gravity of the distribution of each joint marked point passes through the rotation axis; with the increase in the number of marked points, the radius of the measurement points, and the rotation angle, the recognition accuracy of the rotation axis will be improved.

[0028] The further preferred technical solution provided by the present invention is:

[0029] The specific method for calculating the actual kinematic parameters of each joint of the robot according to the camera measurement of the marked points in the said step S5 is:

[0030] The transformation relationship between the (i - 1)-th joint and the i-th joint of the robot is expressed as:

[0031]

[0032] where, θ i represents the joint variable; d i represents the link offset; a i represents the link length; α i represents the link twist angle; β i represents the rotation variable around the Y-axis; cθ i represents the cosine of θ i cosθ i , cαi Denote α i as the cosine of α, cosα i ; sθ i Denote θ i as the sine of θ, sinθ i , sα i Denote α i as the sine of α, sinα i ; Rot represents rotational motion, and Trans represents translational motion;

[0033] Perform total differentiation on , then the transformation matrix error between joints is expressed by Equation (2) as:

[0034]

[0035] where, Δθ i , Δd i , Δa i , Δα i and Δβ i are the corresponding errors of the kinematic parameters θ i , d i , a i , α i and β i ;

[0036] The point cloud data P of the object measured by the scanning measurement system under the coordinate system of the robot can be further expressed as:

[0037]

[0038] where, is the matrix for hand-eye calibration of the robot unit; P1 is the point cloud coordinates of the object measured by the scanning measurement system, expressed as:

[0039]

[0040] where, [x k y k z k (k = 1, 2, …, j) are the coordinates of the light stripe points;

[0041] By image processing, the position and motion trajectory of the marker points are analyzed in real time, so as to accurately measure the key parameters of the robot's attitude, displacement and motion speed in three-dimensional space; at the same time, the camera also measures the marker points pasted in the robot's working space. The marker points in the robot's working space serve as a reference benchmark to help the system accurately construct the coordinate relationship of the entire working area, ensure that the robot can accurately locate and measure the target object during work, and obtain the robot kinematic parameter θ in real time according to the front and back changes of the three-dimensional coordinates of each joint marker pointi ,d i ,a i ,α i and β i 。

[0042] Another preferred technical solution provided by the present invention is as follows:

[0043] In step S3, the scanning measurement system is a monocular cross structure, including a monocular camera and a crosshair structured light. The crosshair structured light is accurately projected onto the object surface, and the monocular camera quickly captures the information of the object. By capturing the light strip of the crosshair structured light device through the monocular camera, the purpose of measurement is achieved.

[0044] The present invention also provides a technical solution for the application of a method for calibrating the kinematic parameter errors of a robot based on real-time tracking of marker points in machine tool calibration.

[0045] Compared with the prior art, the beneficial effects of the present invention include:

[0046] 1. For the method for calibrating the kinematic parameter errors of a robot based on real-time tracking of marker points according to the present invention, the equipment cost is much lower than that of traditional robot calibration equipment such as laser trackers. At the same time, problems such as light occlusion during measurement by laser trackers are avoided, as well as the difficulty in determining the spatial coordinates of a robot moving outside the measurement area. Additionally, the laser tracker can only measure the position of one object at a time, which affects the measurement efficiency.

[0047] 2. For the method for calibrating the kinematic parameter errors of a robot based on real-time tracking of marker points according to the present invention, not only can the specific values of the kinematic parameters of the robot at each time point be determined in real time, but also the robot can be dynamically compensated based on these real-time data, thereby significantly improving the positioning accuracy and motion stability of the robot in dynamic tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0049] Figure 1 It is a schematic diagram of the calibration process of a method for calibrating the kinematic parameter errors of a robot based on real-time tracking of marker points according to the present invention;

[0050] Figure 2 It is a schematic diagram of the measurement process of a method for calibrating the kinematic parameter errors of a robot based on real-time tracking of marker points according to the present invention. Detailed implementation manners

[0051] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0052] Figure 1 It is a schematic diagram of the calibration process of a kinematic parameter error calibration method for a robot based on real-time tracking of marker points according to the present invention. Figure 2 It is a schematic diagram of the measurement process of a kinematic parameter error calibration method for a robot based on real-time tracking of marker points according to the present invention. Combining the attached Figure 1 and Figure 2 As shown in, a kinematic parameter error calibration method for a robot based on real-time tracking of marker points according to the present invention uses a camera measurement unit 5 to track the marker points 2 on the robot 1 in real time, dynamically monitors the actual position and attitude changes of the robot 1 during the movement process, and calculates the error of the kinematic parameters in real time; in this way, not only can the specific values of the kinematic parameters of the robot 1 at each time point be determined in real time, but also the robot 1 can be dynamically compensated according to these real-time data, thereby significantly improving the positioning accuracy and movement stability of the robot 1 in dynamic tasks; this calibration method based on real-time tracking of the marker points 2 can provide strong support for the high-precision operation of the robot 1 in a complex dynamic environment, and promote the development and application of robot technology in more high-precision and high-dynamic application scenarios.

[0053] The specific implementation steps of the above technical solutions are introduced as follows:

[0054] Step S1, paste marker points on the robot 1 and at fixed positions in the working space of the robot 1;

[0055] Marker points 2 are pasted on each joint of the robot 1 to ensure that there are at least 3 non-collinear marker points 2 to avoid the non-unique determination of rigid body motion;

[0056] The center of gravity of the distribution of each joint marker point 2 preferably passes through the rotation axis; of course, as the number of marker points 2, the radius of the measurement point, and the rotation angle increase, the recognition accuracy of the rotation axis will improve.

[0057] Step S2, place the camera positions so that the fields of view of multiple cameras can cover the positions of the marker points pasted on the robot and the working space;

[0058] The camera measurement unit 5 is composed of a triangular bracket 501 and a camera 502. Set the position of the camera measurement unit 5 so that the fields of view of multiple cameras can cover the positions of the marker points pasted on the robot 1 and the working space of the robot 1; when the robot 1 moves in a large range of arbitrary spaces within the feasible region, the camera measurement unit 5 can identify and track each marker point 2;

[0059] Step S3: The robot 1 starts working, driving the scanning system 3 to scan and measure the object 4.

[0060] The robot 1 drives the precise scanning and measuring system 3 to move according to a preset program. Under the drive of the robot, the scanning and measuring system 3 scans and measures the object 4 along a set trajectory.

[0061] The scanning and measuring system 3 has a monocular cross structure, including a monocular camera 301 and a crosshair structured light 302. The crosshair structured light is accurately projected onto the surface of the object 4, and the monocular camera 301 quickly captures the information of the object. By capturing the light strip of the crosshair structured light 302 device through the monocular camera 301, the purpose of measurement is achieved. These data are transmitted and stored in real time, providing a data basis for subsequent detection.

[0062] Step S4: The camera measurement unit 5 tracks the marker points in real time and measures the marker points pasted in the robot and the working space.

[0063] As the robot 1 scans and measures the object 4, the camera 502 of the camera measurement unit 5 records the dynamic changes of each marker point 2 to achieve the purpose of tracking.

[0064] At the same time, the camera measurement unit 5 also measures the marker points 2 pasted in the working space of the robot 1. By replacing the absolute measurement of the moving marker points during the operation of the robot 1 with the relative measurement between the static marker and the moving marker points in the space of the robot 1, the problem of low measurement accuracy caused by low camera accuracy is solved.

[0065] Step S5: Calculate the actual kinematic parameters of each joint of the robot according to the camera measurement of the marker points.

[0066] The transformation relationship T between the (i - 1)-th joint and the i-th joint of the robot 1 i i-1 is expressed as:

[0067]

[0068] where, θ i represents the joint variable; d i represents the link offset; a i represents the link length; α i represents the link twist angle; β i represents the rotation variable around the Y-axis; cθ i represents the cosine of θ i cosθ i , cα i represents the cosine of α i cosα i ; sθ i represents θ iThe sine sinθ i , sα i represents α i The sine sinα i ; Rot represents rotational motion, and Trans represents translational motion.

[0069] Due to the influence of errors in processing, assembly, etc. on the robot 1, there is a certain error between the actual value and the theoretical value of the kinematic parameters of the robot 1. Performing total differentiation, the transformation matrix error between each joint is represented by Equation (2).

[0070]

[0071] Among them, Δθ i , Δd i , Δa i , Δα i and Δβ i are the corresponding errors of the kinematic parameters θ i , d i , a i , α i and β i Therefore, the point cloud data P of the object measured by the monocular line structured light measurement unit 3 in the coordinate system of the robot 1 can be further expressed as:

[0072]

[0073] Among them, is the matrix for hand-eye calibration of the robot unit 1. The hand-eye calibration process uses the method disclosed in the invention patent with the publication number CN 118559755 A and the invention name of a robot calibration method based on multi-line structured light applied by the applicant on June 13, 2024; P1 is the object point cloud coordinates measured by the monocular line structured light measurement unit 3. It can be expressed as:

[0074]

[0075] Among them, [x k y k z k (k = 1, 2, …, j) are the coordinates of the light strip points.

[0076] Through image processing, the position and motion trajectory of the marker point 2 are analyzed in real time, so as to accurately measure key parameters such as the attitude, displacement, and motion speed of the robot 1 in three-dimensional space. At the same time, the camera 502 also measures the marker points 2 pasted in the working space. These marker points 2 serve as reference benchmarks to help the system accurately construct the coordinate relationship of the entire working area, ensuring that the robot 1 can accurately locate and measure the target object 4 during work. According to the front and back changes of the three-dimensional coordinates of each joint marker point, the kinematic parameters θ i , d i , a i , α i and β i .

[0077] Step S6, record the kinematic parameters of the robot 1 joints and the actual kinematic parameters measured and calculated by the camera measurement unit 5, and compensate them;

[0078] Record the kinematic parameters of the robot 1 and the actual kinematic parameters θ i , d i , a i , α i and β i , compare the parameter errors between the actually measured actual kinematic parameters and the actual kinematic parameters recorded by the robot 1 to obtain the actual kinematic parameter errors Δθ i , Δd i , Δa i , Δα i and Δβ i . Substitute these errors into formula (3) to realize the compensated object point cloud coordinates. Through the compensation of the robot kinematic parameters, the real-time compensation of the kinematic parameter errors of the robot 1 at this measurement position is achieved to improve the positioning accuracy of the robot 1 and complete the purpose of the calibration task.

[0079] Step S7, determine whether the object scanning measurement in step S3 is completed. If it is, proceed to the next step S8 to complete the measurement; if not, return to step S3, and the robot 1 continues to work, driving the scanning system 3 to scan and measure the object;

[0080] Step S8, complete the measurement, and realize robot calibration and measurement.

[0081] At the same time, the robot kinematic parameter error calibration method based on real-time marker point tracking provided by the present invention is not only applicable to robots, but also applicable to the calibration and measurement of machine tools.

[0082] The described embodiments are only a part of the embodiments of this application, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts shall fall within the scope of protection of this application.

Claims

1. A method for calibrating the kinematic parameter errors of a robot based on real-time tracking of marker points, characterized in that, The camera measurement unit tracks preset marker points in real time, dynamically monitors the actual position and attitude changes of the robot during movement, calculates the errors of kinematic parameters in real time, determines the kinematic parameters of the robot at each time point in real time, and performs dynamic compensation on the robot based on the real-time data to improve the positioning accuracy and motion stability of the robot in dynamic tasks; The specific steps include: Step S1, paste marker points on the robot and at fixed positions in the robot's working space; Paste marker points on the joints of the robot to ensure that there are at least 3 non-collinear marker points to avoid the inability to uniquely determine rigid body motion; Step S2, place the camera positions so that the fields of view of multiple cameras can cover the positions of the marker points pasted in Step S1; The camera measurement unit consists of a triangular bracket and a camera. Set the position of the camera measurement unit so that the fields of view of multiple cameras can cover the positions of the marker points pasted on the robot and in the robot's working space; when the robot moves in a large range of arbitrary spaces within the feasible area, the camera measurement unit can identify and track each marker point; Step S3, the robot works, driving the scanning measurement system to scan and measure an object; The robot drives the precise scanning measurement system to move according to a preset program. Under the drive of the robot, the scanning measurement system scans and measures the object along a set trajectory; the line structured light of the scanning measurement system is accurately projected onto the surface of the object, and the camera quickly captures the information of the object to be measured, and the data is transmitted and stored in real time to provide a data basis for subsequent detection; Step S4, the camera measurement unit tracks the marker points in real time and measures the marker points pasted on the robot and in the robot's working space; As the robot scans and measures the object, the cameras of the camera measurement unit record the dynamic changes of each marker point to achieve the purpose of tracking; at the same time, the camera measurement unit also measures the marker points pasted in the robot's working space; Step S5, calculate the actual kinematic parameters of each joint of the robot based on the camera measurement of the marker points; Step S6, record the kinematic parameters of the robot joints and the actual kinematic parameters measured and calculated by the camera measurement unit, and compensate them; Record the kinematic parameters of the robot and the actual kinematic parameters measured and calculated by the camera measurement unit. Compare the parameter errors between the actually measured kinematic parameters and the actual kinematic parameters recorded by the robot to obtain the actual kinematic parameter errors, and substitute the actual kinematic parameter errors into the calculation to realize the compensation of the object point cloud coordinates; Through the compensation of the robot kinematic parameters, the real-time compensation of the kinematic parameter errors of the robot at this measurement position is achieved to improve the positioning accuracy of the robot and complete the calibration task; Step S7, determine whether the object scanning measurement in Step S3 is completed. If it is, proceed to the next step S8 to complete the measurement; if not, return to Step S3, and the robot continues to work, driving the scanning measurement system to scan and measure the object; Step S8, complete the measurement, realizing robot calibration and measurement.

2. The kinematic parameter error calibration method of a robot based on real-time tracking of marker points according to claim 1, wherein, In the said step S1, the marked points pasted at the fixed positions of the robot are specifically marked points pasted on the joints of the robot, ensuring that there are at least 3 non-collinear marked points to avoid the non-unique determination of rigid body motion; the center of gravity of the distribution of each joint marked point passes through the rotation axis; with the increase in the number of marked points, the radius of the measurement point, and the rotation angle, the recognition accuracy of the rotation axis will be improved.

3. The robot kinematic parameter error calibration method based on real-time tracking of marker points according to claim 1, characterized in that: In the said step S5, the specific method for calculating the actual kinematic parameters of each joint of the robot based on the marked points measured by the camera is as follows: The transformation relationship between the (i - 1)-th joint and the i-th joint of the robot It is expressed as: Among them, θ i represents joint variables; d i Indicates the offset of the rod; a i represents the length of the rod; α i represents the torsion angle of the rod; β i represents the rotation variable around the Y axis; cθ i represents θ i cosine of cosθ i , cα i Represents α i cosine cosα i ;sθ i represents θ i sine of sinθ i , sα i Represents α i sine of sinα i ;Rot represents rotational motion, Trans represents translational motion; For performing total differentiation, the transformation matrix error between each joint is expressed by Equation (2) as follows: Where Δθ i , Δd i , Δa i , Δα i and Δβ i are the kinematic parameters θ i , d i , a i , α i and β i The corresponding error; The point cloud data P of the object measured by the scanning measurement system in the coordinate system of the robot can be further expressed as: in, is the matrix of hand-eye calibration of the robot unit; P1 is the point cloud coordinate of the object measured by the scanning measurement system, expressed as: Among them, [x k y k z k (k = 1, 2, …, j) are the coordinates of the light stripe points; Through image processing, the positions and movement trajectories of the marked points are analyzed in real time, so as to accurately measure the key parameters of the robot's attitude, displacement, and movement speed in three-dimensional space; at the same time, the camera also measures the marked points pasted in the robot's working space. The marked points in the robot's working space serve as a reference benchmark to help the system accurately construct the coordinate relationship of the entire working area, ensuring that the robot can accurately locate and measure the target object during work. The kinematic parameters θ i , d i , a i , α i and β i .

4. A method for calibrating the kinematic parameter error of a robot based on real-time tracking of marker points according to claim 1, characterized in that In the said step S3, the scanning measurement system is a monocular cross structure, including a monocular camera and a crosshair structured light. The crosshair structured light is accurately projected onto the surface of the object, and the monocular camera quickly captures the information of the object. By capturing the light strip of the crosshair structured light device through the monocular camera, the measurement purpose is achieved.

5. An application of a method for calibrating the kinematic parameter error of a robot based on real-time tracking of marked points as described in any one of claims 1-4 in machine tool calibration.

Images