A point laser plane scanning tracking method

Abstract

This invention discloses a point laser planar scanning and tracking method, specifically comprising: S1: obtaining the installation pose of the point laser sensor in the flange coordinate system through a parameter calibration process; S2: constructing the position expression of the sampling point in the robot coordinate system; S3: continuously measuring at specified intervals during robot movement, creating a point laser distance data queue and a robot pose data queue, and performing staggered matching to construct pose-distance matching data; S4: constructing a surface point cloud data queue based on the pose-distance matching data at a certain moment; S5: selecting the optimal point position from the surface point cloud data queue as a reference point when entering the tracking state; S6: using the reference point to calculate the correction amount of the pose of the end effector mounted on the robot, calculating the command point based on the correction amount, and controlling the robot to move towards the command point as the target point. This invention can achieve real-time data scanning and trajectory correction using a simple point laser device.

Description

Technical Field

[0001] This invention belongs to the field of industrial robot technology, and in particular relates to a point laser planar scanning and tracking method. Background Technology

[0002] Industrial robots are multi-degree-of-freedom, multi-functional, and reprogrammable general-purpose machines that play a vital role in many fields such as electronics, mechanics, and automation. Industrial robots offer advantages such as high precision, good stability, and large load capacity, improving production efficiency and quality while reducing the labor intensity of workers. The development of industrial robots plays a crucial role in improving labor productivity in manufacturing, reducing labor intensity and enterprise production costs, enhancing the international competitiveness and quality of products, improving working conditions and the labor environment, reducing environmental pollution, and achieving energy conservation and emission reduction.

[0003] When industrial robots are used in laser cleaning, spraying, and surface treatment applications, they need to maintain a constant height relative to the surface being processed. Point laser devices, as a commonly used non-contact displacement measurement sensor, have advantages such as small size, high precision, no moving parts, and no impact on the surface being processed, and are widely used in data measurement and correction scenarios. However, point lasers themselves only sample distance information, and in related applications, they are only used to assist other measuring devices in achieving precise positioning and data acquisition. There is a lack of methods to directly use point laser measurement data for trajectory scanning and tracking.

[0004] Application No. 202110798597.9 discloses a method for extracting 3D coordinates from images in a robot scene. This method primarily involves layering XY plane images at different heights (Z) from the target surface. The camera and a point laser are fixedly mounted at suitable positions on the robot's end effector. The robot and its auxiliary axes are moved in a photographing posture to ensure the real-time image's central crosshair falls on the target point. With the camera's line of sight substantially perpendicular to the target surface, the 3D coordinates of a point on any plane can be obtained, which is beneficial for robot path planning. However, in this scenario, the point laser is only used to supplement depth information in the camera image, thereby converting vertical plane point information into 3D point information. Furthermore, this technical solution restricts the robot's measurement posture to be perpendicular to the working surface.

[0005] Application No. 2021104768138 discloses a method, system, and storage medium for precise positioning using a single-line laser combined with point laser. The main process involves scanning the outline of a target object using a single-line laser radar and servo motor before the robotic arm's trajectory planning, generating point cloud data. Processing the point cloud yields the three-dimensional coordinates of the target object's corners or edges. The robotic arm, carrying the point laser, then precisely positions itself within a designated small area. Based on the precisely positioned three-dimensional coordinates, the robotic arm's motion trajectory is planned with millimeter-level accuracy. However, the robot must pre-execute a point scanning process before performing the machining action, and the trajectory must be corrected based on the results. This method is not suitable for scenarios requiring fine-tuning of a known machining trajectory. Summary of the Invention

[0006] Purpose of the invention: In order to solve the problems existing in the prior art, the present invention provides a point laser planar scanning tracking method.

[0007] Technical solution: This invention provides a point laser planar scanning tracking method, comprising the following steps:

[0008] S1: The point laser sensor is fixedly installed at the end of the robot. The point laser sensor can move and rotate with the end of the robot. The installation pose C of the point laser sensor in the flange coordinate system is obtained through the parameter calibration process.

[0009] S2: Based on the installation pose C of the point laser sensor in the flange coordinate system, the pose A of the robot in the robot coordinate system, and the position B of the sampling point of the surface to be measured in the point laser measurement coordinate system, construct the position expression of the sampling point in the robot coordinate system.

[0010] S3: During the robot's movement, continuously measure at specified intervals, create a point laser distance data queue and a robot pose data queue, mismatch according to the actual situation, and then construct pose-distance matching data;

[0011] S4: Based on the pose-distance matching data of the lower position at a certain moment, the installation pose C of the point laser sensor in the flange coordinate system and the expression constructed in step 2, the coordinates of the sampling point in the robot coordinate system at that moment are obtained, and a surface point cloud data queue is constructed.

[0012] S5: When entering the tracking state, the robot takes the planned trajectory point to be executed as the target, selects the optimal point in the surface point cloud data queue according to the proximity principle, and uses it as the reference point for the tracking action at this time.

[0013] S6: Use the reference point to calculate the pose correction of the end effector mounted on the robot, calculate the command point based on the correction, and control the robot to move with the command point as the target point.

[0014] Furthermore, step 1 specifically includes:

[0015] Step 1.1: Move the robot to align the measurement point of the point laser sensor with the fixed reference point, and ensure that the laser data is valid;

[0016] Step 1.2: Based on the requirements of Step 1.1, under the condition that the robot pose is different and the point laser distance is the same, collect n sets of robot position-laser distance data, where the laser distance is the distance between the surface to be measured and the end point of the point laser sensor lens; based on the n sets of robot position-laser distance data, use the least squares method to calculate the position of the fixed reference point in the robot coordinate system;

[0017] Step 1.3: Based on the requirements of Step 1.1, move the robot and collect another set of robot position-laser distance data. When collecting this set of data, the robot's pose and point laser distance are different from those in Step 1.2.

[0018] Step 1.4: Based on the position of the fixed reference point in the robot coordinate system obtained in Step 1.2 and the data collected in Step 1.3, calculate the coordinate values ​​of the point laser distance in Step 1.2 and the laser distance in Step 1.3 in the flange coordinate system. Connect the two coordinates in the flange coordinate system and obtain the installation posture of the point laser sensor based on the direction of the connection. Finally, calculate the installation posture C of the point laser sensor in the flange coordinate system.

[0019] Furthermore, step 2 specifically includes:

[0020] Step 2.1: Based on the robot's pose A in the robot coordinate system and the point laser sensor's mounting pose C in the flange coordinate system, obtain the pose of the point laser sensor in the robot coordinate system.

[0021] Step 2.3: Construct the expression for the position of the sampling point on the surface to be measured in the robot coordinate system: D = C × B × A.

[0022] Furthermore, step 3 specifically includes:

[0023] Step 3.1: During the robot's movement, continuously collect position B and pose A at a specified period, and store the collected position B into the point laser distance data queue in time order, and store the pose A into the robot pose data queue in time order.

[0024] Step 3.2: Based on the system delay of the actual sampling process, compare the data acquisition times of the two queues in Step 3.1, and interpolate the adjacent two data points in the queue with the relatively earlier acquisition time;

[0025] Step 3.3: Match the interpolated queues and output the matched pose-distance data (A, B).

[0026] Furthermore, step 4 specifically involves: performing window averaging filtering on the coordinates of the sampling points calculated in step 4 in the robot coordinate system, calculating the distance between the filtered result and the last point in the surface point cloud data queue, and if the distance is greater than or equal to a preset threshold, then saving the filtered result to the surface point cloud data queue; otherwise, not saving it.

[0027] Furthermore, the proximity principle in step 5 specifically involves: projecting the points in the surface point cloud data queue onto the XOY plane of the end-effector coordinate system to obtain the projection distance between the point and the XOY plane, and selecting the point with the smallest projection distance as the reference point under the current pose.

[0028] Further, step 6 specifically involves: transforming the reference point to the end-effector coordinate system to obtain the projected distance G of the reference point in the Z-axis direction of the end-effector coordinate system; calculating the deviation I, where I = HG, H represents the tracking height between the preset robot trajectory and the surface to be measured; and the end-effector pose is translated by the deviation I in the Z-axis direction of the end-effector coordinate system to obtain the corrected tracking point pose. The corrected tracking point pose is used as the input of the PID controller, the actual robot end-effector pose is used as the feedback of the PID controller, and the output of the PID controller is used as the command point for trajectory tracking to control the robot's movement.

[0029] Furthermore, the measurement point of the point laser sensor is located in front of the tip of the end tool.

[0030] A computer storage medium storing a computer program that, when executed by a processor, implements the aforementioned point laser planar scanning tracking method.

[0031] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the aforementioned point laser planar scanning tracking method.

[0032] Beneficial effects: This invention can achieve real-time data scanning and trajectory correction using a simple point laser device, decoupling the data scanning process from the trajectory correction action. The attitude change during the trajectory tracking process does not affect the measurement results. Since the data is recorded in the form of surface point cloud, the direction of the normal to the measured surface can be inferred from the point cloud data obtained by scanning and combined with the robot's end-effector pose, which facilitates the robot's operation posture in correcting the running direction when tracking the running trajectory. Attached Figure Description

[0033] Figure 1 This is the control flowchart of the present invention;

[0034] Figure 2 This is a structural diagram of a six-joint robot according to an embodiment of the present invention. Detailed Implementation

[0035] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0036] like Figure 1 As shown, the point laser planar scanning tracking method of the present invention specifically includes:

[0037] S1: The installation pose of the point laser sensor in the robot's end-effector coordinate system is obtained through a parameter calibration process; specifically:

[0038] S11: The robot point laser is fixedly installed at the end of the robot and can move and rotate with the end of the robot.

[0039] S12: Manually move the robot to align the point laser measurement point with the fixed reference point, and ensure that the laser data is valid.

[0040] S13: As required by step S12, under the condition that the robot pose is different and the point laser distance is the same, collect four sets of robot position-laser distance data (since the laser is emitted from the origin of the point laser measurement coordinate system along the Z+ axis, the laser distance is the distance between the surface to be measured and the origin of the point laser measurement coordinate system, which takes the end point of the point laser sensor lens as the origin). The position of the fixed reference point in space in the robot coordinate system can be calculated by using the least squares method.

[0041] S14: Move the robot according to the requirements of step S12. Under the condition that the pose and laser distance are different from those in S13, collect another set of robot position-laser distance data. Combine the position of the fixed reference point and the fifth set of robot position-laser distance data, the two positions of the two laser distances in the flange coordinate system can be calculated. The installation posture of the point laser sensor can be obtained by connecting the two positions. The installation posture C of the point laser in the flange coordinate system can be calculated by the installation posture, the position in the flange coordinate system and the laser distance value.

[0042] This step does not require the robot tool end to contact an external fixed point to measure its specific position. It only requires aligning with the fixed point and collecting five sets of robot flange position and point laser distance information as required to complete the installation posture calibration. It is easy to operate and has high measurement accuracy.

[0043] S2: Using the distance data between the surface to be measured and the installation position of the point laser sensor measured by the point laser sensor, the installation pose of the point laser sensor, and the pose of the robot end flange coordinate system, construct an expression for the position of the sampling point in the robot coordinate system;

[0044] S21: Based on the current pose A of the robot flange in the robot coordinate system and the installation pose C of the point laser sensor in the flange coordinate system obtained in step S14, obtain the current pose of the point laser sensor in the robot coordinate system.

[0045] S22: Based on the results of S21 and the position B of the surface to be measured in the point laser measurement coordinate system (that is, the position of the sampling point in the point laser measurement coordinate system), the position D of the sampling point in the robot coordinate system is obtained by the expression F(A,B,C)=C*A*B, D=F(A,B,C)=C*A*B.

[0046] S3: During the robot's movement, continuously measure at a specified cycle to create a point laser distance data queue and a robot pose data queue, and match them in a staggered manner according to the actual situation to eliminate sampling delay error;

[0047] S31: To ensure the reliability of the tracking effect, the planar scanning position should be earlier than the tracking action position. Therefore, in the trajectory movement direction during programming, the measurement point of the point laser sensor is required to be earlier than the tip of the end tool.

[0048] S32: During the robot's movement, continuously collect point laser measurement values ​​B (i.e., the position of the surface to be measured in the point laser measurement coordinate system) and robot pose A at specified intervals, and store the collected results in the point laser distance data queue and robot pose data queue in chronological order.

[0049] S33: Based on the system delay of the actual sampling process, compare the data acquisition times of the two queues in S32, and interpolate the two adjacent data in the queue with the relatively earlier acquisition time; if the time of the distance data queue is earlier than that of the robot pose data queue, then interpolate the distance data queue.

[0050] S34: Match the misaligned queues after interpolation in step S33 to compensate for the sampling delay error of the two types of data, and output the matched result (A, B).

[0051] This step does not require taking values ​​when the robot is stationary. By using an interpolation mismatch matching method, the error caused by sampling delay is eliminated, and the accuracy of measuring the spatial position of sampling points during robot movement is improved.

[0052] S4: After calculating the sampling points and filtering them according to the minimum spacing parameter of the sampling points, the point cloud data of the measured surface is obtained;

[0053] S41: Using the pose-distance matching data (A,B) output from step S34 and aligned with the time axis, read the installation pose C obtained in step S14 and substitute it into the scan point calculation equation created in step S22 to obtain the scan point data D to be processed.

[0054] S42: Perform sliding window averaging filtering on the scan point data D, calculate the distance between the filtered result and the last sampling point in the surface point cloud data queue, and save the result into the surface point cloud data queue E when the distance is greater than the required minimum spacing parameter of the sampling points.

[0055] S5: When entering the tracking state, the robot takes the planned trajectory point to be executed as the target, selects the optimal point in the point cloud according to the proximity principle, and uses it as the reference point for the tracking action at this time.

[0056] S51: Read the robot target point pose in the planned teaching trajectory (the target point pose is the end effector pose in the robot coordinate system).

[0057] S52: Convert the surface point cloud data queue E output in step S42 to the end tool coordinate system.

[0058] S53: Sort the transformed points according to their projection distance on the XOY plane of the end tool coordinate system, and take the point with the smallest projection distance as the reference point P under the current pose.

[0059] This step does not require maintaining the robot's scanning posture. The optimal reference point is selected by sorting the vertical projection distance, which eliminates the influence of changes in the distance between the point laser sampling position and the tip of the end tool, greatly improving the applicability of the tracking process.

[0060] S6: By comparing the reference point with the target trajectory point interpolated by the teaching program, the correction amount of the robot end effector can be calculated. Then, the correction amount is superimposed on the planned trajectory point to be executed, and the trajectory tracking process is completed.

[0061] Transform point P into the end-effector coordinate system obtained in step S51, and obtain its projected distance G along the Z-axis of the end-effector coordinate system. G is the distance between the robot target point and the surface of the tracked object. The robot trajectory tracking function specifies the tracking height H that the robot trajectory and the surface of the tracked object should maintain, so the deviation that should be corrected is I = HG. Translate the robot target point pose along the Z-axis of the end-effector coordinate system by the deviation I to obtain the corrected tracking point pose. Use the tracking point pose as the input to the PID controller, the actual robot end-effector pose as the feedback to the PID controller, and the output of the PID controller as the final command point for trajectory tracking to control the robot's movement.

[0062] This embodiment uses a six-axis industrial robot and a point laser device for data scanning and tracking. During operation, the measurement position is approximately 5 cm ahead of the tool tip in the trajectory direction. The workpiece is a power battery tray with discontinuous surfaces. Another point laser device was installed at the tool position to verify this method. The measurement direction of this point laser is consistent with the tool's Z-axis direction. The robot used in this embodiment is as follows... Figure 2 As shown.

[0063] The method of this embodiment enables a six-axis industrial machine to perform scanning tracking at a speed of 5 m / min on a discontinuous splicing plane (with a drop of 5 mm). The actual effect of the scanning tracking process is shown in Table 1.

[0064] Table 1

[0065] Tracking error RMS mean Maximum value Point laser tracking 0.26mm 0.24mm 2.67mm

[0066] The device mounted on the robot tool can measure the actual distance between the tool tip and the plane, and the tracking deviation can be obtained by subtracting the tracking height set in the plane tracking command. As can be seen from Table 1, the present invention achieves a relatively good tracking effect.

[0067] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the present invention will not describe the various possible combinations separately.

Claims

1. A point laser plane scanning tracking method, characterized by, Includes the following steps: S1: Fix the point laser sensor on the end of the robot. The point laser sensor can move and rotate with the end of the robot. The installation pose C of the point laser sensor on the flange coordinate system is obtained through the parameter calibration process. S2: Based on the installation pose C of the point laser sensor in the flange coordinate system, the pose A of the robot in the robot coordinate system, and the position B of the sampling point of the surface to be measured in the point laser measurement coordinate system, construct the position expression of the sampling point in the robot coordinate system. S3: During the robot's movement, continuously measure at specified intervals, create a point laser distance data queue and a robot pose data queue, mismatch according to the actual situation, and then construct pose-distance matching data; S4: Based on the pose-distance matching data of the lower position at a certain moment, the installation pose C of the point laser sensor in the flange coordinate system and the expression constructed in step 2, the coordinates of the sampling point in the robot coordinate system at that moment are obtained, and a surface point cloud data queue is constructed. S5: When entering the tracking state, the robot takes the planned trajectory point to be executed as the target, selects the optimal point in the surface point cloud data queue according to the proximity principle, and uses it as the reference point for the tracking action at this time. S6: Use the reference point to calculate the correction amount of the pose of the end effector mounted on the robot, calculate the command point based on the correction amount, and control the robot to move with the command point as the target point; S6 specifically involves: transforming the reference point to the end-effector coordinate system, obtaining the projected distance G of the reference point in the Z-axis direction of the end-effector coordinate system, calculating the correction amount I, I=HG, where H represents the tracking height between the preset robot trajectory and the surface to be measured, and the end-effector pose is translated by the Z-axis correction amount I in the end-effector coordinate system to obtain the corrected tracking point pose. The corrected tracking point pose is used as the input of the PID controller, the actual robot end-effector pose is used as the feedback of the PID controller, and the output of the PID controller is used as the command point for trajectory tracking to control the robot's movement.

2. The method of claim 1, wherein, Specifically, S1 is: Step 1.1: Move the robot to align the measurement point of the point laser sensor with the fixed reference point, and ensure that the laser data is valid; Step 1.2: Based on the requirements of Step 1.1, under the condition that the robot pose is different and the point laser distance is the same, collect n sets of robot position-laser distance data, where the laser distance is the distance between the surface to be measured and the end point of the point laser sensor lens; based on the n sets of robot position-laser distance data, use the least squares method to calculate the position of the fixed reference point in the robot coordinate system; Step 1.3: Based on the requirements of Step 1.1, move the robot and collect another set of robot position-laser distance data. When collecting this set of data, the robot's pose and point laser distance are different from those in Step 1.

2. Step 1.4: Based on the position of the fixed reference point in the robot coordinate system obtained in Step 1.2 and the data collected in Step 1.3, calculate the coordinate values ​​of the point laser distance in Step 1.2 and the laser distance in Step 1.3 in the flange coordinate system. Connect the two coordinates in the flange coordinate system and obtain the installation posture of the point laser sensor based on the direction of the connection. Finally, calculate the installation posture C of the point laser sensor in the flange coordinate system.

3. The method of claim 1, wherein, Specifically, S2 is: Step 2.1: Based on the robot's pose A in the robot coordinate system and the point laser sensor's mounting pose C in the flange coordinate system, obtain the pose of the point laser sensor in the robot coordinate system. Step 2.3: Construct the expression for the position of the sampling point on the surface to be measured in the robot coordinate system: D=C×B×A.

4. The point laser planar scanning and tracking method according to claim 1, characterized in that, Specifically, S3 is: Step 3.1: During the robot's movement, continuously collect position B and pose A at a specified period, and store the collected position B into the point laser distance data queue in time order, and store the pose A into the robot pose data queue in time order. Step 3.2: Based on the system delay of the actual sampling process, compare the data acquisition times of the two queues in Step 3.1, and interpolate the adjacent two data points in the queue with the relatively earlier acquisition time; Step 3.3: Match the interpolated queues by misalignment and output the matched pose-distance data (A, B).

5. The point laser planar scanning and tracking method according to claim 1, characterized in that, Specifically, S4 involves: performing window averaging filtering on the coordinates of the sampling points obtained in step 4 in the robot coordinate system, calculating the distance between the filtered result and the last point in the surface point cloud data queue, and saving the filtered result to the surface point cloud data queue if the distance is greater than or equal to a preset threshold; otherwise, not saving it.

6. The point laser planar scanning and tracking method according to claim 1, characterized in that, The proximity principle in S5 specifically means: projecting the points in the surface point cloud data queue onto the XOY plane of the end-effector coordinate system, obtaining the projection distance between the point and the XOY plane, and selecting the point with the smallest projection distance as the reference point under the current pose.

7. The point laser planar scanning and tracking method according to claim 1, characterized in that, The measurement point of the point laser sensor is located in front of the tip of the end-effector.

8. A computer storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements a point laser planar scanning tracking method as described in any one of claims 1-7.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements a point laser planar scanning tracking method as described in any one of claims 1-7.

Images