CN-122008240-A - Path planning method and system for robot calibration task

CN122008240ACN 122008240 ACN122008240 ACN 122008240ACN-122008240-A

Abstract

The invention relates to the technical field of motion control and calibration of industrial robots, in particular to a path planning method and a path planning system for a robot calibration task, wherein the method comprises the steps of setting parameters of a laser tracker and a target ball in the calibration task, establishing a robot kinematic model according to DH parameters of the robot, and defining a dynamically updated task coordinate system based on the parameters Task constraint vector And (2) planning in a robot joint space by adopting a random sampling path planning framework of an FR algorithm, iteratively generating feasible path nodes meeting task constraint and having no collision, and extracting key points of joint space data of the original path generated in the step (S2) by adopting an RDP algorithm to generate a calibration instruction sequence suitable for PTP motion of the robot joint space. The invention has environmental adaptability, practicability and high efficiency, can effectively solve the problems of complex joint limit and interference of working space obstacles (including ground and workpieces), and remarkably improves the efficiency of calibration task preparation.

Inventors

SHANG WEIWEI
QIN CHANGYUAN
ZHANG BIN
CHU ZHAOQI
XIAO YONGQIANG
YUAN HAOLUN

Assignees

埃夫特智能机器人股份有限公司
中国科学技术大学

Dates

Publication Date: 20260512
Application Date: 20260401

Claims (9)

1. A path planning method for a robot calibration task is characterized by comprising the following steps: Step S1, setting parameters of a laser tracker and a target ball in a calibration task, establishing a robot kinematic model according to parameters of a robot DH, and defining a dynamically updated task coordinate system based on the parameters Task constraint vector ; S2, planning by adopting a random sampling path planning framework of an FR algorithm in a robot joint space, and iteratively generating feasible path nodes meeting task constraint and having no collision; and S3, extracting key points of the joint space data of the path generated in the step S2 through an RDP algorithm, and generating a calibration instruction sequence suitable for PTP motion of the robot joint space.
2. A path planning method for robot calibration task according to claim 1, wherein the parameters set in step S1 comprise the fixed position of the laser tracker in the robot-based coordinate system Fixed offset of target ball center under robot end flange coordinate system Euler angle of target ball mirror surface under robot terminal flange coordinate system 。
3. A path planning method for robot calibration tasks as claimed in claim 2, wherein the task coordinate system in step S1 The definition is as follows: the origin of the task coordinate system is the position of the center of the target sphere in the robot base coordinate system The X axis and Y axis of the task coordinate system are defined according to the right hand rule and the suitability of the current movement direction, and the Z axis of the task coordinate system is the slave Pointing to Is a vector of the vector (b).
4. A path planning method for robot calibration task according to claim 3, wherein the task constraint vector in step S1 Is defined as 。
5. The method for planning paths for calibration tasks of robots according to claim 1, wherein the planning procedure of the FR algorithm in step S2 is as follows: step S21, defining tree nodes as joint angle vectors Two trees are initialized And , Is the root node of , Is the root node of ; Step S22, entering a main loop, namely randomly sampling nodes Assume that the current active tree is Find its nearest neighbor node ; Step S23, slave node Directional node Direction expansion step length to obtain node ; Step S24, calling FR algorithm to calculate and obtain feasible nodes meeting constraint ; Step S25, performing collision detection, if the node is To a feasible node Motion process of (a) and feasible nodes If no collision exists, the feasible nodes are obtained by using the RRT-Connec algorithm Adding in ; Step S26, switching the activity tree to be Attempt to get from Medium-distance feasible node The nearest node expands to the node until the node collides with an obstacle and reaches a feasible node Or reaching the upper limit of the expansion step number; step S27, if the two trees are successfully connected through a certain expansion, the planning is successful, and a slave node is obtained by backtracking To the node Is used to determine the initial path of the path(s).
6. The method for planning the path for the calibration task of the robot according to claim 5, wherein the step S24 comprises the following specific steps: Step S241, computing node Corresponding task coordinate system Pose of current tail end target ball ; Step S242, calculating an error between the pose of the tip in the task coordinate system under the current pose and the desired pose defined by the task constraint vector ; Step S243, constraint jacobian matrix based on task Pseudo-inverse of (2) Calculating joint adjustment And performing iterative correction; Step S244, iterate to Or reach the maximum iteration number to obtain feasible nodes meeting the constraint Wherein Represented as the convergence threshold of the FR algorithm.
7. The method for path planning for a robot calibration task of claim 1, wherein the collision detection in step S25 is based on a predefined geometrical model of the robot body, end tool and environmental obstacle.
8. The path planning method for the robot calibration task according to claim 1, wherein the specific flow in the step S3 is as follows: S31, calculating a first-order difference of an original path point sequence, obtaining Euclidean distances between adjacent path points, and determining a simplified threshold value of an RDP algorithm based on the Euclidean distances; And S32, for a given path segment, calculating the vertical distance from all internal points to a chord formed by connecting the head and tail points of the segment, if the maximum vertical distance exceeds a simplification threshold, dividing the path segment into two subsections by taking the furthest point as a dividing point, recursively processing, and if the maximum vertical distance does not exceed the simplifying threshold, discarding all internal points in the segment, and only keeping the head and tail points.
9. A path planning system for a robot calibration task, characterized by performing a path planning method for a robot calibration task according to any one of claims 1 to 8.

Description

Path planning method and system for robot calibration task Technical Field The invention relates to the technical field of motion control and calibration of industrial robots, in particular to a path planning method and a path planning system which generate no collision, no loss of measurement signals and smooth motion for automatic calibration tasks of robots under the constraint of external measurement equipment (a laser tracker). Background The absolute positioning accuracy of the industrial robot is a key for ensuring the operation quality of the industrial robot. In order to improve the precision, a high-precision external measuring device such as a laser tracker is generally required to calibrate parameters of the robot. In this process, the target ball mounted on the robot end flange must be continuously aligned with the laser emitting head of the laser tracker throughout the motion to keep the laser path uninterrupted, a requirement called "no-light loss" constraint. This is essentially a motion planning problem that also requires to meet specific end pose constraints under complex joint constraints, environmental obstacle constraints. Traditional calibration path generation relies primarily on cartesian space constraints of the calibration points and manual adjustment of the robot pose by the operator. The method has low efficiency, has higher experience requirements on operators, and is difficult to ensure the optimality and the safety of the path, especially in complex scenes such as large change of the robot configuration or installation of lengthening tools. A common approach is to use Point-to-Point (Point to Point) instructions of the robot controller to make a setpoint movement in the joint space. However, the simple joint space linear motion cannot ensure that the tail end gesture meets the complex geometric constraint of continuous alignment laser, laser loss is very easy to cause the calibration process to be interrupted, and the reliability and the automation degree are low. Some existing general motion planning algorithms, such as fast-expanding random trees (Rapidly Exploring Random Tree) and variants thereof, while capable of solving obstacle avoidance problems in high-dimensional space, typically deal with position space or simple pose constraints, which are not deep coupled with complex, dynamically updated task constraints such as "end-specific vector pointing to a fixed point". The Chinese patent CN202211605825.7 discloses an RRT-connect algorithm which integrates process knowledge, realizes the rapid path planning of the industrial robot for a specific process, although the method improves the path searching efficiency and the planning quality by introducing expert knowledge, an expert mechanism of the method depends on a preset process rule, and the method cannot ensure the satisfaction of the calibration task constraint from a planning source. Therefore, a method and a system for planning a calibration path capable of automatically and reliably generating a calibration path satisfying the constraint of no light loss, avoiding collision and having smooth movement are needed, so as to improve the efficiency, success rate and application range of automatic calibration of a robot. Disclosure of Invention In order to solve the technical problems, the invention provides a path planning method and a system for a robot calibration task. The technical problems to be solved by the invention are realized by adopting the following technical scheme: A path planning method for a robot calibration task comprises the following steps: Step S1, setting parameters of a laser tracker and a target ball in a calibration task, establishing a robot kinematic model according to parameters of a robot DH, and defining a dynamically updated task coordinate system based on the parameters Task constraint vector; S2, planning by adopting a random sampling path planning framework of an FR algorithm in a robot joint space, and iteratively generating a feasible path which meets task constraint and has no collision; and S3, extracting key points of the joint space data of the path generated in the step S2 through an RDP algorithm, and generating a calibration instruction sequence suitable for PTP motion of the robot joint space. As a further improvement of the invention, the parameters set in step S1 include the fixed position of the laser tracker in the robot-based coordinate systemFixed offset of target ball center under robot end flange coordinate systemEuler angle of target ball mirror surface under robot terminal flange coordinate system。 As a further development of the invention, the task coordinate system in step S1The definition is as follows: the origin of the task coordinate system is the position of the center of the target sphere in the robot base coordinate system The X axis and Y axis of the task coordinate system are defined according to the right hand rule and the suitability of the current movem