Redundant robot trajectory planning method

Abstract

The present invention relates to an intelligent manufacturing field, and more particularly relates to a redundant robot trajectory planning method. The robot is characterized in that the step comprises establishing a redundant robot kinematics and dynamics model, and obtaining the end effector joint angular velocity with the mathematic relation between the posture; adopting closed-loop pseudo-inverse method solving joint angles for trajectory planning; solving joint angles, constructing a constraint redundant robot trajectory planning target function through a firefly algorithm and solving thetarget function of the minimum value to correct the accumulated errors generated when integrated is the integration error; and improving firefly algorithm includes individual initialization, decisiondomain update, the location update and fluorescein updates the four step, wherein the location update procedure by adopting a generalized reverse learning mechanism to perform the location update. The method is high in robustness, accuracy, real-time performance and effectiveness.

Description

technical field [0001] The invention relates to the field of intelligent manufacturing, in particular to a redundant robot trajectory planning method. Background technique [0002] With the development of robotics, artificial intelligence, sensor technology, and electronic information technology, industrial robots are playing an increasingly important role in the field of intelligent equipment manufacturing. Traditional industrial robots can no longer meet the needs of more sophisticated and complex tasks. A redundant robot is an anthropomorphic mechanical arm, which has good self-motion characteristics, and is easy to deal with problems such as torque optimization and joint limit avoidance, so that it can better adapt to complex working conditions; usually, the number of joint degrees of freedom of a redundant robot It is larger than the dimension describing the operation space, and there are infinitely many sets of solutions in its inverse kinematics. Therefore, the diffi...

AI Technical Summary

Problems solved by technology

[0003] At present, the methods of redundant robot trajectory planning mainly include numerical iteration method, geometric method and algebraic solution method; among them, the numerical iteration method converts the robot inverse kinematics problem into a function optimization problem, and uses artificial intelligence algorithm for optimization calculation. This method is easy to operate , but the calculation amount is large and the speed is slow; the geometric method obtains the analytical solution under specific conditions by deriving the relationship between the pose of the end effector and each joint angle, which has the advantages of fast calculation speed and high calculation accuracy, but the mathematical modeling process is complicated. Lock; the algebraic solution uses the Jacobian matrix and its deformation to solve the differential relationship between the joint velocity and the end-effector linear velocity and angular velocity. The common ones are closed-loop pseudo-inverse method (CLP), gradient descent method, etc. Accumulated errors will affect the accuracy of trajectory tracking

Images