Path Planning Of Manipulator In Complex Environment Based On Improved RRT Algorithm

In recent years,with the guidance of the "Made in China 2025" policy,robot technology has achieved rapid development.It is playing an increasingly important role in industrial production,medical equipment,warehousing logistics,national defense construction and other fields.Among industrial robots,manipulator is the most widely used.However,in traditional industrial applications,the manipulator is mainly used for a single assembly line,and its operation relies on the operator's teaching and programming,which is inefficient and poorly adaptable.With the advent of the intelligent age,the increasingly complex environment urgently requires the manipulator to autonomously plan paths to adapt to different environments.For this reason,this article takes the AUBO-i5 industrial manipulator as the research object,and aims to propose a path planning method with strong versatility,simple application,convenient and efficient,and capable of autonomous obstacle avoidance.The main work of this paper is as follows:(1)Taking the AUBO-i5 manipulator as the research object,the overall plan of the manipulator path planning is designed,and the task is decomposed into several subtasks of manipulator modeling,path planning algorithm,collision detection,and path optimization.The focus is on the manipulator model based on the DH(Denavit-Hartenberg)parameter method and its forward kinematics equations,as well as the simplified model of the manipulator and obstacles based on the geometric envelope method;Finally,a polynomial fitting method based on least squares is given,which lays an important realization basis for the path planning of the AUBO-i5 manipulator.(2)Aiming at the problem of low sampling efficiency of RRT algorithm and low quality of sampling points,a path planning method based on the fusion of low discrepancy sequence and RRT(Sobol-RRT)is proposed.Based on the target biased RRT(Biased-RRT),the algorithm introduces the Sobol sequence sampling method,which enhances the sampling uniformity in the target search direction,effectively avoids redundancy and repeated sampling points,improves the speed of the algorithm and space exploration capability.Then a sampling point selection strategy based on sampling pool is proposed,which adopts concurrent sampling strategy to construct sampling pool and optimize sampling points.It improves the quality of sampling points and further enhances the convergence performance of the algorithm.(3)Aiming at the problem of Sobol-RRT node redundancy and easy collision of nodes near obstacles,a Sobol-RRT algorithm based on multi-index node filtering(SN-RRT)is proposed.The algorithm introduces a filtering strategy based on node sparsity,sets node distribution sparsity parameters,and does not expand the node which is larger than node sparsity,so that the node distribution is more uniform and the number of nodes is reduced.Then a node filtering strategy based on collision probability and collision direction angle is proposed.It records node information including the number of successful and failed expansion,and collision direction angles,then filters out nodes that are prone to collision,which improves the efficiency of path planning.(4)Build the manipulator experimental platform for algorithm application experiments.The platform designs an experimental scene and establishes a virtual reality scene to synchronize the operation of the manipulator.The results show that the manipulator runs smoothly and the joint angle does not change suddenly,which proves the feasibility of the proposed algorithm applied to the manipulator.There are 55 figures,11 tables and 84 references in this paper.