Research On Path Planning Of Catenary Insulator Cleaning Robot

Catenary insulator is an important part of electrified railway.Due to its long-term exposure to the outdoors,the surface is easy to gather dust and other dirt particles,which weakens the electrical characteristics,causes pollution flashover accidents and endangers the safety of train operation.Therefore,the regular cleaning of contact line insulator is indispensable.However,the traditional manual cleaning method is inefficient and can not meet the needs of insulator cleaning task in modern railway system.The research of robot automatic cleaning technology is imperative.Path planning is the key to the cleaning task of insulator cleaning robot.The path planning of insulator cleaning robot includes two-dimensional plane path planning of intelligent mobile chassis and three-dimensional path planning of robot manipulator.The fast extended random tree algorithm based on sampling collects the obstacle information through collision detection,avoids the modeling of the moving space environment,and is suitable for high-dimensional space path planning,which is the basic algorithm of this paper.Firstly,the robot for insulator cleaning of overhead contact system is introduced.Based on the D-H method,the structural modeling of the manipulator arm is analyzed.The mathematical expression of position and altitude of the manipulator and the transformation of the linkage coordinate system are described.The forward and inverse kinematics equations of the robot are derived and verified.Secondly,in the process of robot path planning,three kinds of common space geometry envelope models are introduced.By simplifying the robot and environmental obstacles,the problem of collision detection between each joint of robot manipulator and obstacles is transformed into the problem of position relationship judgment between linear segment and space geometry,which is analyzed and solved by using the knowledge of space geometry solution.Thirdly,aiming at the shortcomings of the basic Rapidly-exploring random tree algorithm,an improved sampling algorithm based on bi-directional asymptotically optimal fast extended random tree(RRT*-Connect)is proposed.Combined with the advantages of Artificial potential field algorithm in path planning,through intelligent sampling of random state points,potential field guidance is added in the process of sampling point generation,and the path starting point and target point are used as the attraction poles to expand the random tree alternately,which makes the optimization directional,reduces redundant search,and improves the path generation efficiency.In the process of selecting and rewiring the parent node,the previous points on the random tree of the adjacent points of new node are added,and the connection cost is optimized by using the trigonometric inequality.Finally,the improved RRT*-Connect algorithm is simulated and verified in Matlab platform.By comparing and analyzing the simulation results of the proposed algorithm and the basic RRT* and RRT*-Connect algorithm,the improved RRT*-Connect algorithm is superior to the basic RRT* and RRT*-Connect algorithm in the quality of planning path,which proves the efficiency and rationality of the improved RRT*-Connect algorithm.