Research And Application Of Path Planning Approach For Mobile Robot Autonomous Navigation

As a master of current science and technology,mobile robot integrates advanced technology of electronic information,computer science and other disciplines.Its software system is composed of perception system,decision and planning system,control system and other modules.One of the key and difficult technologies is decision and planning.This paper first introduces an autonomous mapping method combining with end-to-end path planning technology,and improves the end-to-end path planning algorithm,and finally verifies its feasibility through simulation experiments.For the cleaning robot required by the project,this paper proposes a bidirectional full ergodic path planning method based on backtracking method,which can fully cover the free area in the workspace.Grid method is used to build global map for mobile robot.The boundary detector of the improved adaptive step size rapid exploring random tree(RRT)algorithm is used to detect the boundary of the unknown region.The optimal boundary points obtained after screening are assigned to the robot by the task allocator.The robot takes them as the target points,and uses the RRT* algorithm based on the target to replace the A* algorithm,Planning a collision free path to move to the corresponding target point,combining with dynamic window approach(DWA)to avoid obstacles.When the mapping is completed,the robot diameter is taken as the size of a grid,and the map is further divided into a new grid map.Taking the current position of the robot as the starting point,according to the approximate aspect ratio of the obtained map,the appropriate traversal order is selected,and the comb traversal method is used to cover the workspace.When the robot moves to the dead zone,the best backtracking point is selected from the list of backtracking points,and the OHA*(orientation heuristic A*)algorithm is used to plan a transfer path from the dead zone to the backtracking point,Starting from the backtracking point,a new traversal motion is started until the backtracking point list is empty and the full coverage is completed.Finally,a simulation room and a lidar simulation robot are built on the ROS platform to simulate the improved scheme.The simulation results show that the method can complete the detection of the whole workspace,and the improved scheme is feasible.In order to get the average performance data,the generated random map is used to do several experiments on the complete traversal work.The experiments show that the bidirectional complete traversal path planning algorithm based on Backtracking has certain advantages in planning speed,total length of coverage path and coverage rate.