CFP RRT for an automatic robot in the narrow space and the trapped goal environment
Menghomg Bun, Cherdsak Kingkan, Warcc Kongprawechnon, Hiroki Nakahara · 2021
RRT algorithm is a popular obstacle avoidance path planning algorithm of the sampling-based algorithm to tackle the path planning issues of the robot's high degree of freedom in the high dimensional working environment. Moreover, it has been developed in many versions to improve its performance depending on the working space, but it still has been some weak performance in the trapping goal environment. To gain the efficiency of the previously improved RRT algorithm, this paper is studied a CFP-RRT algorithm by combining three different techniques of the previous algorithm. First, the changing sampling area (CSA) technique is used to reduce the number of random nodes, which are lengthier than the distance between extension nodes and a goal node. Second, the following obstacle (FO) technique is utilized to find quickly the motion planning in the trapped goal environment by the tree is grown following obstacle to encounter the goal. Third, the heuristic probabilistic biased goal (PBG) technique is used heuristic probabilistic to select the extension node equation, to avoid searching nodes for invalid areas of the target gravity equation, and changing the biased goal factor in the target gravity equation is applied to prevent the local minimum problem. Finally, the result of the proposed algorithm is evaluated and verified via the performance comparison of the previous RRT algorithm, i.e., the average number of nodes in the tree, the average number of nodes in the path, the average length of the waypoints, and the average time computation.