RRT-A* Motion planning algorithm for non-holonomic mobile robot

The RRT algorithm can deal with the motion planning problems in consideration of non-holonomic differential constraints, but it does not taken into consideration the optimal path problem in planning process. Random selection of nodes leads to every planning cost relatively different because of only use of metric function in new node selection. In this paper, an improved heuristic RRT-A* algorithm is proposed for robot motion planning with non-holonomic constraints. In this algorithm, the cost function of A-Star(A*) is introduced into the RRT algorithm to optimize the performance. Meanwhile, several metric functions are used as the heuristic information functions respectively to measure the performance of different metric function. The simulation results shown that the Manhattan heuristic information function based RRT-A* planning algorithm is better than the other improved RRT algorithms in optimization path and computational cost.