Very fast algorithm for nonholonomic optimal path planning and replanning, generalities and details

This paper presents an original method for nonholonomic mobile robot path planning. The method is based on A* searching of a graph built in a configuration space. Very good heuristic approximation of the real cost of the path accelerates path searching significantly. Due to dynamic modifications of the graph, the path replanning procedure is much faster than brute-force replanner. The paper presents general assumptions of the method, detailed algorithms of important procedures (especially the path replanning procedure) and a new form of configuration space and open list. The simulation and experimental tests on car-like robot as well as the comparisons of this method to the others are also included.