Integrating topological map for autonomous navigation in unstructured environments

To address the local path planning problem for autonomous land vehicle in unstructured environments, this paper will incorporate the topological representation of the local map and a fast A* search. Topological connection of the local map will serve as a navigation function for the computing of heuristic value. To extract the topological representation, polar histogram is constructed iteratively. Outlets of temporal expanded state can be found by wave peak of polar histogram. To improve the efficiency of computing the heuristic value based on topological map, a lazy computing process is adopted. Experimental results show that the proposed method can well resolve the local path planning problem for autonomous land vehicle in unstructured environments, even if the localization of the vehicle may not be very accurate.