Mobile robot navigation system based on Probabilistic Road Map (PRM) with Halton sampling of configuration space

The paper uses Probabilistic Road Map (PRM) based path planning algorithm which composes Halton point sets to improve mobile robot navigation capabilities. The Halton set obtains a good coverage of robot environment with points that is better than using grid based methods. Within PRM learning phase Halton point sets are used to randomly generate the robot configurations which constitute probabilistic road maps. The shortest path between an initial and a final configuration was founded out using A* algorithm adopted for query phase of PRM. The influence of number of Halton points and a distance between adjacent nodes for the current configuration to the path planning are analyzed. For purpose of map building of an unstructured environment a new histogramic based method is applied. In order to implement PRM based planning algorithm the whole navigation system is designed and implemented. The effectiveness of the proposed navigation system was demonstrated in both simulation and experimental modes.