FAST MARCHING METHOD BASED PATH PLANNING FOR WHEELED MOBILE ROBOTS

The paper presents a path planning approach for wheeled mobile robots in obstructed environments. The trajectories of moving objects have to be carefully planned in order to obtain a near-shortest smooth path at still acceptable computational complexity. The combined approach is therefore proposed which utilizes search algorithm A* as well as methods of numerical solving of a particular form of partial differential equation - an eikonal equation. The use of related fast marching method enables to derive smooth trajectories within the shortest path corridor identified by the heuristic search algorithm while keeping the on-line computational burden relatively low. To illustrate the basic idea our investigation is limited to situation with static obstacles, e.g. buildings in the area which is crossed by autonomous vehicles. The proposed approach operation is validated by experimental results on a differential mobile robot.