Abstract: | This dissertation describes the path planning/control method of a car-like mobile robot. The proposed two-stage path planner consists of the Global path planner and the Local path planner. The Global path planner finds collision-free paths from an environmental map, which features universal consideration of the topological configuration of all obstacles by the method of Maklink Graph. The Local path planner/controller linearises the robot dynamics by using the chained form and accomplishes a closed-loop control. This satisfies the non-holonomic constraints anad obtains robustness toward disturbances. Obstacle avoidance experiments show that the proposed method can successfully plan the collision-free path in a cluttered environment and navigate the robot to the goal with high trajectory accuracy.
|