Near optimal path planning for vehicle with heading and curvature constraints

A method of path planning is presented based on a discretization of the state space, aiming at taking into account of heading and curvature constraints for vehicles which are capable of forward and backward motion in this paper. A search space is created with the heading dimension appended to the grid. Motion coding is achieved efficiently through the connections of states using arcs and Bezier curves. Therefore, an Extended Control Set (ECS) is attained and the feasibility of paths in ECS is also proved out. Vehicle path planning in the search space can be transformed into heuristic search in which ECS is used. The use of Look-Up Table (LUT) has accelerated the search efficiency of the A* algorithm. Simulation results show that the proposed method can successfully generate near optimal paths without post-processing, and they satisfy the requirements of constraints on heading and curvature.