A motion strategy for a mobile robot with holonomic and nonholonomic constraints

Presented is a novel motion strategy for a mobile, car like robot that is subject to kinematic constraints. The algorithm operates on the original obstacles, without needing to generate the configuration space obstacles for the dimensioned robot. The path for the dimensioned robot is generated in three stages: (i) the shortest path problem for a point robot is solved; (ii) free space relative to the point robot shortest path is locally evaluated by minimum distance computations; (iii) the point robot shortest path is locally modified to account for the size and kinematic constraints of the mobile robot. If the shortest point robot path fails to be modified into a feasible path for the robot, the process is repeated with a second candidate point robot path, and SO on until a feasible path is generated. Thus the proposed strategy combines a global scheme for point robot path generation with a local scheme for free space evaluation and point robot path modification. The algorithm is computationally efficient, being of computational time O(nk +nlogn) where n is the total number of vertices, including the two ends, and k is the number of obstacles. The algorithm has been tested in computer simulations, demonstrating its ability to automatically generate paths which may include reversals.