A probability-based path planning method using fuzzy logic

In this paper, we improve the path-of-probability (POP) algorithm to generate better and smoother paths for robotic systems. The POP algorithm generates a full path with the maximum probability to reach a target. Although the POP algorithm has been successfully applied to path generation of robotic systems, it has the limit that the intermediate paths are generated using a finite and small number of candidate paths. This results in non-smooth (or zigzag pattern) paths which are not desirable for practical use. In addition, this approach sometimes fails to generate a path that reaches a target. This can be overcome by increasing the number of candidate paths, but it increases the computation time too. To improve the POP method, we use a fuzzy logic. The fuzzy logic enables us to obtain a continuous set for the intermediate paths from the discrete candidate paths. By applying it, smoother paths from the starting point to the target can be obtained directly without any additional path smoothing because the intermediate paths are chosen from a continuous set. Fundamentally, the fuzzy logic makes a mapping from a finite and discrete set of candidate paths to a continuous set of paths. We apply the proposed method to path planning for a two wheeled robot and a flexible needle. The simulation results confirm the performance of the proposed approach.