OGBPS: Orientation And Gradient Based Path Smoothing Algorithm For Various Robot Path Planners

Abstract

It is significant to plan a smooth trajectory for high-speed wheeled mobile robots working in a cluttered environment. The trajectory generated by most of the previously proposed path planners is not smooth enough for robot motion, especially under kino-dynamic constraints. An improved smoothing algorithm is proposed in this work as a solution for most of the previously proposed path planners to deal with the rugged paths, which may cause abrupt and angular turns of robots. The improved solution we proposed could be applied to many mainstream path planners (like Theta*, A*, RRT, RRT*, RRT#, SORRT*, PRM) as a post-smoothing algorithm, which is called orientation and gradient-based path smoothing (OGBPS). The OGBPS algorithm is derived from both orientation-angle-based and gradient-based path deformations to obtain a high-quality path. The objective of path deformations in this work is to improve path smoothness, lower maximum curvature and path length. Sufficient simulation experiments are well conducted to demonstrate the effectiveness of our approach. It is verified that the proposed algorithm can improve the quality of the previous path while respecting the kino-dynamic constraints through experiments. The simulation results indicate the advantages (smaller maximum curvature and smaller path length) of the proposed algorithm compared with several state-of-the-art smoothing algorithms.