An Autonomous Vehicle Motion Planning Method based On Dynamic Programming

In this article, we propose a new approach to solve local motion planning problem in autonomous vehicle field. Instead of generating several primitives and choosing the optimal trajectory from that, we get the future T time step horizon's vehicle state, and fit that into clothoid local trajectory by solving a dynamic programming problem. The trajectory generated is smooth enough and meet both ego vehicle's dynamics and traffic regulation's constraints. Iterate this process, the ego autonomous vehicle can obtain its future trajectory continuously. This method has a good performance in the simulation using Matlab and Prescan. This method has fast computing speed to apply on autonomous vehicle driving in real road online compared to other methods.