Motion planning using cooperative perception on urban road

In this paper, we consider motion planning with long-range sensing information provided by cooperative perception. Firstly, we develop a general framework to reflect sensing uncertainty and transmission delay into motion planning. The Bayesian filter is utilized for perception belief fusion, which is then formulated into a cost function for optimal planning. With the cost map, we leverage the optimal property of RRT* framework and propose a long-term perspective planning algorithm to exploit the benefits introduced by long-range sensing. Finally, we demonstrate our proposed methods for a self-driving vehicle featured with cooperative perception. The experiment result shows that the proposed approach is able to improve the planning performance and is applicable to real-time implementation.