Design and implementation of a trajectory planning algorithm using Dynamic Programming method on a Hybrid Electric Vehicle

ABSTRACT In the recent years the automotive industry is in a revolution towards the electrified and intelligence mobility. The autonomous driving, which is the possible solution for the traffic congestion, accidents, and emission issues, is quickly becoming one of the hottest topics in the automotive field. However, the technology is still not mature enough for commercialization due to several technical challenges. The trajectory planning is one of them. Many researches regarding the trajectory planning have already been developed in the robotic field, however when it comes to the automotive field, it becomes extremely challenging due to the critical requirements on the system’s real-time performance and on the trajectory generation with multi-objective optimization(Collision avoidance, occupant’s comfort, fuel economy, etc). The two targets are usually contradictory since the planning algorithms with better optimization performance will degrade the real-time performance. In this thesis an approach is proposed to solve this problem. First an off-line optimization based on Dynamic Programming method is implemented on a Hybrid Electric Vehicle to obtain the optimum control sequence considering multi-optimization targets. Then a rule-based planner is established based on the results of the off-line optimization to realize the real-time application. Finally, the results obtained from different algorithm are compared (PID, MPC, DP, rule-based) and it proved that this approach could be a feasible solution for the contradictory requirements on real-time and optimality of the trajectory planning algorithm.