Overtaking or Merging? Eco-Routing Decision and Speed Trajectory with Full Terrain Information 2018-01-0038
With vehicle platooning becoming an important research field in recent years, it is now imperative to introduce platoons as part of the dynamic environment, considering overtaking and merging possibilities. This article studies optimal speed trajectories and longitudinal control with optimized energy efficiency for an autonomous vehicle with several preceding platoons and full terrain information. It aims at improving the energy efficiency of vehicles with Advanced Driver Assistance Systems (ADAS). A forward discrete dynamic programming (DDP) algorithm with distance as the discretization basis is used to derive speed trajectories in the trade-off between air drag reduction and energy saved by utilizing the road slope information. The problem is decomposed into decisions whether to overtake or to merge into the nearest platoon with the assumption of sufficient distance among platoons. During the process, speed choices and cost function reflect interactions among the controlled vehicle, the platoons, and the road. Simulations confirm that the energy consumed can be reduced significantly when tracking optimal trajectories compared to driving by the existing model predictive control (MPC) tracking strategy and linear quadratic regulator (LQR). For a trip of 500 m, if the primary velocity of the ego vehicle is 7.5 m/s and the initial state of a preceding platoon is 10 m from the ego car, 8 m/s, the energy cost saving of the proposed solution can be up to 3.45% compared by LQR. When there are two platoons whose original states are 10 m, 7 m/s, and 35 m, 7.5 m/s respectively, the energy saving of the ego car with initial velocity of 8 m/s is 3.07% compared by LQR and 1.99% by MPC. Simulations indicate the potential of energy efficiency of the proposed method for further studies with more sophisticated conditions. Sensitivity analysis has shown that the energy saving is not sensitive to the initial speed of the ego car when it is relatively low.