Hierarchical energy management strategy based on adaptive dynamic programming for hybrid electric vehicles in car-following scenarios
Wenbin Tang, Yaqian Wang, Xiaohong Jiao, Lina Ren
Abstract
In order to improve the driving safety and fuel economy of hybrid electric vehicles (HEVs) in car-following scenarios, this paper proposes a hierarchical energy management strategy (EMS) based on adaptive dynamic programming (ADP). In the upper-layer speed planning, the speed prediction model of the preceding vehicle is established utilizing a back-propagation neural network (BPNN), and the host vehicle speed is planned based on heuristic dynamic programming (HDP) real-time through the predicted preceding vehicle speed and the current state of the host vehicle. In the lower-layer energy management control, the EMS based on dual heuristic dynamic programming (DHP) is used to realize the power allocation of each power source according to the planned speed to reduce fuel consumption. In both HDP and DHP algorithms, BPNN is used to implement the critic network (CN) and action network (AN) to obtain a more accurate control variable and faster convergence speed. By comparing the results with various existing EMSs in the different driving cycles, the simulation results show that the proposed EMS ensures real-time planning speed and good following performance while reducing energy consumption.