Hierarchical control based on SPAT unidirectional heterogeneous communication topology for intelligent connected hybrid electric vehicle platoon

The rapid development of intelligent connected vehicle technology provides new solutions for eco-driving urban transportation. To address the problem of vehicle platoons stopping and waiting at consecutive signalized intersections and to investigate the following performance and fuel economy of platoons under unidirectional heterogeneous communication topologies, this study proposes a hierarchical control strategy for intelligent connected hybrid electric vehicle (HEV) platoons based on signal phase and timing (SPAT) and unidirectional heterogeneous communication topologies. The upper-level controller establishes a target speed model based on SPAT information and solves the optimal target speed using a model predictive control (MPC) algorithm, considering the platoon’s unidirectional heterogeneous communication topology and the variation in vehicle aerodynamic drag. The lower-level controller employs a deep deterministic policy gradient (DDPG) algorithm for energy management to achieve optimal power allocation, addressing the limitation of deep Q-learning (DQN) in handling continuous high-dimensional state spaces. Simulation results demonstrate that vehicle platoons with unidirectional heterogeneous communication topologies can satisfy various traffic constraints, achieving excellent following and passing performance. Compared with DQN, the DDPG-based energy management strategy improves fuel economy by 6.45%, 6.23%, 6.19%, and 6.25% under PF, PLF, TPF, and TPLF communication topologies, respectively.