Deep reinforcement learning based direct torque control strategy for distributed drive electric vehicles considering active safety and energy saving performance

Distributed drive electric vehicles are regarded as a broadly promising transportation tool owing to their convenience and maneuverability. However, reasonable efficient allocation of torque demand four wheels is challenging task. In this paper, deep reinforcement learning-based distribution strategy proposed guarantee the active safety energy conservation. The task explicitly formulated Markov decision process, in which vehicle dynamic characteristics can be approximated. actor-critic networks utilized approximate action value policy functions for better control effect. To continuous output further stabilize learning twin delayed deterministic gradient algorithm deployed. motor efficiency incorporated into cumulative reward reduce consumption. results double lane change demonstrate that handling stability performance. addition, it improve transient response eliminate static deviation step steering maneuver test. For typical maneuvers, direct significantly improves average reduces loss by 5.25%–10.51%. Finally, hardware-in-loop experiment was implemented validate real-time executability strategy. This study provides foundation practical application intelligent algorithms future vehicles.