Multi-dimensional Attitude Cooperative Hierarchical Control Considering Zero Moment Point

Abstract

Vehicle ride comfort and handling stability constitute fundamental technical indicators in vehicle dynamics. Active roll control, as a pivotal technology for enhancing vehicle lateral dynamic performance, has demonstrated efficacy under steady-state steering conditions. However, under complex driving scenarios, it encounters significant challenges including response lag in high-frequency dynamic conditions and the absence of a unified stability evaluation framework. Furthermore, during complex maneuvering operations, competitive conflicts emerge among multiple attitude control objectives including roll control, pitch control, and ride height adjustment, resulting in suboptimal actuator resource allocation and potentially precipitating control failure. To address these multidimensional attitude coordination challenges, this paper proposes a Zero-moment point based multidimensional Attitude Coordinated Control methodology (ZACC). This approach extends ZMP theory from robotics to vehicle dynamics, establishing a unified vehicle stability evaluation framework that achieves coordinated optimization of roll, pitch, and ride height attitudes. The methodology incorporates a comprehensive objective function encompassing comfort and stability criteria, employs enhanced deep reinforcement learning for intelligent actuator resource allocation, and implements model predictive control based on an 11-degree-of-freedom vehicle dynamic model. Simulation results demonstrate that ZACC significantly enhances ride comfort and handling performance while maintaining vehicle stability.