Coupled Control of Preview Active Suspension and Longitudinal Dynamics for Autonomous Vehicle

Autonomous vehicles (AVs) equipped with an array of advanced sensors gather road preview information, presenting new opportunities to enhance ride comfort. To simultaneously improve both the vertical and longitudinal ride comfort of vehicles, a dual timescale model predictive control (MPC) preview active suspension system (ASS) and longitudinal dynamics coupled controller is developed. On a short time scale, the coupling control of the vehicle’s ASS and longitudinal acceleration is achieved using road preview information, enhancing both vertical and longitudinal ride comfort, thereby improving response speed. On a longer time scale, road prediction information obtained via Gaussian processes (GPs) is utilized for vehicle speed planning, aiming to mitigate vertical excitations caused by road profile variations while minimizing frequent speed changes. However, when road preview information is continuously used as disturbance predictions in MPC, it undermines the recursive feasibility and stability of MPC. To address this, a scaling method is devised to account for disturbances incorporated into the predictive model. Theoretical foundations ensure both recursive feasibility and asymptotic stability. The effectiveness and advantages of the dual timescale MPC preview active suspension and longitudinal dynamics coupled control are validated through simulations and bench tests.