Adaptive MPC Control for In-Wheel-Motor-Drive Electric Vehicles With Semi-Active Suspension Based on Binocular Vision Road Identification

This article presents an adaptive model predictive control (AMPC) approach to address increased body vibration resulting from unbalanced electromagnetic forces for semi-active suspension (SAS) systems in-wheel-motor-drive electric vehicles (IWMD EVs). In particular, the proposed framework consists of three key steps. Firstly, the Magneto-Rheological (MR) shock absorber characteristics have been analyzed and compared with the simulation and bench testing. Then, the dynamic model of a half-vehicle SAS system with IWMD EVs is obtained. Secondly, an AMPC preview control framework is formulated based on the road roughness and distance information obtained from binocular vision. The size of the control step is dynamically adjusted based on road preview data, and the weight matrix within the objective function is adaptively modified through the genetic algorithm according to the acceleration of the mass sprung by the suspension. This SAS control system enables optimal suspension adjustments in various road conditions. Simulation and hardware-in-the-loop experiment studies verify that the proposed framework performs better than existing methods in typical road conditions. The proposed AMPC control strategy significantly enhances ride comfort in IWMD EVs.