Real-Time Optimized Model Predictive Control of an Active Roll Stabilization System with Actuator Limitations

Abstract

Active roll stabilization systems are used to improve both the ride dynamics and ride comfort. For that, the measureable information about road disturbance should be used to calculate the control variable for the actuators at the front and rear axle. Even without previewed disturbance information the whole car dynamics can be modelled and provide future states of the controlled system which consequently can be regarded in the calculation in advance. By the framework of a Model Predictive Control, the actuator limitations can be included. Additionally, the movement of the car body and each wheel is regarded, so that an optimal allocation of the control variables on both actuators takes place. With both aspects, namely the actuator limitations and the optimization itself, a high potential for ride comfort improvement is generated.