Explicit Nonlinear MPC of an Automotive Electromechanical Brake

Abstract: Electromechanical brakes (EMB) have great potential for automotive applications due to performance, manufacturing and environmental benefits. One key performance criterion is the ability to track brake clamp forces requested by higher level controllers such as anti-lock braking and electronic stability systems. Prior EMB controllers have utilised architectures including cascaded proportional-integral (PI) control and linear model predictive control (MPC), although only with suboptimal results. In this paper, an explicit nonlinear constrained MPC is proposed for the EMB. The explicit control law is obtained by minimising a quadratic performance criterion. The solution is computed offline and saved to memory to avoid the computational expense of online optimisation. The control law is implemented in simulation using a lookup table and its effectiveness is demonstrated. The effect of model parameter variation on control performance is discussed, and its impact on the controller implementation is investigated with a view to determining the most suitable parameters for online adaptation. Model parameter adaptation within the explicit MPC framework is also investigated. 1