Model Predictive Control of Magnetic Automotive Actuators

Magnetically actuated mass-spring-damper systems are common in automotive systems as components of various actuation mechanisms. They are characterized by nonlinear dynamics, tight performance specifications and physical constraints. Due to these reasons, model predictive control (MPC) is an appealing control framework for such systems. In this paper we describe different MPC approaches to control the magnetically actuated mass-spring-damper system. The MPC controller based on the complete system model achieves very good performance, yet it may be too complex to be implemented in standard automotive microcontrollers. Hence, we consider the possibility of decoupling the electromagnetic subsystem from the mechanical subsystem, assuming that the electromagnetic dynamics, controlled by an inner-loop controller, are much faster than the mechanical dynamics. Based on a previous feasibility study, we implement a control architecture in which the MPC optimizes only the dynamics of the mechanical subsystem, and we test it in closed-loop simulations with the nonlinear system. The resulting control system achieves lower performance, but it is simple enough to be implemented in an automotive microcontroller.