Horizon-1 Predictive Control of Automotive Electromagnetic Actuators

Electromagnetically driven mechanical systems are characterized by fast nonlinear dynamics that are subject to physical and performance constraints, which makes controller design a challenging problem. Although model predictive control (MPC) is well suited for dealing with constraints, the fast dynamics of electromagnetic (EM) actuators render most standard MPC approaches impractical. This paper proposes a horizon-1 MPC strategy that can handle both the state/input constraints and the computational complexity limitations associated with EM actuator applications. A flexible Lyapunov function is employed to obtain a nonconservative stability guarantee for the horizon-1 MPC scheme. Moreover, an invariant region of attraction is provided for the closed-loop MPC system. The simulation results obtained on a validated model of an EM engine valve actuator show that performance is improved with respect to previous strategies, and that the proposed algorithm can run within a sampling period in the order of a millisecond.