Data-driven predictive control of idle speed control for SI engine

As we know the automotive idle speed control (ISC) is a highly nonlinear vehicle control problem with complicated dynamic characters. In this paper, the control variables are throttle angle and spark advance angle, data-driven predictive control is selected to design the controller which elegantly combines identification method of subspace with model predictive control(MPC). In automobile, there are actuator constraints on throttle and spark ignition and the optimal object is to make the predictive engine speed tacking with the reference. The predictive model can be obtained by subspace identification by using intput-output data, by using MPC, control problem comes down to optimization problem with input and state constraints. The proposed control method in engine is well validated on a platform of enDYNA. Finally, there are effective and promising results which indicate that the selected method is feasible.