Closed-loop subspace Predictive Control for Linear Parameter Varying systems (i) - the nominal case

The paper presents a new data driven predictive control approach for a special set of nonlinear systems. A Linear Parameter Varying (LPV) subspace based identification technique is combined with predictive control approach without computing the parameter dependent state space matrices. Therefore, the Subspace based Predictive Control for LPV systems (SPC LPV) is a candidate for joint nonlinear identification and predictive control as a model independent technique. Based on an identified nonlinear input/output predictor, the SPC LPV algorithm formulates a (constrained) optimal and predictive control problem without the explicit knowledge of the model parameters. Finally, a nonlinear system is controlled by an input/output based optimal control law. The proposed approach is applied on a real time environment on a DC motor.