MPC based optimal input design for nonlinear system identification

A combined nonlinear model predictive control with extended Kalman filter strategy has been proposed for optimal input design. As the designed controller depend on the identified parameters, the achievable performance highly depends on the quality of the identified information. The degradation in achieving the desired control performance is quantified b y introducing an optimality criteria which minimize the error covariance matrix of the identified parameters. The major contribution is using the information of the system parameter at every sample time to improve the control performance at next time step. The the performance of the proposed algorithm is verified by numerical simulations for a example system.