Closed loop optimal experiment design for on-line parameter estimation

This paper focuses on the problem of closed loop on-line parameter identification for dynamic systems. An approach for the combined on-line optimal experiment design and model parameter identification is presented. Based on the observation theory and the model based predictive control theory, this approach aims to solve an optimal constrained control problem. During the designed experiment, the optimal time-varying input applied is computed at each current time to maximize the sensitivities of the model outputs with respect to the unknown model parameters which are also estimated on-line. The approach does not require to measure all the process state. Moreover constraints may be specified to maintain the system behavior in a prescribed region. A case study of chemical process is used to illustrate the developed approach.