Iterative Identification for A Class of Closed-loop Systems Based on A Greedy Algorithm

A compressive sampling matching pursuit (CoSaMP) iterative algorithm is proposed in this paper to identify parameters and time-delays of a class of closed-loop systems where the forward channel is a CARMA model. Due to the unknown time-delays of both the feedback controller and the controlled plant, a high dimensional identification model with a sparse parameter vector is derived by using an overparameterized method. Then combining the CoSaMP algorithm with the iterative idea, the parameter vector is estimated and the unmeasurable noise items are updated in each iteration. Finally, the parameters of the feedback controller are extracted based on the model equivalence principle and time-delays are estimated according to the sparse characteristic of the parameter vector. The proposed method can simultaneously estimate the parameters and time-delays from a small number of sampled data. The simulation results illustrate that the proposed algorithm is effective.