Parametric Continuous-Time Blind System Identification

In this paper, the blind system identification problem for continuous-time systems is considered. A direct continuous-time estimator is proposed by utilising a state-variable-filter least squares approach. In the proposed method, coupled terms between the numerator polynomial of the system and input parameters appear in the parameter vector which are subsequently separated using a rank-1 approximation. An algorithm is then provided for the direct identification of a single-input single-output linear time-invariant continuous-time system which is shown to satisfy the property of correctness under some mild conditions. Monte Carlo simulations demonstrate the performance of the algorithm and verify that a model and input signal can be estimated to a proportion of their true values.