Skip to Main Content
A unified approach is developed for identification of linear time-invariant systems. It is shown that, given the experimental frequency-response data of the system, the plant can be identified using a simple, numerically reliable algorithm. Further, an error bound is derived for exponentially stable systems when the frequency-response data are corrupted by bounded noise. An example is presented to illustrate the proposed algorithm.