Cyclic Spectral Based Approach to Closed-Loop Identification
Abstract
The direct closed-loop identification algorithm based on cyclic spectral is considered in frequency domain in this paper. It is illustrated that the input and output signals obtained through inter-sampling technique have cyclic spectral, and the cyclic spectral contain plant model information, then the plant model can be estimated just from the cyclic spectral of the plant input and output signals. It is clarified that the test signal is not necessary, and the new algorithm does not depend on the controller structure. The effectiveness of the proposed algorithm is also demonstrated through some numerical examples.