A novel identification method based on frequency response analysis

Abstract

Motivated by the fact that many classical identification methods cannot be directly used for a closed-loop control system or have to meet high requirements, a novel method is proposed for application to both open- and closed-loop identification by introducing a damping factor in the frequency response domain identification. Based on the mathematical analysis of the system, we proposed a novel particle swarm optimization (NPSO) algorithm to obtain an optimal estimate of the model by searching for the optimal damping factor. In addition, considering that many methods are confined to some special test signals as system inputs, the method presented here is proved to obtain the optimal estimated model by choosing the optimal damping factor for any continuous signals. With deep analysis and evidence, the method is improved so that the input signal can maintain output convergence and a second-order derivative in the s-domain is available. Simulations of the identification of open- and closed-loop systems are given to demonstrate the effectiveness and merits of the proposed algorithm in the paper.