A Hybrid Modified PSO System Identification Method Based on the Asynchronous Time-Dependent Learning Factor

摘要

In this paper, the system identification method to Hammerstein model is studied. Considered that the identification accuracy of the standard particle swarm optimization (PSO) is limited and the local optimal problem is easily occurred at later stage, the standard PSO and its initial value setting is firstly discussed. Then, a modified PSO combined with the methods of asynchronous time-varying learning factor and linearly decreasing time-varying weight is put forward to obtain the optimal solution in the whole parameter space. Finally, the comparison experiments are done to verify the accuracy and the advantage of noise resistance of the proposed method.