Blind Identification of Weak Nonlinear System Based on Normalized Kurtosis

This paper studies normalized kurtosis and its application in weak nonlinear system identification. According to the definition of normalized kurtosis and its properties, the memory effect and nonlinear order influences on normalized kurtosis are derived theoretically, and simulation results are given. Accordingly, this paper proposes a method in weak nonlinear system identification by using normalized kurtosis variation with varying system characteristics. Finally, this paper describes the potential advantage of proposed method.