An Extended Synchronization Method to Identify Slowly Time-Varying Parameters in Nonlinear Systems

作者：Zhang, Xiaoxu; Xu, Jian

来源：IEEE Transactions on Signal Processing, 2018, 66(2): 428-437.

DOI：10.1109/TSP.2017.2770092

• Parameter estimation
• online identification
• synchronization observer
• nonlinear system
• time-varying parameter

A synchronization observer is proposed to identify any given nonlinear systems with slowly time-varying parameters. In the process of parameter identification, the existing synchronization method (SM) provides the parameter estimations continuously in a small updating rate to obtain two state memories. These memories are then utilized to impulsively correct the parameter estimation in a narrow time The state memories also provide a function of noise resistance. With these modifications, the SM is extended to be called an extended synchronization method (ESM). We prove that such parameter identification process provided by the ESM is convergent in terms of the Lyapunov stability and the contractivity analysis. As efficiency demonstration of the ESM, we consider the Lorenz chaotic system and the high-dimensional Chua's system as two examples to online identify the time-varying parameters polluted by Gaussian white noise. The results show that the ESM not only enhances parameter tracking ability and noise resistance, but also gives parameter estimation more accurate than the traditional SM. The conclusions also indicate a bright prospect of employing the ESM to relative problems in nature and engineering.

• 出版日期2018-1-15
• 单位同济大学