Release Time:2021-06-01  Hits:

Indexed by: Conference Paper

Date of Publication: 2021-04-26

Page Number: 1337-1342

Key Words: switched nonlinear systems; EM algorithm; weighed multi-innovation least squares; key-term separation principle; Hammerstein models

Abstract: This study aims to determine how to deal with the identification of switched nonlinear systems (SNSs) with multiple Hammerstein models. The identification tasks of SNSs are model detection (MD) and parameters identification (PI). At the stage of MD, Expectation-Maximization (EM) algorithm is used to confirm the switched time and the dwell time of SNSs. At the stage of PI, the weighted multi-innovation least square (WMILS) algorithm is proposed to obtain the final parameter estimates for each subsystem. Finally, the effectiveness of the proposed methods is verified through simulation example.