複数個の未知ジャンプパラメータを含む非線形系の識別同定法

DOI Web Site Web Site

書誌事項

タイトル別名
  • A Recognition and Identification Method for the Nonlinear System Containing Unknown Jump Parameters
  • フクスウコ ノ ミチ ジャンプ パラメータ オ フクム ヒセンケイケイ ノ シ

この論文をさがす

抄録

In this paper, a recognition-identification method is proposed for the nonlinear systems of which the dynamics have unknown pattern variations. Such systems are formulated by considering the parameters in the system configuration as unknown jump parameters, and the identification problem can be reduced to the estimation problem of the augmented system containing these parameters in the state vector.<br>Using the parallel processing of extended Kalman filter and the generalized likelihood ratio test based on the state observations, we can effectively detect the time at which a parameter jumps, recognize the parameter, estimate the parameter and, of course, estimate the state. Especially in our method, a sensitive filter algorithm and an iterative scheme are used in an extended Kalman filter to improve the estimation accuracy of the jump quantities.<br>This method is developed from the failuer detection method considered by C. N. Shen, A. S. Willsky, et al. for linear systems. We apply their method to nonlinear systems. This method may be applicable also to the system with more varieties of patterns than those of the jump type by introducing the multi-hypotheses. Satisfactory results are obtained in the numerical study for the 2nd order system with a jump type pole and jump type gain parameters.

収録刊行物

詳細情報 詳細情報について

問題の指摘

ページトップへ