Passive state estimation for fuzzy jumping neural networks with fading channels based on the hidden Markov model

• Published in: Physica A Vol. 535; p. 122437
• Main Authors: Wang, Xuelian, Xia, Jianwei, Wang, Jing, Wang, Jian, Wang, Zhen
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.12.2019
• Subjects: Fading channels; Fuzzy Markov jump neural networks; Hidden Markov model; Passive state estimation; Fading channels; Hidden Markov model; Passive state estimation; Fuzzy Markov jump neural networks
• ISSN: 0378-4371; 1873-2119
• DOI: 10.1016/j.physa.2019.122437

The passive state estimation issue of the fuzzy Markov jump neural networks with fading channels is investigated in this paper. In view of the fact that it is difficult for the estimator to completely obtain the mode information of the neural networks in the actual situations, the hidden Markov model is utilized to depict the mode mismatching phenomenon between the networks and estimator. Moreover, aimed at the circumstance that data are transmitted over fading channels, and an improved discrete-time Rice fading model with the mode-dependent channel coefficients is employed. The principal goal is to devise a state estimator which can ensure that the error system realizes the stochastic stability and satisfies the passive performance index. By establishing a suitable mode-dependent Lyapunov–Krasovskii functional, sufficient conditions guaranteeing the realization of the designed state estimator are presented. The advantages and effectiveness of the adopted design scheme are verified by an illustrated example. •A fuzzy state estimator is designed for the FMJNNs for the first time.•The HMM is used to explain the relation between the estimator and system mode.•Using the Rice fading channels, the effect of signal fading is more reasonable.