A mode-dependent approach to state estimation of recurrent neural networks with Markovian jumping parameters and mixed delays. Academic Article

abstract

• This paper is concerned with the problem of state estimation of recurrent neural networks with Markovian jumping parameters and mixed delays. A mode-dependent approach is proposed by constructing a novel Lyapunov functional, where some terms involving triple or quadruple integrals are taken into account. The advantage is that as many as possible of the Lyapunov matrices are chosen to be mode-dependent. Several design criteria are established under which the estimation error system is globally exponentially stable in the mean square sense. The gain matrices of the state estimator can be then found by solving a set of coupled linear matrix inequalities. It is shown in theory that better performance can be achieved by this approach. Furthermore, by introducing some scaling parameters, this approach is effectively employed to deal with the state estimation problem of the neural networks with complex dynamic behaviors, to which some existing results are not applicable.

authors

• Huang, Tingwen

published proceedings

• Neural Netw

author list (cited authors)

• Huang, H. e., Huang, T., & Chen, X.

citation count

• 68

complete list of authors

• Huang, He||Huang, Tingwen||Chen, Xiaoping

publication date

• October 2013

publisher

• Elsevier  Publisher

published in

• Neural Networks  Journal

keywords

• Algorithms
• Computer Simulation
• High-order Integral Terms
• Markov Chains
• Markovian Jumping Parameters
• Mode-dependent Criteria
• Neural Networks, Computer
• Recurrent Neural Networks
• State Estimation
• Time Delays
• Time Factors

PubMed Central ID

• 23708670

Digital Object Identifier (DOI)

• 10.1016/j.neunet.2013.04.014

start page

• 50

end page

• 61

volume

• 46

URL

• http%3A%2F%2Fdx.doi.org%2F10.1016%2Fj.neunet.2013.04.014