Название: Stability Analysis and State Estimation of Memristive Neural Networks
Автор: Hongjian Liu, Zidong Wang, Lifeng Ma
Издательство: ‎CRC Press
Год: 2021
Страниц: 237
Язык: английский
Формат: pdf (true)
Размер: 10.48 MB

In this book, the stability analysis and estimator design problems are discussed for delayed discrete-time memristive neural networks. In each chapter, the analysis problems are firstly considered, where the stability, synchronization and other performances (e.g., robustness, disturbances attenuation level) are investigated within a unified theoretical framework. In this stage, some novel notions are put forward to reflect the engineering practice. Then, the estimator design issues are discussed where sufficient conditions are derived to ensure the existence of the desired estimators with guaranteed performances. Finally, the theories and techniques developed in previous parts are applied to deal with some issues in several emerging research areas.

In the context of neural networks, synapses are essential elements for computation and information storage, which needs to remember its past dynamical history, store a continuous set of states, and be “plastic” according to the synaptic neuronal activity. All these cannot be accomplished by a resistor in traditional recurrent neural networks (RNNs). When the resistors are replaced by the memristors, the resulting memristive neural networks (MNNs) could rather completely solve these problems. Meanwhile, the implemented MNNs could be more efficient than the traditional RNNs when applied in brain emulation, combinatorial optimization, knowledge acquisition, and pattern recognition. As such, the dynamics analysis problems, such as stability and synchronization for MNNs, have recently received considerable research attention and a rich body of relevant literature has been available for different kinds of MNNs. It should be mentioned that almost all results obtained so far have been exclusively for continuous time MNNs and the corresponding results on discrete-time memristive neural networks (DMNNs) have been much fewer.

The book:

• Unifies existing and emerging concepts concerning delayed discrete memristive neural networks with an emphasis on a variety of network-induced phenomena
  
• Captures recent advances of theories, techniques, and applications of delayed discrete memristive neural networks from a network-oriented perspective
  
• Provides a series of latest results in two popular yet interrelated areas, stability analysis and state estimation of neural networks
  
• Exploits a unified framework for analysis and synthesis by designing new tools and techniques in combination with conventional theories of systems science, control engineering and signal processing
  
• Gives simulation examples in each chapter to reflect the engineering practice