Adaptive Periodic Event-Triggered Stabilization of Switched Neural Networks Under the Merging Signal Scheme
针对切换神经网络，提出一种自适应周期事件触发机制，通过合并信号统一分析异步与同步切换下的指数镇定问题，并构建含环状泛函的Lyapunov泛函降低保守性。
This article is concerned with the exponential stabilization of switched neural networks (SNNs) with asynchronous switching. To save the limited bandwidth effectively, an adaptive periodic event-triggered mechanism (APETM) with a novel adaptive rule is excogitated, in which the threshold function is updated at each sampling instant to quickly respond to system changes and a tuning parameter is introduced to increase the adjustable range of the threshold function. A merging signal is constructed and then the closed-loop system is established to carry out stability analysis for asynchronous and synchronous switching cases within a unified framework. Then, based on the merging signal, a corresponding Lyapunov functional that includes a looped functional is constructed. This looped functional helps to reduce conservatism of the stability criterion. Finally, examples, including relevant discussions and comparisons, are shown to illustrate the efficacy of our developed results.