Ding Sanbo, Wang Yong, Xie Xiangpeng
IEEE Trans Cybern. 2023 Aug;53(8):5380-5386. doi: 10.1109/TCYB.2021.3131475. Epub 2023 Jul 18.
This article investigates the event-triggered synchronization control problem of discrete-time neural networks (DNNs) in the case of periodic sampled-data. A discrete-time periodic event-triggered mechanism is adopted to evaluate the measurements, which avoids formulating the triggering function in a continuous manner and saves energy consumption. Under this framework, an event-triggered dynamic output-feedback controller is designed to achieve the goal of synchronization. A piecewise Lyapunov functional is constructed to analyze the sawtooth-like pattern of sampled-error signals. Thereafter, the synchronization criteria are formulated for the considered DNNs. The co-designed issue is further discussed for the control gains and triggering parameter. Finally, a simulation example is presented to show the effectiveness of the proposed method.
本文研究了周期采样数据情况下离散时间神经网络（DNN）的事件触发同步控制问题。采用离散时间周期事件触发机制来评估测量值，这避免了以连续方式制定触发函数并节省了能量消耗。在此框架下，设计了一种事件触发动态输出反馈控制器以实现同步目标。构造了一个分段李雅普诺夫泛函来分析采样误差信号的锯齿状模式。此后，为所考虑的DNN制定了同步准则。进一步讨论了控制增益和触发参数的协同设计问题。最后，给出了一个仿真例子以说明所提方法的有效性。
