混合时滞复值神经网络的事件触发状态估计北大核心CSCD

研究了事件触发机制下混合时滞复值神经网络的状态估计问题.首先基于测量输出设计了事件触发机制,有效降低了估计器更新的频率.在触发机制中引入了等待时间,以此避免了采样中的Zeno现象.运用Lyapunov方法和复值矩阵的性质,建立了估计误差系统全局渐近稳定的充分性判据,并基于线性矩阵不等式技巧给出了复值增益矩阵K的求解算法.最后的数值例子验证了理论成果的正确性和有效性.

Event-Based State Estimation of Complex-Valued Neural Networks With Mixed Delays

The event-based state estimation problem was investigated for a class of complex-valued neural networks with mixed delays. Based on the measurement output, a novel event-triggering scheme was introduced to reduce the frequency of updating while ensuring the estimation performance. A waiting time was first employed to avoid the Zeno phenomenon. By means of the Lyapunov direct method and some properties of complex-valued matrices, a sufficient criterion was established to guarantee the globally asymptotic stability for the error system. The weighted parameters and gain matrices were designed with resort to the feasible solution of matrix inequalities. A numerical simulation example illustrates the effectiveness of the proposed method.