This paper addresses the issue of ${\mathcal{H}}_{\infty }$ synchronization of switched neural networks via an observer-based sliding mode control scheme. With regard to the switched neural networks, the persistent dwell-time switching law is introduced to govern the switchings among subsystems, which is a more general switching law. Furthermore, in view of the difficulty in the measurement of most disturbances in various applications, a disturbance observer is constructed. Subsequently, an integral sliding surface and a sliding mode control law are constructed based on the disturbance observer states to guarantee the reachability of the sliding surface. Then, based on the Lyapunov stability and sliding mode control theories, sufficient conditions are derived to ensure that the sliding mode dynamics are globally uniformly asymptotically stable with a prescribed ${\mathcal{H}}_{\infty }$ performance. Finally, the validity of the proposed method is verified by a numerical example.

本文解决了 ${\mathcal{H}}_{\infty }$通过基于观察者的滑模控制方案实现开关神经网络的同步。关于切换神经网络，引入了持续的停留时间切换定律来控制子系统之间的切换，这是一种更为通用的切换定律。此外，鉴于在各种应用中大多数干扰的测量困难，因此构造了干扰观察器。随后，基于扰动观测器状态构造一个完整的滑动表面和一个滑模控制律，以保证滑动表面的可到达性。然后，基于Lyapunov稳定性和滑模控制理论，得出了足够的条件以确保滑模动力学在规定的条件下全局一致渐近稳定${\mathcal{H}}_{\infty }$表现。最后，通过数值算例验证了所提方法的有效性。