This article investigates the asynchronous distributed finite-time $H_{\infty }$ filtering problem for nonlinear Markov jump systems over sensor networks under stochastic attacks. The stochastic attacks, called two-channel deception attacks, exist not only between the Markov jump plant and the sensors but also among the sensors. It is assumed that the mode of the filter relies on, but is asynchronous with, that of the Markov jump plant. First, we establish a filtering error system that combines the Markov jump plant with the asynchronous filtering system. Then, we present an asynchronous distributed filter, which ensures the filtering error system mean-square finite-time bounded and satisfies a prescribed $H_{\infty }$ performance level under the two-channel attacks. Finally, an example is given to illustrate the effectiveness of the presented filter.

中文翻译：

---

具有隐马尔可夫切换和两通道随机攻击的传感器网络中的异步分布式有限时间 H∞ 滤波

本文研究了异步分布式有限时间 $H_{\infty }$随机攻击下传感器网络上非线性马尔可夫跳跃系统的滤波问题。随机攻击，称为双通道欺骗攻击，不仅存在于马尔可夫跳跃工厂和传感器之间，而且存在于传感器之间。假设滤波器的模式依赖于马尔可夫跳跃工厂的模式，但与之异步。首先，我们建立了一个将马尔可夫跳跃工厂与异步过滤系统相结合的过滤误差系统。然后，我们提出了一种异步分布式滤波器，它保证了滤波误差系统均方有限时间有界，并满足规定的 $H_{\infty }$双通道攻击下的性能水平。最后，给出一个例子来说明所提出的过滤器的有效性。