In this paper, we consider the \(H_{\infty }\) filtering problem for two-dimensional (2D) discrete-time Markov jump systems (MJSs) described by the Roesser model in a networked environment. Due to the network communication between the plant and the filter, the measurements may be subject to cyber-attacks. To this end, a Bernoulli distributed stochastic variable is employed to model the phenomenon of deception attacks. For the reason that the probability of occurrence of this random variable is not exactly known in practice, this value is supposed uncertain while the majority of the existing results consider it to be completely known. By constructing a new structure of the Lyapunov functional and using some zero equalities, a new \(H_{\infty }\) performance analysis condition for the filtering error system is established. Moreover, a novel condition is developed for the filter design, which can unify the design of both parameter-dependent and parameter-independent filters in a single framework to cover both cases of absence and presence of mode transition information. All results are presented in terms of linear matrix inequalities (LMIs). In the end, the usefulness of the aforementioned filtering method is verified by a numerical example with simulation.
