Event-based distributed filtering against deception attacks for sensor networks with quantization effect

This paper considers a distributed secure filtering problem for a category of time-varying system subject to uncertainty and model-reality mismatch, two-stage deception attacks and bandwidth limitation. Both deception attacks between sensor and corresponding estimator and among estimators appear randomly. To alleviate communication burden, a quantization strategy is introduced before transmitting measurement and estimation signals. An event-triggered mechanism is employed for each estimator node thus only necessary data are transmitted to its neighbour sensors when a setting event occurs. The desired target of the problem to be handled is to devise a series of time-varying filters such that the H_∞ secure performance is guaranteed against random deception attacks over a finite time horizon. Sufficient conditions ensuring the existence of time-varying filters under effect of complex factors are derived, where filter gains are obtained by finding the solution of a sequence of recursive matrix inequalities online. Simulation results in both numerical example and industrial continuous-stirred tank reactor system are given to show the validity of the presented methodology.