Particle filtering for a class of cyber-physical systems under Round-Robin protocol subject to randomly occurring deception attacks

In this paper, the particle filtering problem is studied for a class of general nonlinear cyber-physical systems with non-Gaussian noises under Round-Robin protocol (RRP) subject to the randomly occurring deception attacks. In order to prevent the data from collisions and alleviate the communication overhead for the shared network with limited resources, the RRP is introduced in the sensor-to-filter channel to schedule the multiple sensors with a predefined transmission order. Under the RRP, only one sensor can be granted the access to the shared channel for measurement transmission at each time instant. A Bernoulli-distributed stochastic variable is utilized to describe the characteristic of random occurrence of deception attacks initiated by the adversaries. A RRP-based particle filtering algorithm is developed by establishing a modified likelihood function, where the statistical property of the randomly occurring deception attacks is exploited and the RRP-induced effect on the filter design is reflected. Finally, an illustrative example regarding the target tracking problem is provided to verify the feasibility and effectiveness of the developed particle filtering scheme.