Distributed fusion estimation for non-linear networked systems with random access protocol and cyber attacks

This study investigates the distributed fusion estimation problem for multi-sensor non-linear networked systems with random access protocol (RAP) and cyber attacks. For each sensor, the RAP is applied to determine which channel can enter the communication networks at each transmission instant. The process of transmitting measurements to a corresponding local estimator (LE) via the wireless network may be attacked by a vicious attacker. Hence, a detector based on cross-correlation of the causal sequence is proposed in the LE. Moreover, a distributed fusion estimation algorithm based on a sequential covariance intersection fusion technique is designed utilising local estimations calculated by modified unscented Kalman filter. A sufficient condition is obtained to guarantee the boundedness of fusion estimation error. Finally, a numerical simulation is given to demonstrate the effectiveness of the theoretical results.