DoS attack schedules for remote state estimation under multi-sensor round-robin protocol

In this paper, we investigate the optimal scheduling of denial-of-service (DoS) attacks for remote state estimation in cyber–physical systems with multiple sensors and channels. To avoid data conflicts, only one of the sensors can access the wireless communication channels and transmit its data packets to the remote estimator which may also experience congestion caused by a malicious DoS attacker. Given the energy limitations of the attackers, decisions must be made at each time step about whether to initiate an attack on the communication channels. We obtain an inherent dynamic characteristic of the error covariance at the remote estimator and propose an optimal attack schedule by concentrating on two scenarios: (i) smart sensors that can execute a Kalman filter algorithm in real time, and (ii) non-intelligent sensors that are restricted to collecting and transmitting only raw measurement data. Finally, the theoretical results are illustrated by numerical examples.