Optimal Sensor Scheduling under Denial-of-Service Attack for Remote State Estimation with Energy Constraint

This paper investigates the sensor scheduling problem under denial-of-service (DoS) attack. Specifically, a system is in consideration where a battery-powered sensor sends the data packet to the remote estimator over a wireless channel under the successive DoS attack. Due to the limited power budget, the sensor can transmit M times over the finite time horizon T. The issue of the optimal sensor scheduling for remote estimation to minimize the average expected estimation error has been studied. Most of the existing works considered the channel without attack such that it is assumed that no packet dropouts will happen, which is, however, no longer rational in the presence of DoS attack. And to the best of our knowledge, some sensor scheduling problem under DoS attack for remote estimation over the finite time horizon has not been addressed. To fill this gap, a novel analysis approach is proposed, based on which, we derive the conditions under which the optimal sensor scheduling is found. Numerical example is presented to substantiate the theoretical results.