A Detection-Estimation Strategy for Delayed Systems under Spoofing Attack

This paper investigates the linear state estimation problem for networked system with random delay and the packet sequence is unavailable to the estimator, which is due to spoofing attack destroying the time stamp of the packet. We consider a transmission mechanism that the packet is sent only once, as such the packet arrives at the remote estimator may be the current packet or the previous one or even empty. In this paper, we propose a detection-estimation strategy, which consists of the following two parts: Firstly, we detect the unavailable time stamp under the maximum a posteriori probability (MAP) estimation criterion; Secondly, using the estimated sequence of the time stamp, we propose a filtering algorithm which consists of smooth and estimate. Numerical examples are provided to illustrate the proposed algorithm.