Detection of data tampering attack in FIR system identification with binary-valued observations

This paper addresses the security detection of data tampering attack in identification of finite impulse response systems with quantized observations. First, we designed a detection algorithm for data tampering attack based on the prior information of the system and gave its online detection form. Second, the introduction of the concepts of the false judgment rate and the missed judgment rate gave a specific evaluation index for the performance of the detection algorithm. Considering the high nonlinearity of binary quantized and the difficulty of obtaining distribution function, we provided calculation methods for two indexes and gave expressions under large samples. Then, the effects of the data length, the prior information of the system parameter, and the attack strategy on the detection algorithm are discussed in this paper. Finally, the theoretical results are verified by numerical simulation.