2013/06/10 by Erhan Bayraktar, Bayraktar, Erhan, Lifeng Lai +1 · 1 citation
Computer Science · Decision Sciences · Medicine · #Advanced Statistical Process Monitoring #Data-Driven Disease Surveillance #Distributed Sensor Networks and Detection Algorithms #FOS: Computer and information sciences #FOS: Electrical engineering #FOS: Mathematics #Information Theory (cs.IT) #Optimization and Control (math.OC) #Probability (math.PR) #Systems and Control (eess.SY) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.1306.2086
openalex publication_date 2013/06/10 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We introduce and solve the problem of Byzantine fault tolerant distributed quickest change detection in both continuous and discrete time setups. In this problem, multiple sensors sequentially observe random signals from the environment and send their observations to a control center that will determine whether there is a change in the statistical behavior of the observations. We assume that the signals are independent and identically distributed across sensors. An unknown subset of sensors are compromised and will send arbitrarily modified and even artificially generated signals to the control center. It is shown that the performance of the the so-called CUSUM statistic, which is optimal when all sensors are honest, will be significantly degraded in the presence of even a single dishonest sensor. In particular, instead of in a logarithmically the detection delay grows linearly with the average run length (ARL) to false alarm. To mitigate such a performance degradation, we propose a fully distributed low complexity detection scheme. We show that the proposed scheme can recover the log scaling. We also propose a centralized group-wise scheme that can further reduce the detection delay.