Robust Spectrum Monitoring in Cognitive Radio Networks With Uncertain Traffic Information

Without interfering wireless networks, passive spectrum monitoring is important for network diagnosis and radio frequency management in spectrum-sharing wireless networks. Most of the related works focused on the sniffer-channel assignment problem, i.e., assigning each wireless sniffer a proper operating channel, with the aim of tracking the target signals or data packets. These approaches were usually designed for the scenarios, where the behaviors of malicious or suspect wireless users are known. In this paper, we focus on the problem of spectrum patrolling in a cognitive radio network, in which the sniffers have no specific targets, but try to patrol the spectrum of interest over a temporal-spatial region. Once the periodicity or regularity of the wireless traffics is identified, a patrol path will be developed for routine patrolling. The path planning problem is formulated as a robust reward maximization problem with uncertain channel information. We propose both optimal and sub-optimal algorithms to determine the route of spectrum patrolling and validate it through numerical simulations. Simulation results show that our proposed algorithm can achieve the maximal reward even with unknown information of the users' activities.