Collaborative Spectrum Sensing in Tactical Wireless Networks

In this paper, we propose an algorithm for channel sensing, collaboration, and transmission for networks of Tactical Communication Systems that are facing intrusions from hostile Jammers. Members of the network begin by scanning the spectrum for Jammers, sharing this information with neighboring nodes, and merging their respective sets of observation data into a decision vector containing the believed occupancy of each channel. A user can then use this vector to find a vacant channel for its transmission. We introduce the concept of nodes sharing these vectors with their peers, who then merge them into super-decision vectors, allowing each node to better identify and select transmission channels. We consider fading scenarios that substantially limit the reliability of the users' observations, unless they cooperate to increase the trustworthiness of their sensing data. We propose a pseudo-random channel selection algorithm that strikes a balance between sensing reliability with the number of channels being sensed. Simulation results show that the proposed system improves the network's overall knowledge of the spectrum and the rate of Jammer-free transmissions while limiting added computational complexity to the nodes of the network, despite the Jammers' unpredictable nature.