Mobile-based Collaborative Compressive Spectrum Sensing for Cognitive Radio Networks

Spectrum sensing task is an essential operation in cognitive radio networks. As such, this paper considers the impact of primary mobile users whose location and dynamic spectrum use can be tracked and sensed by collaborative secondary users. We evaluate the probability of detection for two simplistic random mobility models by acquiring compressed signal measurements at the fusion center. The collaborative signal acquisition is to reduce the computation at the secondary user side. Combined with the recovered signal via compressed sensing, a localization technique by Kalman filtering is used to track a primary mobile user in the network region. For a random mobility model, it is shown that better spectrum sensing performance can be achieved with a high probability of detection. Further, for pseudo-static movements having pause time factor, the detection performance increases. Simulation results are given to corroborate the approach used in the evaluation.