Spectrum Sensing Based On Compressed Sensing

Spectrum sensing has become a key technology of cognitive radio. Wireless sensor network based is a good method. For large wireless sensor networks, the primary users are relatively fewer than the number of channels. Because of deployment cost, the number of sensors is limited, and due to energy constraint, not all the sensors are turned on all the time. This paper introduces a system model of spectrum sensing based on compressive sensing (CS). The primary users are considered to be sparse in the network. The number of sensors in work can be greatly reduced, which is much smaller than the total number of channels. Restricted Isometry Property (RIP) will ensure that the original spectrum signal can be reconstructed from less measurement data with high probability. Fuse center can recover the signal with Orthogonal Matching Pursuit (OMP). Differing from the previously works, spectrum signals are considered as continuous value. In the way, users can choose the desired channels by adjusting the threshold. Simulation results show that the scheme performs effectively. If the signal is operated an orthogonal transformation before measuring, the reconstructed results are better.