Sliding-Window-Based Detection for Spectrum Sensing in Radar Bands

Recently, radar frequency bands have been drawing much attention as possible candidates for cognitive radio, since their utilization was revealed to be relatively low. One important functionality of a cognitive radio system is the spectrum sensing, which detects the presence of a primary user (PU) signal. The pulse radar signal, which is the PU signal in radar bands, is characterized by its sparsity in the time domain. Exploiting this sparsity, we present a spectrum sensing technique that combines the intermediate detections from multiple small sliding sensing windows of the received radar signal into a final decision. We also show by computer simulation that it can achieve improvement of the sensing performance in comparison with the conventional single-window-based spectrum sensing approach.