A real time spectrum scanning technique based on compressive sensing for cognitive radio networks

This paper describes a real time spectrum scanning method using software defined radio (SDR) units to perform and evaluate the wideband spectrum occupancy survey. The proposed method is based on compressive sensing in order to reduce the scanning time as the conventional spectrum scanning requires a great deal of processing time. It consists of performing spectrum scanning on the compressed measurements through the compressive sensing framework. Energy, autocorrelation, and Euclidean distance based sensing techniques were used to perform the spectrum sensing on the compressed signals instead of the entire signals. Also, an SNR estimation technique is used to enhance the scanning performance. The results of the experiments were analyzed and compared in terms of processing time, number of sensed channels, SNR, occupancy, detection rate, and false detection rate. Through analyzing the results, the proposed method is faster than the conventional spectrum scanning.