Effective autocorrelation-based spectrum sensing technique for cognitive radio network applications

Spectrum sensing based on detection techniques enables cognitive radio networks to detect vacant frequency bands. The spectrum sensing gives the opportunity to increase the radio spectrum channels re-utilization. However, the main challenge in spectrum sensing is the simplicity of the considered detection approach and the amount of prior information needed to make an accurate decision. This paper proposes a novel sensing technique based on the autocorrelation function. This novel approach is based on the speed of convergence to zero of all autocorrelation coefficients. This technique shows the highest probability of detection for the same probability of false alarm target at low signal-to-noise ratio (SNR) compared with many standard detection techniques. The proposed method has been implemented using GNU Radio software and SDR (software-defined radio) platforms. The experimental results show the effectiveness of the proposed method under real scenarios.