An advanced blind cyclostationary spectrum sensing algorithm for cognitive radio networks

A multitude of methodologies, based on the detection of cyclostationary features (CFD), are available to researchers seeking to identify unoccupied spectrum channels within cognitive radio (CR) networks. Notwithstanding the inherent difficulties of wireless environments, such as those involving a low signal-to-noise ratio, CFD has demonstrated considerable potential. However, addressing signal variation and system complexity remains a primary area of research. This paper introduces a novel CFD algorithm that employs the autocorrelation function as a preprocessing step to enhance the received signal characteristics and distinguish noisy signals from noise. Subsequently, a straightforward cyclostationary detection approach is applied. The objective of this blind cyclostationary spectrum detection technique was to reduce the algorithmic complexity and enhance the detection efficiency. This paper presents optimized parameters for a blind cyclostationary detector and offers an evaluation of its performance in simulation environments. The results demonstrate a minimum 3.5 dB enhancement in detection performance relative to the benchmarking techniques. Furthermore, the SDR implementation of the proposed method in the receiving part of a transmitter/receiver FM broadcasting system, using two USRPs cards connected to two laptops running the GNU Radio platform, serves to validate its effectiveness in real-time scenarios.