A Software Radio Implementation for Spectrum Hole Sensing in Cognitive Mobile Networks

This paper presents a software radio implementation for spectrum hole sensing in cognitive mobile networks. We propose a multi-channel spectrum sensing method for spectrum hole detection, exploiting Fast Fourier Transform (FFT) filter bank (FFTFB). Our method tests multiple frequency bands at the same time, detecting multiple spectrum holes at once. We modify a recently introduced two-dwell sensing method, implementing the second stage under a software radio approach (i.e. using a digital reprogrammable hardware). In particular, a filter bank implementation varying the length of the FFT is deeply investigated. FFTs with a small number of coefficients, that can be computed with fast and low complexity algorithms (multiplication free), are exploited and their performance discussed in terms of required complexity. The obtained results have evidenced the efficiency of such an approach for cognitive mobile networks