SDR Implementation of Spectrum Sensing for Wideband Cognitive Radio

This paper provides experimental results of the edge detection and spectrum sensing algorithms for wideband cognitive radio networks which are recently proposed in [1] using software defined radio (SDR) platform. The considered algorithms employ ratio based test statistics for detecting the edges of all sub-bands and generalized energy detection (GED) for examining the status of each sub-band. In particular, we validate the theoretical detection and false alarm probabilities of the edge detection and GED algorithms of [1] experimentally for a number of practically relevant parameters such as sensing time and bandwidth. We also compare the performances of these algorithms with and without calibrating the Cognitive Radio Device (CRD). Through extensive experiments, we have found that the theoretical performances claimed in [1] can be achieved reliably just by performing appropriate calibration at the CRD. Moreover, we also verify that the considered detection algorithms are robust against noise variance uncertainty, carrier frequency and timing offsets.