Spectrum sensing based on entropy estimation using cyclostationary features for cognitive radio

Abstract

This work presents a new spectrum sensing technique based on entropy estimation of autocorrelation estimates of received signal at different cyclic frequencies. The performance of the proposed entropy detection is compared with cyclostationary detection based on spectral coherence function (SCF) and energy detection methods. Both the algorithms are verified under single node and multinode/cooperative environment. Sensing performance of both the algorithms are analyzed using Monte-Carlo methods. Simulation result discloses that, entropy detection algorithm detect signals of signal-to-noise ratio (SNR) upto 26dB whereas SCF method detects signal upto 23dB using seven nodes in cooperation for detection probability P d≥0.9 and false alarm probability P fa≤0.1. The proposed sensing algorithm is also implemented in Virtex-4 (XC4VSX35-10FF668) Field Programmable Gate Arrays. © 2012 IEEE.