Energy Detector for Spectrum Sensing using Robust Statistics in non-Gaussian Noise Environment

In Cognitive Radio, spectrum sensing enables dynamic spectrum access. Conventional energy detector is a blind algorithm, where the detection performance depends on the estimation accuracy of noise variance that deteriorates in a non-Gaussian noise environment. Hence, in this work, we propose median absolute deviation statistics to estimate the noise variance for threshold evaluation. Further, we estimate the energy of the received signal using the Huber cost function to mitigate the effect of non-Gaussian channel noise. The detection performance of the proposed algorithm is compared with conventional energy detector and deep learning-based detectors on the RadioML2016.10A dataset, signal-to-noise ratio (SNR) ranging from -20 dB to +18 dB. The simulation results demonstrate that the proposed algorithm outperforms conventional energy detection and is at par with deep learning-based detection.