A centralized cooperative study of spectrum sensing based on energy detection in cognitive radio networks
Abdullah Hussein Talib, Aseel Hameed Al-Nakkash, Ahmed Ghanim Wadday · AIP conference proceedings · 2024
As next-generation wireless communication technologies advance and spectrum resources become scarcer, Cognitive Radio Networks (CRN) must incorporate learning and reasoning capabilities.Spectrum Sensing (SS) in particular has emerged as the primary target for investigation to solve the problems of CRN.Deep learning techniques have been used to add new aspects to SS.This paper is based on a statistical analysis of the classic energy detection scheme, which is solely dependent on the number of samples and secondary users' signal-to-noise ratio.The detection of channel occupancy is based on well-established analytical techniques such as Maximum Ratio Combining (MRC) and AND/OR rules.Furthermore, the deep learning Long Short Term Memory (LSTM) algorithm is used for SS and compared to analytic techniques.the results demonstrate that the LSTM technique presents a better trade-off between the probability of false alarms and the probability of missed especially for fading channels with accurate detection of 87.4%.