Deep Learning applied for Spectrum Sensing using multiple and concurrent stages

Abstract Spectrum sensing techniques have many challenges and as one as the most challenges are the noise and interference rejection. Since the occurrence of noise power uncertainty cause the degradation of the performance of the spectrum detector. One of the most techniques of spectrum sensing is the Energy Level detection it could be used with deep learning network to distinguish between presence of signal and noise to this end, we will introduce a comparison between AlexNet, SqueezeNet ResNet101 and LSTM neural networks. To test them in different situations. . ...