Suggesting a New Approach on Identifying Degree of Separability in Signal Detection, for Using in Channel Estimation

Abstract

Signal Detection Noise removal, is a very important issue in channel estimation, and increasing performance of signal transformation in cognitive networks. Therefore it is necessary to have a criterion for evaluating the degree of correctness and reliability of the signals. Nowadays neural networks has very important role in calculations and if it combined with statistical methods they will produce perfect results in separability detection.   In this paper, we used the separability degree as a criterion for separating and identifying noise from the main signal. We use statistical Hypotheses and declare some statistical  thresholds  for signal validity to get the signal more suitable by increasing noise detection quality. This method supposes two states for our signal that are false detection of weak signal, and correct detection of the main signal. All these will be done by statistical_neural methods.