| Faced with the growing scarcity of spectrum resources,cognitive radio technology has been proposed.Wireless awareness is especially important in cognitive radio.In addition,the powerful tool of deep learning has achieved excellent results in various fields,and this paper introduces the deep learning method into the research of cognitive radio,focusing on solving the wireless environment perception problem based on deep learning.This thesis first introduces the research background,communication model and deep neural network,deduces the expression of the received signal,and analyzes the common problems in the wireless environment perception.Secondly,this thesis discusses the signal-to-noise ratio estimation problem of received signals,and proposes a signal-to-noise ratio estimation algorithm based on deep learning and eye diagram.The algorithm estimates the signal-to-noise ratio based on the clarity of the eye diagram under different signal-to-noise ratios.It still has higher accuracy at low signal to noise ratios.Then,this thesis uses depth learning to solve the channel coding recognition problem of received signals,and proposes a convolutional code recognition algorithm based on TextCNN.The algorithm can be regarded as a natural language processing problem by word vector representation,which greatly enhances channel coding recognition.performance.Finally,based on the convolutional code recognition research,this thesis also proposes a deep learning-based Turbo code recognition algorithm,which distinguishes the Turbo code according to the difference of the component encoder.The experimental results show that the proposed method is based on the high bit error rate.The algorithm has higher recognition accuracy than traditional algorithms. |
