Weak Signal Detection Method Based on Bayesian Neural Network

With the development of The Times, the high efficiency and privacy of modern communication become increasingly important. How to distinguish weak signal from noise in the case of insufficient prior information becomes particularly important. The traditional modulation recognition process requires manual extraction of signal features, and the process is complicated and the unstable factors are too large, which cannot meet the requirements of high precision and high time of modern communication. With the development of deep learning, machine learning has made great breakthroughs in the direction of radio signal processing due to its obvious classification and decision advantages, and signal recognition has a higher accuracy than before. More and more researches begin to use machine learning method to extract features directly from the original data of communication signals, which ensures the real-time and robustness of automatic modulation recognition of communication signals. Naive Bayes is one of the classical algorithms of machine learning and one of the few classification algorithms based on probability theory. This method is still effective in the case of less data and can deal with multiple kinds of problems. Based on Bayesian network algorithm, this paper studies the weak signal identification method of wireless telecommunication below 0db. Aiming at the problem that the generalization ability of traditional model is insufficient, this paper proposes a weak signal recognition method based on Bayesian network. The experimental results show that the recognition rate of this method is 40 percent higher than that of different Bayesian neural networks in the recognition of radio weak signals, which improves the recognition ability of the whole recognition model for weak signals not in training concentration.