低信噪比条件下信号调制体制识别算法研究通信与信息系统专业论文.docx

哈尔滨工业大学工学硕士学位论文 哈尔滨工业大学工学硕士学位论文 II II Abstract Modulation recognition is a technique to automatically classify the modulation typeand parameter of a received communication signal corrupted by noises. It is a rapidly evolving signal analysis area and has found a variety of applications in modern communication systems.With rapidly developing of communication technology, the environment of communication becomes more and more complex and the SNR (Signal to Noise Ratio) of modulated signals becomes lower and lower. It results in that the electronic surveillance, electronic warfare and spectrum management become more and more difficulty. How to improve the performance of modulation recognition algorithms in low SNR becomes an important theme for their applications. So the main purpose of this dissertation is to seek the perfectly performance algorithms especially atlow SNR. Recently, several modulation identification techniques have been developed. In general, they can be categorized as a decision-theoretic approach and a statistical pattern recognition approach. The decision-theoretic approach is optimal in the sense that it maximizes the averaged probability of successful classification. But the difficulty lies within the implementation of such a decision rule. The pattern recognition approach is generally a feature-based recognition algorithm that extracts the features in the frequency domain (e.g., spectrum estimation) or in the time domain (e.g., instantaneous envelope, frequency, phase) from the raw signals. This approach is more realistic in the sense that it requires less a priori knowledge of received signals and can be applied under various radio channel environments. Based on the work of others，by the analysis of spectrum of signals, the modulation recognition and parameters estimation of many common used modulated communication signals are investigated this dissertation. The main works can be summarized as follows: 1．A novel recognition scheme based on the sp