The Research On Automatic Recognition Of Communication Signal Modulation Based On Cognitive Radio

Modulation recognition of communication signals is one of the key technologies in cognitive radio system. Any communication signals must know the modulation and parameters of the signal to receive and demodulate. The traditional communication radio or system is designed for a single system with specific modulation and bandwidth. Its application scope is very limited and not suitable for the multi-modulation, multi-service system at present. Therefore, the recognition technology of modulation signals will play a very important role in the continuous development of cognitive radio, associated with diversification of communication modulation signals.The purpose of this research is to analyze the modulation from received signals in condition of noise interference in the premise of unknown modulation information. The significance of this study is that the current recognition work of modulation signals focuses on the classification recognition between digital and analog signals, in digital or analog signals, as well as single-carrier and OFDM signal. However, more study about the identification of mixed-signal set has not so much. Therefore, for the above situation, the paper implements a new multi-feature binary-decision-tree recognition algorithm and new feature is proposed.The main work in this paper can be summarized as follows:(1) First, the structure of software radio and cognitive radio as well as the key technology at present are discussed. The existing modulation signal recognition algorithm is reviewed and the characteristics of existing algorithms are analyzed. (2) Using cyclostationary algorithm to identify digital and analog signals, extract three characteristics to complete the identification process.(3) Because of the high complexity, computational load and poor real-time in cyclostationary algorithm, a new characteristic parameter and new multi-feature binary-decision-tree recognition algorithm are proposed. The average probability of correct classification is better through Matlab simulation in low SNR conditions. The computation load of the method is low and suitable for military CR.