Research On The Modulation Identification Of Radio Signal

The automatic classification and identification of modulation scheme of communication signals is an essential function of communication signal processing, and an important part in electronic countermeasures also. It is a rapidly evolving area of signal analysis. The automatic identification of signal modulations has been applied to many fields such as signal identification, interference identification, radio interception and monitoring, etc. The objective of automatic modulation identification is to decide the modulation type and estimate the modulation parameters without any priori knowledge about signal contents. In recent years, interest from the academic and military research institutes has focused around the research and development of modulation recognition algorithms. This thesis is concered with presenting new algorithm for automatic modulation recognition of communication signals.The theory of signal modulation and recognition is introduced first. And then the parameter estimation of communication signals is discussed. Finally, three kinds of modulations recognizers and a spectrum correlation recognition algorithm, based on the policy-making theory, are proposed.Aimed at the basic characteristics of signals both in time domain and in frequency domain, the analogous recognizer is used to distinguish the analog modulations, the digital recognizer to the digital modulation, and the composite recognizer both to analog and to digital modulations, without knowing the type of signals in advance. The decision flows of these algorithms are simulated in software environment of a computer, and revisions are made to these algorithms according to empirical results.In view of signal characteristic about spectrum correlation and based on the spectrum correlation theory, this article has studied a modulation recognition algorithm which may distinguish AM, CW, FM, DSB, VSB, LSB, USB, 2PSK, 4PSK, 2ASK, 4ASK, 2FSK and 4FSK modulated signals. Under a lOdB of signal to noise ratio, experimental results indicat an average recognition rate of 95.7%, and the lowest recognition rate of 86.3%.