Research On Detection & Recognition Of Digital Modulation Signals In Short-wave Channel

In recent years, digital modulated signal technology has been a new power in the field of communication signals. It is not only widely used in the military communication realm, but also in the field of civilian and commerce. Therefore, the technology of getting and detecting the information in the modulated signals is the focus on the contest of information. Up to now, at least most of the domestic work about the detection and recognition is still finished by manual work, which does not fulfill the application requirement and badly hurts health. This paper stands on the requirement of practical application, focusing on the research of solving the difficulty and key problems in the detection and recognition of digital modulated signals.Firstly, the paper introduces and analyzes the characters of modulated signals, noise and speech signals. Especially, we introduce the modulated signals needed to be detected, including 39-Lines signals, QYT6 signals and 8-frequency quantified signals and so on; in the course of detection, we use the basic structure in pattern recognition—tree model. We take two characters: one is short-time standard deviation in time field; the other is frequency spectrum evenness in frequency field. After many imitated experiments and practical running, the algorithm has the strongpoint of simple computing, little running time, high detection rate(≥90%) and low missing rate(≤2%), which basically satisfies the requirement of real time detection.Secondly, in the research of digital modulated signals recognitions, which mainly focuses on the recognition of ASK, FSK and PSK signal. The algorithm just uses two characters: one is the number of statistical frequency peak value in histogram of the power spectrum; the other is signal phase parameter R. Using the two characters, we first recognize FSK signal and then recognize ASK signal. Compared with the former work, the algorithm is simpler and has higher recognition rate, which comes to 95 percent when SNR≥15dB. The experiments prove that the algorithm is effective and utilitarian. At last, the paper introduces the function and the design of detection system software. The system can save the modulated signal and noise signals separately for the further process. Meantime, it also saves the original files as copy files.