Research On Detection And Modulation Analysis Techniques For Co-channel Multi-signal

With the rapid development of wireless communication technologies, radio spectrumresources are in short supply, the modulated signals in the same frequency band are becomingmore and more crowded. It is possible that there are multiple independent time-frequencyoverlapped signals being received at the same time, which is called Co-channel Multi-Signal.Because of time-frequency overlapping, those traditional methods which are used to processsingle signal are no longer able to be used in general. This brings huge challenges to signalanalysis and demodulation without prior knowledge in communication investigation. On thebasis of analyzing the characteristics of co-channel multi-signal, this thesis realized blinddetection and modulation analysis of those types overlapped by BPSK, QPSK,8PSK,8QAMand16QAM. The main content includes signal detection, source number estimation, modulationparameters estimation and modulation recognition.The main work and innovative algorithms proposed are summarized as follows:1. Combining with the concept and signal model of co-channel multi-signal, the timedomain instantaneous characteristics, frequency domain characteristics, higher order statisticscharacteristics and cyclostationary characteristics are analyzed. The processing flow ofco-channel multi-signal blind detection and analysis are given briefly.2. For the existence detection of co-channel multi-signal, an algorithm based on fractal boxdimension is proposed. The detection threshold is set by theoretical deduction. The influence oftime-frequency overlapping is discussed; theoretical analysis and simulation results proved thatthe algorithm can solve the problem perfectly. Compared with energy detection, thecomputational complexity is lower.3. For the number estimation of co-channel multi-signal received by single channel, analgorithm based on cyclic cumulants is presented. For the single carrier signals, the carrierfrequency number equals to the source number when the carrier frequencies of all independentsignals are different. According to the relationship between carrier frequency and cyclicfrequency, the algorithm is realized by searching the number of cyclic frequency. The problemthat it is complex to set threshold to estimate cyclic frequency is solved emphatically, thealgorithm doesn’t need to set threshold anymore, and can be used to existence detection as well,so it is more flexible in implementation.4. An algorithm for modulation parameters estimation of co-channel multi-signal based oncyclic cumulants and cyclic spectrum is proposed. The carrier frequency of each independentsignal is estimated by the cyclic frequency of cyclic cumulants firstly, and then the symbol rate isestimated by the cyclic frequency on carrier frequency profile of cyclic spectrum. It not only canestimate the modulation parameters of each independent signal in low SNR accurately, but alsocan solve the matching problem between parameters effectively.5. An algorithm for modulation recognition of co-channel multi-signal based on high-ordercumulants is presented. It is proved that independent signals can be separated in cumulants domain if their carrier frequencies are all different. Then the symbol synchronization ofco-channel multi-signal is studied, it can be concluded that the maximum mean power symbolsynchronization method can be used for symbol synchronization of each signal on condition thattheir over-sampling ratios are all different and irreducible. Subsequently, cumulants of eachsignal are calculated after being synchronized. According which, the feature parameters areextracted to recognize the modulation type one by one. This method can classify moremodulation types, so the application range is wider.