Research On Demodulation Techniques Of Paired Carrier Multiple Access Signals

Targeting at the Demodulation of Paired Carrier Multiple Access(PCMA) signals, somesignificant problems are researched in depth in this thesis, which include the modulationrecognition of PCMA signals, the frequency offset estimation and initial phase estimation of8PSK modulated PCMA signals, the blind separation of cooperative and non-cooperativereceived PCMA signals and so on. The main work and innovative points completed in the thesiscan be summarized as follows:1. In non-cooperative satellite communication system, a recognition scheme is given torecognize the specific modulation mode of PCMA signals from the signal set, which containscommon digital modulated signals and PCMA signals. Four characteristics which are notsensitive to the frequency offset are extracted to construct a tree-based classifier to accomplishthe identification. Simulation results show that, the classifier is not sensitive to the timing errorneither, and high correct recognition rate can be achieved in common condition ofcarrier-to-noise ratio(C/N) for the PCMA signals, whose amplitude mixing ration is high than0.7.2. Aiming at the frequency offset estimation and initial phase estimation of8PSK modulatedPCMA signals, when two source signals are coded asynchronously(the positions of sync codes intwo signals are not overlapped) and the frame length is known, a data-aided frequency offsetestimation algorithm is derived from the idea of maximizing the likelihood probability functionfirstly, and the effective estimation range is given. After realizing the frequency offset estimation,a data-aided initial phase estimation algorithm is given, and the feasibility of proposed algorithmis verified by the simulations.3. To separate the cooperative received PCMA signals, a joint linear MinimumMean-Square Error(MMSE) equalization and soft decoding iteration demodulation/decodingalgorithm is proposed, which utilizes the error correcting ability of channel coding afteraccomplishing the channel parameters estimation and the interference cancellation. In the courseof equalization, the algorithm exploits the feedback prior statistic after decoding to improve theperformance of demodulation, which special emphasis on the information interaction betweenequalization and decoding, and the impact of the parameters estimation error on the performanceof demodulation. Simulation results show that, with Turbo Product Codes(TPC) whose subcodeis (64,57,4) expand BCH codes resorted, at a bit error rate of10-4, a gain of about2dB inperformance can be achieved for QPSK modulated signals after1iteration, when compared with Per-Survivor Processing Maximum Likelihood Sequences Estimation(PSP-MLSE) algorithmwhose channel is initialized accurately.4. A PSP iterative separation algorithm based on the idea of BCJR is proposed for thePCMA signals in non-cooperation communication. To start with, in order to improve theseparation performance of PSP algorithms which have been existed so far, the BCJR algorithm isexpanded from the view of non-binary based on MAP criterion, by which BCJR-PSP iterativeseparation algorithm that can provide soft output is given, then the separation performance isimproved further by the way of oversampling. And furthermore, to diminish the computationcomplexity of existed PSP algorithms, when any one of the two source signals can beapproximately optimally sampled, the number of non-zero taps of the corresponding equivalentchannel filter can be seen as one, then the trellis can be constructed only for the other sourcesignal, thus the BCJR-PSP soft output iterative separation algorithm based on the synchronouslysampled is proposed in this thesis.5. Targeting at the separation of non-cooperative received PCMA signals, the constellationof PCMA signals can be seen as the higher-order modulated signal’s constellation when thedifference between two source signals’ timing error is small enough. Then the iterative scheme ofTurbo equalization is constructed, which includes four modules: initial blind equalization, SISOdemapping, Decision Feedback Equalization(DFE) and MMSE iterative equalization. Simulationresults show that, when the difference between two source signals’ timing error is small enough,the separation performance can be the best if the amplitude mixing ratio ranges from0.4to0.7for QPSK modulated PCMA signals; and the separation performance will be the best when theamplitude mixing ratio is0.6for8PSK modulated PCMA signals.