Research Into Non-data Auxiliary SNR Evaluation Methods For Communication Signals

The SNR in communication signals is an important index reflecting the quality of the communication signals and plays an indispensible role in many research areas of the communication signal processing. Therefore, the research into SNR evaluation methods is of great significance. Based on the study of the current SNR evaluation methods in communication signals, this thesis has come up with new SNR evaluation methods applicable to various occasions. The detailed work includes the following four areas.1. It presents a detailed analysis of the backgrounds and application occasions of SNR evaluation, the classification of SNR evaluation methods, relationships and application effects among different algorithms as well as existing problems in communication signals.2. Oriented toward the QPSK signals, it has come up with a SNR joint evaluation method based on the multiple-statistical-volume approach. The evaluation accuracy of the DF algorithm is high in low SNR, but its evaluation performance deteriorates in high SNR. However, the evaluation accuracy of M2M4is very high even in high SNR. This thesis uses the ML iterative SNR evaluation algorithm as the startup algorithm, which is selected the basis for further algorithm operation through the M2M4and DF to conduct accurate SNR evaluation section by section, thus expanding the SNR evaluation scope of the original algorithm.3. As for MQAM, this thesis uses the high-order statistical volume M2and M4to set up a related format for SNR. The least square is used to get a new SNR evaluation method. The experimental simulation has shown that this method has a higher accuracy than theM2M4M6in the entire SNR evaluation scope.4. As the method to use the statistical volume for the evaluation of the MQAM SNR is not ideal in low SNR, this thesis has proposed that the eignfunction derivative should be used to evaluate the signal power, and that the statistical volume should be used to evaluate the noise power, thus forming a new SNR evaluation method. This method has made use of the eignfunction which is featured by good SNR evaluation in low SNR, making up the shortcoming of high-order statistical volume-based SNR evaluation in low SNR.5. This thesis has analyzed the code-aided SNR EM iterative evaluation algorithm. After the analysis of the algorithm model, this thesis has proposed that the non-code-aided SNR EM iterative algorithm should be taken as an exception to combine the two algorithms for a unified representation. As this algorithm has a complex operation, this thesis uses the mutual complementary work process between SNR and LDPC to simplify the operation by introducing the embedded evaluation algorithm. Simulation results show that the enhanced algorithm has a low operation complexity while maintaining the evaluation performance of the original algorithm.