Based On The Cycle Stability Of Digital Signal Modulation Identification And Parameter Estimation Research

The modulation recognition is one of the key technologies of the non-cooperativecommunication, such as electronic countermeasures and software defination radio,therefore it is such a significant discipline and orientation in the research ofcommunication field. Obviously, without enough knowledge in transmitter and verifiedparameters about receivers, for example signal power, carrier frequency and phase offsetsand so force, it is hard to get the modulation of blind identification. What is more, for thesake of the multipath fading, frequency-selective and time-varying channels in the realscenarios the task is even harder.One of the statistical characteristics of the digital signal is its certain periodicity,called the cyclostationary feature. For the cyclostationary features of different types ofsignals have significant distinctions, cyclostationary characteristic of the signal applied tothe modulation recognition has its unique advantages: the structural recognition featurewhich is simple and intuitive to distinguish; identifying amounts of signals without prioriinformation; good recognition performance in low SNR conditions; estimating of themodulation parameters of signal precisely.With this in mind, the author provides a comprehensive survey of differentmodulation recognition techniques in a systematic way and comes up with a recognitionmethod based on the theory of cyclostationarity which is used on2ASK, BPSK, QPSK,8PSK,16QAM,2FSK,4FSK, MSK, OQPSK nine common digital signal modulationrecognition and parameter estimation techniques. Firstly, I construct classificationcharacteristics according to the cyclic spectrum signal, the cycle of accumulation, as wellas the signal power spectrum, and then delve on a detail study to extract impact of thechanges based on in-band signal to noise ratio, the amount of data, sampling rate, and thefilter roll-off factor. Secondly, the study related to the cyclic spectrum estimates based onthe signal carrier frequency, character rate, and single-frequency carrier frequency signalestimates based on power spectrum and SNR is done. Finally, the modulation recognitionprogram about nine signal design is designed and simulation analysis work is done. The simulation results of seven common modulation show that, without knowing carrierfrequency, symbol rate, and non-integer sample rate of the signal, when the band SNRgreater than or equal5dB, the inter-class correct recognition rate is more than80%, andwhen the band SNR greater than or equal8dB, the inter-class correct recognition rate ismore than90%.On the other hand, the intra-class correct recognition rate onMPSK/MQAM is more than90%when the band SNR greater than or equal15dB.