The Study Of OFDM System Modulation Recognition Methods Based On Wavelet Neural Networks

The modulation recognition is very important in communications; it is used in many fields such as signal affirmance, interference identifying, frequency surveillance, electron counterwork and military threat analyses.OFDM is one of the multi-carrier modulations, it divides the channel into some quadrature sub-channels, and then the high speed data stream is transformed to parallel low speed sub-data streams, these sub-data streams are modulated to the sub-channels to transmit. OFDM has high utilization ratio of spectrum, low cost and can resist the multi-path interference and frequency-selective fading, so it has become the hotspot of study after CDMA. Thus it is very necessary to study the methods of identifying the modulation types of OFDM system, if or not correctly identify the modulation types is related to the correct or incorrect demodulation and information receiving.Wavelet transformation has the favorable property of time-frequency localization, while the neural network has the excellence of self-study, self-adaptation, high stabilization and error acceptability. Using neural network can improve the automatization and intelligence of recognition.Based on above, wavelet transformation is used to analyze the four signals BPSK,QPSK,16QAM,64QAM of OFDM system in this thesis, then extract the character parameters and input these character parameters to the neural network, train the neural network, export the classification result when the error meets the requirement at last. Finally I simulate the method with Matlab and analyze its performance.