Research on QPSK Intelligent Receiver Model Based on Deep Learning
Aimed at the interference factors of noise in communication channels and poor performance of quadrature phase shift keying(QPSK)receiver demodulating received signals.In view of the problem.a QPSK intelligent receiver model based on deep learning is studied.The QPSK intelligent receiver model is composed of the long and short-term memory(LSTM)neural network and fully connected layer.With the help of the memory structure in the recurrent neural network,LSTM is also used to extract the char-acteristic of the temporal correlation for the received signal.Simulation experiments is conducted under the condition of a signal-to-noise ratio of 0 to 7 dB,the experimental results show that under the influences of Gaussian white noise,in-phase and quadrature im-balance,and frequency deviation interference factors,and compared with the traditional hard decision method,the bit error rate of the proposed QPSK intelligent receiver model with the signal-to-noise ratio of 0 to 7 dB is significantly reduced.Among them,the bit er-ror rate of QPSK intelligent receiver model at the signal-to-noise ratio of 7 dB is as low as 0.010 9％,which is only about 1/7 of the bit error rate of the traditional hard decision method.In the conditions of frequency deviation and IQ imbalance,the bit error rate of QPSK intelligent receiver model at the signal-to-noise ratio of 7 dB is as low as 0.014 7％and 0.019 8％,respectively,which are much lower than the bit error rate of the traditional hard decision method under the same condition.Therefore,the proposed QPSK intelligent receiver model can significantly improve the detection performance of the receiver.
deep learningLSTM neural networkfully connected layerQPSK modulationintelligent receiver
万方数据
维普
