基于深度学习的井下信号调制识别研究

Research on Coal Mine Wireless Signal Modulation Recognition Based on Deep Learning

为识别矿井Nakagami-m衰落信道下的无线信号调制方式,研究了基于深度学习的端到端调制识别方法。首先对接收端IQ信号提取实部和虚部数据作为数据集,并搭建组合深度神经网络模型(CLDNN)对11种井下无线信号进行识别。仿真结果表明,当信噪比(SNR)为0时,平均正确识别率为75.3%,当SNR为5 dB以上时,平均正确识别率可达到92.4%以上,相比于经典的深度学习调制识别方法,所提出的端到端深度神经网络模型可以更准确识别矿井无线信号。

In order to identify the modulation mode of wireless signal in Nakagami-m fading channel,an end-to-end modulation recognition method based on deep learning is studied.Firstly,the real part and imaginary part data are extracted from the IQ signal at the receiving end as the data set,and build a combined deep neural network model(CLDNN) to identify 11 types of underground wireless signals.The simulation results show that when the signal-to-noise ratio(SNR) is 0,the average correct recognition rate is 75.3%,and when the SNR is more than 5 dB,the average correct recognition rate can reach more than 92.4%.Compared with the classical deep learning modulation recognition method,the proposed end-to-end deep neural network model can identify mine wireless signals more accurately.