To overcome challenges such as signal interference,high peak-to-average power ratio,limited bandwidth of light-emitting diodes(LED),and nonlinear effects in multiple-input multiple-output(MIMO)-orthogonal frequency-division multiplexing(OFDM)visible light communication(VLC)systems,a VLC receiver based on convolutional neural network(CNN)is proposed.The receiver can learn from the distorted signal at the receiver and the original signal at the transmitter to achieve signal demodulation in MIMO-OFDM visible light systems,effectively improving the system's ability to suppress nonlinear distortion and having lower complexity.The experimental results show that compared with the least square receiver,the CNN receiver can effectively compensate for the linear and nonlinear distortions of the signal as well as the inter-user signal interference,and the average bit error rate is improved by more than an order of magnitude.At the same time,it can effectively over-come the problem of LED bandwidth limitation,and the bit transmission rate is increased by 53％compared with the least square receiver.
关键词
多输入多输出/正交频分复用/卷积神经网络/信号补偿/可见光通信
Key words
multiple-input-multiple-output/orthogonal frequency division multiplexing/convolutional neural network/signal compensation/visible light communication
维普
万方数据