| Summary: | The signal distortion caused by the limited bandwidth of Light Emitting Diode (LED) and its nonlinear electro-optic conversion severely limit the transmission performance of Visible Light Communication (VLC). Traditional Message Passing Algorithm (MPA) has limited ability to resist signal nonlinear distortion in Sparse Code Multiple Access (SCMA) -VLC. In order to solve the above problems, the paper proposes a Convolutional Neural Network (CNN) receiver, which can effectively improve the anti-nonlinear distortion ability of SCMA-VLC system by feature extraction, classification and judgment of the received signal. The results show that, in the case of strong nonlinear distortion, CNN receiver has better Bit Error Rate (BER) performance than MPA receiver, which effectively expands the dynamic range of bias voltage and reduces the Signal-to-Noise Ratio (SNR) requirements of the system.
|
|---|