Super Gaussian Kernel-aided Deep Neural Network Equalizer Utilized in Visible Light Transparent Transmission from Free-space to Fiber.

Qiuyi Lu,Jiang Chen,Yiheng Zhao,Nan Chi
DOI: https://doi.org/10.1109/wcsp49889.2020.9299871
2020-01-01
Abstract:In this paper, we experimentally demonstrate a novel Super Gaussian kernel-aided deep neural network (SGK-DNN) equalizer utilized in blue LED CAP 16QAM visible light transparent transmission from 10cm free-space to a 100m fiber. The SGK-DNN equalizer can effectively mitigate the strong nonlinear distortion. The introduce of Super Gaussian kernel can reduce the required training iterations to 55%. And the data rate closes to 900Mbps with the bit error rate (BER) below the 7% forward error correction (FEC) limit of 3.8 × 10-3. As far as we know, this is the highest data rate ever reported for blue LED visible light transparent transmission from free-space to fiber.
What problem does this paper attempt to address?