Interference Mitigation for Faster-Than-Nyquist Multiband CAP Visible Light Communication

Visible light communication (VLC) based on light-emitting diodes (LEDs) is expected to be a key technology for next-generation (6G) networks owing to the property of providing illumination and communication. However, commercial LEDs are mostly designed for illumination or display, so that bandwidth limitations and nonlinear effects induced by the devices are major bottlenecks for VLC system. Faster-than-Nyquist (FTN) signaling is an alternative to realize high spectral efficiency (SE) with low-order modulation formats by inducing nonorthogonality in the time or frequency domain, namely, time compression (TC) and frequency compression (FC). For the first time, we propose and experimentally demonstrate a FTN multiband carrierless amplitude and phase (CAP) VLC system. The inter-symbol interference (ISI) and interchannel interference (ICI) terms are analyzed theoretically, and neural network (NN)-based dirty paper coding and waveform-to-symbol decoding schemes are proposed to address the combination of ISI and ICI. The experimental results prove the feasibility of the proposed scheme. A compression ratio of 50% is achieved for the FTN 16QAM multiband CAP VLC system. The average achievable SE exceeds 6.34 bit/s/Hz using FTN 16QAM with the 3.8×10 ^-3 BER threshold.