Distribution-Adaptive Quantization for Ultra-Low-Resolution Transmitter in IM/DD VLC Links

We propose and experimentally demonstrate a distribution-adaptive quantization scheme, termed the fuzzy-minimized mapping (FMM) quantizer, for high-speed ACO-OFDM VLC transmitters employing low-resolution DACs. The FMM quantizer combines fuzzy c-means (FCM) clustering with minimum mean-square error (MMSE) refinement to adaptively optimize quantization under non-Gaussian signal distributions. Experimental results over a 1 m free-space VLC link using a red VCSEL transmitter show that FMM outperforms conventional Lloyd–Max and round-off (RO) quantizers, achieving a BER of (2.3 × 10⁻³) for 16-QAM at 0.75 GBaud with 2-bit resolution and supporting 64-QAM at 0.625 GBaud (3-bit) with a BER of (2.8 × 10⁻³). Constellation diagrams further illustrate these gains, confirming the FMM quantizer’s practical effectiveness for Gbps-level VLC with low-resolution transmitter hardware.