可见光通信（VLC）由于其传输速度快和保密性强等诸多优势已成为了一种行之有效的水下通信技术。光在水下传播时容易发生散射现象，导致光传播的能量随传输距离的增大而减少，因此改善接收端响应度是实现水下VLC的关键和核心。为此，文章首先针对探测器前端设计了一种新型的全息波导聚光系统，可有效提升水下探测器的信噪比；然后，构建了具有弱光探测能力的低成本硫化锑薄膜探测器。在上述基础上搭建VLC系统，实验结果显示，其接收端探测器的暗电流仅为10^-7 A/cm²，3 dB带宽为220 kHz，响应度达0.2 A/W，且在蓝绿光波段的外量子效率超过70%。因此，新型硫化锑薄膜探测器结合全息波导聚光系统可以实现针对可见光的弱光探测，是一种十分有效的水下可见光接收系统。

Although the 5G wireless network has made significant advances, it is not enough to accommodate the rapidly rising requirement for broader bandwidth in post-5G and 6G eras. As a result, emerging technologies in higher frequencies including visible light communication (VLC), are becoming a hot topic. In particular, LED-based VLC is foreseen as a key enabler for achieving data rates at the Tb/s level in indoor scenarios using multi-color LED arrays with wavelength division multiplexing (WDM) technology. This paper proposes an optimized multi-color LED array chip for high-speed VLC systems. Its long-wavelength GaN-based LED units are remarkably enhanced by V-pit structure in their efficiency, especially in the “yellow gap” region, and it achieves significant improvement in data rate compared with earlier research. This work investigates the V-pit structure and tries to provide insight by introducing a new equivalent circuit model, which provides an explanation of the simulation and experiment results. In the final test using a laboratory communication system, the data rates of eight channels from short to long wavelength are 3.91 Gb/s, 3.77 Gb/s, 3.67 Gb/s, 4.40 Gb/s, 3.78 Gb/s, 3.18 Gb/s, 4.31 Gb/s, and 4.35 Gb/s (31.38 Gb/s in total), with advanced digital signal processing (DSP) techniques including digital equalization technique and bit-power loading discrete multitone (DMT) modulation format.

多输入多输出（Multiple-input-multiple-output，MIMO）可见光通信（Visible Light Communications，VLC）系统接收端需精确的信道状态信息用以解调信号，而常用的最小二乘算法对噪声敏感，估计误差较大，难以保证可靠性。基于信道稀疏特性，利用压缩感知方法进行MIMO-VLC信道估计，提出一种基于离散傅里叶变换（Discrete Fourier Transform，DFT）的稀疏度预测自适应匹配追踪（DFT Based Prediction-sparsity Adaptive Matching Pursuit，DFT-SAMP）算法。首先，通过DFT的稀疏度预测方法对信道冲激响应的稀疏度进行预估计，将估计的稀疏度作为算法初始步长，以快速逼近真实稀疏度，提高算法效率；其次，采用SAMP算法重构信道冲激响应，提高信道估计准确性，保证通信可靠性。基于2×2的MIMO-VLC系统信道估计实验结果表明，导频数为32时，本文算法相较于最小二乘算法在误码率满足前向纠错的误码率阈值（3.8×10^-3）时所需的信噪比降低4.5 dB；利用DFT-SAMP算法进行信道估计，在保证可靠性的同时，运行效率相比SAMP算法提升约69%，为MIMO-VLC系统信道估计提供了更为有效的方式。