在水下无线光通信系统中，水的吸收、散射以及湍流等效应将造成了信道估计与信号检测的困难，从而导致了通信误比特率升高，甚至无法通信。针对水下复杂信道下光通信信道估计与信号检测的难题，提出了基于机器学习的信道估计与解调算法，研究其在直流偏置光-正交频分复用光通信系统中水下信道估计及信号检测性能。首先基于所提信道估计与解调算法（深度神经网络和无监督学习的k-means星座解调器）完成复杂信道频率响应、二次均衡及误比特分析的仿真建模。其次对比传统最小二乘法、线性最小均方误差信道估计算法及最小距离解调算法，完成复杂信道下光通信信噪比增益等研究。仿真结果中，在湍流闪烁指数为0.18，距离为10 m的水下信道中，当子载波为8阶正交幅度调制（Quadrature Amplitude Modulation，QAM），误比特率为10-5时，所提信道估计算法相比最小二乘法与线性最小均方误差估计分别有＞6 dB与1 dB的信噪比增益。另外采用所提信号检测算法，对比传统算法有＞1 dB的信噪比增益。仿真结果表明，所提出基于机器学习的信道估计与解调算法可提高复杂信道水下光通信的性能。研究结果为远距离高速复杂水下光通信系统设计提供一定参考。

In underwater wireless optical communication systems, the effects of water absorption, scattering, and turbulence make channel estimation and signal detection challenging, leading to increased communication bit error rates and even communication failure. To address the difficulties of channel estimation and signal detection in complex underwater channels for optical communication, a machine learning-based channel estimation and demodulation algorithm is proposed, and its performance in underwater channel estimation and signal detection in direct current biased optical-orthogonal frequency division multiplexing optical communication systems is studied. Firstly, based on the proposed channel estimation and demodulation algorithm (deep neural networks and unsupervised learning k-means constellation demodulator), simulation modeling of complex channel frequency response, second-order equalization, and bit error analysis is completed. Secondly, studies on the signal-to-noise ratio gains in complex optical communication channels are conducted, comparing traditional least squares, linear minimum mean square error channel estimation algorithms, and minimum distance demodulation algorithms. In the simulation results, in an underwater channel with a turbulence scintillation index of 0.18 and a distance of 10 meters, the proposed channel estimation algorithm provides a signal-to-noise ratio gain of >6 dB and 1 dB compared to the least squares and linear minimum mean square error estimation for 8-QAM subcarriers at a bit error rate of 10-5. Additionally, using the proposed signal detection algorithm, a signal-to-noise ratio gain of >1 dB is achieved compared to traditional algorithms. The simulation results demonstrate that the proposed machine learning-based channel estimation and demodulation algorithm can improve the performance of complex underwater optical communication channels. The research results provide a reference for the design of long-distance, high-speed complex underwater optical communication systems.

TN929.1

国家自然科学基金项目（面上项目，重点项目，重大项目）