Free-space optical communications systems suffer from turbulent propagation of light through the atmosphere, attenuation, and receiver detector noise. These effects degrade the quality of the received state, increase cross-talk, and decrease symbol classification accuracy. We develop a state-of-the-art generative neural network (GNN) and convolutional neural network (CNN) system in combination, and demonstrate its efficacy in simulated and experimental communications settings. Experimentally, the GNN system corrects for distortion and reduces detector noise, resulting in nearly identical-to-desired mode profiles at the receiver, requiring no feedback or adaptive optics. Classification accuracy is significantly improved when these generated modes are demodulated using a CNN that is pre-trained with undistorted modes. Using the GNN and CNN system exclusively pre-trained with simulated optical profiles, we show a reduction in cross-talk between experimentally-detected noisy/distorted modes at the receiver. This scalable scheme may provide a concrete and effective demodulation technique for establishing long-range classical and quantum communication links.

自由空间光通信系统遭受光在大气中的湍流传播，衰减和接收器检测器噪声的困扰。这些影响降低了接收状态的质量，增加了串扰，并降低了符号分类的准确性。我们结合开发了最新的生成神经网络（GNN）和卷积神经网络（CNN）系统，并展示了其在模拟和实验通信环境中的功效。实验上，GNN系统可校正失真并降低检测器噪声，从而在接收器处产生几乎与所需模式相同的轮廓，无需反馈或自适应光学器件。当使用未经失真模式预训练的CNN对这些生成的模式进行解调时，分类精度会得到显着提高。使用专门用模拟光学轮廓进行预训练的GNN和CNN系统，我们显示了接收器上实验检测到的噪声/失真模式之间的串扰减少。该可扩展方案可以提供用于建立远程经典和量子通信链路的具体且有效的解调技术。