DBTF-Net: dual-branch decoupled network for laser-spot restoration and channel sensing in hybrid turbulence–fog FSO systems
yuxuan yan, Hui Yang, Zhe Niu, Qiuyan Yao, Mingyuan Wu, Buzheng Wei, Jie Zhang · Journal of Optical Communications and Networking · 2026
In recent years, free-space optical (FSO) communication technology has become a core solution for building next-generation satellite internet and achieving global seamless coverage, owing to its high bandwidth and strong anti-interference capability. However, in complex atmospheric channels, the nonlinear coupling between turbulence and fog produces hybrid degradation, which severely distorts the received laser spots and makes accurate restoration difficult. In this paper, we propose a dual-branch decoupled network (DBTF-Net) for laser-spot restoration and channel sensing in turbulence–fog mixed FSO systems. This model utilizes dual-branch decoders to separately handle turbulence-induced phase distortion and fog-induced intensity attenuation while incorporating a cross-branch attention mechanism to achieve decoupling separation of physical effects. Simulation results show that, compared with other models, the proposed DBTF-Net demonstrates significant advantages in laser-spot restoration accuracy, channel parameter inversion capability, and adaptability to complex environments, effectively solving the hybrid degradation problem caused by the combined effects of turbulence and fog.