Deep Learning and LPF-Based Decision Technique in Free-Space Optical Communication Links
Year: 2025
Type: article
Abstract: This study proposes a deep learning (DL)-based adaptive threshold decision (ATD) technique for free-space optical (FSO) communications. A low-pass filter (LPF) was adopted to extract channel state information (CSI) from the received signal. However, a dynamic cutoff frequency is required for LPF under various turbulence channels and data rates. Therefore, the DL technique, a fully connected neural network (FCNN), was introduced to improve the capability of LPF-based CSI extraction. Two techniques, cascaded LPF and FCNN based CSI estimation, and the FCNN-based adaptive cutoff frequency of the LPF are proposed and evaluated in the simulation. The simulation results showed that the performance of the proposed FCNN-based adaptive cutoff frequency of the LPF technique was close to that of the traditional ATD with a perfect CSI. In addition, it exceeded the performance of cascaded LPF and FCNN-based CSI estimation techniques.
Authors: Yan Gao, Yan-Qing Hong, Chao-Yue Zhai, Xiao-Xue Ren
Institution: Shenyang University of Technology
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4412041098