Optical Beam Steering in FSO Systems Supported by Computer Vision

Abstract

Free-Space Optical (FSO) communications have emerged as a viable solution to complement conventional communication technologies, mainly due to their high transmission rates, ease of implementation, low cost, and unlicensed spectrum. However, FSO technology faces significant challenges in achieving the initial alignment between the transceivers without pre-alignment or extensive manual configuration. The manual alignment process is extremely time-consuming and current automated solutions typically employ an infrared (IR) beam followed by an IR camera, which, besides increasing the transceiver cost, also requires an initial pre-convergence stage resorting to complementary coarse alignment technologies (e.g. Global Positioning System (GPS)). To tackle this challenge, we propose a novel methodology for the automatic alignment of FSO systems that does not require pre-convergence and avoids the use of costly IR cameras. The proposed alignment method is based on the application of computer vision (CV) during the transmission of a visible beacon and the utilization of pointing, acquisition and tracking (PAT) mechanisms. This solution is cost-effective, as the initial coarse alignment is carried out using visible light equipment, such as a standard CMOS camera and a visible beacon. Furthermore, it does not require any additional pre-convergence stage, since the alignment is based on characteristics acquired through CV during the transmission of the visible beacon, specifically the angle of inclination and the count of the number of pixels associated with the beacon. Additionally, the alignment process allows for a maximum angular misalignment of ±3.3 degrees between the optical antennas (OA), corresponding to a coverage area of $5.3 \text {m}^{2}$ at a distance of 20 meters. Finally, using this method, the successful alignment of an FSO system is demonstrated in a free-space link with 20 meters distance for different ambient light conditions, with real-time transmission at 16 Gbps.

Keywords

• Free-space optics
• pointing
• acquisition and tracking mechanisms
• automatic alignment
• computer vision