Design and hardware implementation of a robust deep vision-based FSO-PAT system
Applied Optics · 01.07.2026 · 65 (19) · 6469
DOI: 10.1364/ao.595667 ↗Kurz gefasst
Fachartikel in der Zeitschrift Applied Optics, online erschienen am 22.06.2026. Verfasst von Erfan Kohantorabi und 1 weiteren Autor:innen. Die Arbeit nennt 33 Literaturangaben.
Abstract (Originalsprache)
Free-space optical (FSO) communications provide substantial bandwidth without licensing requirements, yet their dependability is markedly impaired by beam misalignment stemming from atmospheric turbulence and structural vibrations. Sustaining a consistent line-of-sight is essential to mitigate connection disruptions. This study delineates the development and deployment of a resilient, economical vision-centric pointing, acquisition, and tracking framework employing off-the-shelf components, including a Raspberry Pi and Picamera2 module. Diverging from conventional centroid-tracking approaches, the proposed framework integrates a YOLO-driven deep learning methodology for precise target localization, conferring resilience to ambient interference and fluctuating illumination. To counteract the characteristic mechanical instability and oscillatory “hunting” in standard servo assemblies, a discrete regulation paradigm is adopted. This paradigm merges a calibrated proportional (P) controller with an adaptive “Move-and-Stop” activation protocol, effectively suppressing minor oscillations by disengaging servo torque upon target attainment within a designated tolerance zone. Moreover, a latency-minimized software structure is utilized to refine the image capture and computation workflow, assuring real-time efficacy and obviating video delays. Empirical assessments affirm that the framework attains dependable target fixation with limited overshoot, offering a practical, budget-conscious alternative for adaptive FSO alignment.
Zitieren
APA
Kohantorabi, E., & Amirabadi, M. A. (2026). Design and hardware implementation of a robust deep vision-based FSO-PAT system. Applied Optics, 65(19), 6469. https://doi.org/10.1364/ao.595667
Vancouver
Kohantorabi E, Amirabadi MA. Design and hardware implementation of a robust deep vision-based FSO-PAT system. Applied Optics. 2026;65(19):6469. doi:10.1364/ao.595667
BibTeX
@article{Kohantorabi2026,
author = {Kohantorabi, Erfan and Amirabadi, Mohammad Ali},
title = {Design and hardware implementation of a robust deep vision-based FSO-PAT system},
journal = {Applied Optics},
year = {2026},
volume = {65},
number = {19},
pages = {6469},
doi = {10.1364/ao.595667}
}
EUVIMED