Beam Tracking Aided by Complexity-Reduced MobileNetV2 for Water-Air OWC With Waves

• Published in: IEEE photonics technology letters Vol. 36; no. 24; pp. 1469 - 1472
• Main Authors: Xu, Anzi, Di, Yujie, Yue, Xiangyu, Chen, Lian-Kuan
• Format: Journal Article
• Language: English
• Published: New York IEEE 15.12.2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Acoustic beams; Acoustic distortion; Algorithms; beam tracking; Bit error rate; Cameras; Deep learning; Laser beams; Micromechanical devices; Optical distortion; Real time; Real-time systems; Tracking systems; Water-air optical wireless communication; Wireless communication
• ISSN: 1041-1135; 1941-0174
• DOI: 10.1109/LPT.2024.3485897

Wave-induced beam wander degrades communication performance, which causes a high bit error rate (BER) and packet loss rate (PLR) in water-air OWC. We propose and experimentally demonstrate a real-time beam tracking system utilizing an image pre-processing algorithm and a MobileNetV2-based model to mitigate wave-induced effects in water-air OWC. The tracking system is optimized and compared with the ResNet-based tracking system in terms of tracking and communication performance. For non-tracking cases, the PLR is significantly higher, with values exceeding 20%. With a 1.3-m air path, a zero-packet-loss link is realized by the MobileNetV2-based tracking method at a data rate of 800 Mbit/s and a wave average slope changing rate (ASCR) of 28 deg/sec.