Protocols for Air-water Communication and Underwater Localization using Nonlinear Optoacoustic Links

Abstract

While numerous studies have been dedicated to tackling the challenges of underwater communication and localization, little attention has been paid to effectively establishing direct communication links between aerial and deep underwater nodes (UWN). This dissertation exploits the optoacoustic effect to devise a new solution for directly reaching and localizing underwater nodes from the air. The contribution mitigates the shortcomings of conventional methods which require the involvement of a surface gateway with dual modems to establish radio and acoustic links in the air and underwater, respectively. Deploying such a gateway introduces logistical challenges and security vulnerabilities. Unlike radio, acoustic, and visual light, optoacoustic links involve two distinct signal types, optical (laser beam) in air and acoustic in water. Therefore, the laser beam needs to be modulated such that the resulting acoustic signals can be accurately demodulated to retrieve the data. Such a communication technique is unique. To address challenges in optoacoustic communication, we propose two modulation techniques: Vapor Cloud Delayed-Differential Pulse Position Modulation (VCD-DPPM) and Optical Focusing-based Adaptive Modulation (OFAM). OFAM includes both stationary (1D) and dynamic (3D) focusing for a single laser transmitter, providing stable signal generation even with vapor cloud buildup. The validation results show that VCD-DPPM, OFAM-1D, and OFAM-3D achieve data rates approximately 5.12, 6, and 4.4 times higher than on-off keying (OOK). These techniques also improve power efficiency by 137% over OOK. Machine learning techniques have also been leveraged in the demodulation process for increased robustness, where 94.75% demodulation accuracy is achieved with the Random Forest model. To handle multipath interference in high data rate applications, we have developed a deep learning approach using U-Net for signal equalization and ResNet for symbol detection, achieving 96.6% and 91.7% accuracy, compared to 72.94% and 65.30% with traditional peak detection. The proposed air-water wireless communication protocols are further leveraged to extend the GPS service to the underwater environment by remotely localizing UWN. In our approach, GPS coordinates are transmitted from the air to the UWN via creating an underwater temporary acoustic transmitter (plasma) through the optoacoustic process. We analyze the process of controlling the shape and size of the plasma to control the acoustic signal duration and directivity. First, we have developed a Receive Signal Strength (RSS) based localization method using shorter-more spherical shaped plasma. However, the generated acoustic signal shows more directivity as the plasma shape elongates to achieve higher localization range. Therefore, we have utilized the directive nature of the signal and developed a fully connected deep neural network (DNN) based localization method by determining the receiver angle relative to plasma. The simulation results with laboratory constructed dataset show the effectiveness of our approach. Both of our approaches achieve better accuracy compared to traditional techniques without using surface or underwater anchor nodes.