Ph.D. Pre-Synopsis Open Seminar · Electronics and Communication Engineering
Investigations on Learning-driven Data Routing and AUV-aided Data Collection Frameworks for Wireless Underwater Sensor Networks.
Abstract
The growing importance of underwater monitoring, ocean exploration, and Internet of Underwater Things (IoUT) applications has increased the demand for reliable, energy-efficient, and sustainable communication mechanisms in Wireless Underwater Sensor Networks (WUSNs). However, the distinctive characteristics of underwater acoustic communication, including limited bandwidth, high propagation delay, dynamic channel conditions, energy constraints, and intermittent connectivity, pose significant challenges to efficient data routing and collection. Conventional routing approaches often struggle to adapt to changing network conditions, coordinate forwarding decisions among distributed nodes, and sustain network operation under constrained energy resources. Furthermore, efficient data collection from underwater sensor nodes remains challenging due to communication limitations and the need to maintain network endurance. Addressing these challenges, this research investigates learning-driven data routing and AUV-aided data collection frameworks for WUSNs. The routing investigations encompass an adaptive distributed deep Q-learning-based protocol, a cooperative multi-agent routing framework using QMIX, and a federated deep Q-learning-based approach to facilitate collaborative routing-policy learning across distributed nodes. Complementing these routing approaches, we investigate an AUV-assisted multi-modal data collection framework that integrates clustering, opto-acoustic communication, and optical wireless power transfer for reliable data retrieval and cluster-head energy replenishment. Collectively, these investigations aim to enhance routing adaptability, forwarding coordination, learning collaboration, data collection reliability, and energy sustainability, contributing towards efficient and dependable communication frameworks for next-generation underwater sensing and IoUT applications.