My Ph.D. dissertation.
Underwater wireless networks provide many opportunities for exploration, environmental monitoring, and military applications. Future growth of underwater networks seems promising as much of the Earth's oceans remain unexplored. However, research and development in this field comes with unique challenges due to the new operating environment. Many of these changes act networking protocols and require new ideas to achieve the best performance from network deployments. A particularly fruitful area of research focuses on providing communication functionality to underwater devices. Limited energy and communication resources available to underwater devices make protocol optimization of prime importance.
Our work addresses this issue by providing networking support to underwater wireless networks. We do that by exploiting the unique characteristics of the underwater environment through protocol design, in contrast to much of the previous work, which focused on minimizing the degradation from environmental characteristics. Limited energy resources and low data rates focus our attention at the MAC and routing layers, as they have the greatest potential to improve network performance.
We begin our investigation of underwater networks by evaluating a hybrid MAC protocol that combines TDMA and unscheduled access to improve energy efficiency. Additionally, we explore the use of network state distribution to help in dynamic channel access. Our analysis and simulation work shows that a hybrid protocol improves energy efficiency in many cases and provides a higher throughput at a particular delay than either protocol individually. Our work also shows that devices must share and use network state properly to yield any benefit. We examine two protocols that shared state information, but only one yields a benefit over protocols that do not share state.
Our work continues as we formulate and investigate a novel channel scheduling scheme that exploits the underwater channel characteristics. We formulate the new scheduling scheme, named Staggered TDMA, as an optimization problem and explore its performance through numerical evaluation. Staggered TDMA overlaps time slots assigned to nodes to greatly reduce the time nodes spend idle. Our results show Staggered TDMA outperforms traditional TDMA and provides users with the option of focusing on energy efficiency at a lower data rate or a higher data rate with lower energy efficiency. We then develop Staggered TDMA into a distributed MAC protocol, the Staggered TDMA Underwater MAC Protocol (STUMP). When compared through numerical results, STUMP yields higher throughput and lower latency than traditional TDMA. Further exploration shows STUMP scales well with synchronization error and provides insight into how CDMA affects protocol operation.
We continue to expand upon STUMP by integrating it with a routing protocol, which results in STUMP with routing (STUMP-WR). STUMP-WR performs link selection and link scheduling in a single distributed protocol using the scheduling mechanisms designed for STUMP. Evaluation in a network simulator shows that STUMP-WR performs well in underwater environments and yields high energy efficiency and throughput and low latency when compared to TDMA and other proposed random access protocols. Our evaluation also covers design issues such as the use of CDMA and protocol convergence time.
STUMP uses a specific technique for channel scheduling, but the unique characteristics of the underwater channel allow for a spectrum of scheduling options. Therefore, we explore a range of scheduling methods to find the optimal choice for underwater networks. We evaluate five scheduling methods for use in underwater networks and evaluate their performance through numerical and simulation studies. Our work shows that scheduling links provides the best tradeoff between scheduling state overhead and protocol performance across several metrics of interest.
Download Paper PDF Format ~1.4 MB