Underwater networks are a field that has been gathering attention. Land-based methods of network construction from discovery to advanced routing are all well established. Due to the unique constraints of operating in an underwater environment, many of these tried-and-true approaches need modification if they function at all. Peer discovery and lowlevel networking have been dealt with by previous research. In this thesis we describe and evaluate eight possible routing schemes with different system-knowledge requirements. We show that with a minimal set of information it is possible to achieve near-optimal results with energy costs considerably lower than centralized optimal algorithms. We demonstrate this by constructing and evaluating a custom simulation environment in MATLAB. This is implemented in a mixed procedural and array-centric approach. Simulated networks are structured on a line topology. All nodes a spaced along the horizontal axis at a random depth. It is assumed that neighbor-discovery has been completed before the simulation starts, and all nodes have access to a global list of connected neighbors. We demonstrate the effectiveness of distributed algorithms in an ideal environment, leading to the conclusion that near-optimal results can be achieved with local information only.
| Identifer | oai:union.ndltd.org:pacific.edu/oai:scholarlycommons.pacific.edu:uop_etds-1810 |
| Date | 01 January 2012 |
| Creators | O'Rourke, Michael J. |
| Publisher | Scholarly Commons |
| Source Sets | University of the Pacific |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | University of the Pacific Theses and Dissertations |
Page generated in 0.002 seconds