Bandwidth-Aware Routing Tree (BART) for Underwater 3-D Geographic Routing

Abstract

The limited bandwidth and power resources in underwater sensor networks have made geographic routing a favorite choice. While many detouring strategies in geographic routing do not work for 3-D underwater network topology, spanning tree routing can efficiently find a detour when used for such networks. However, the effectiveness of tree routing depends largely on the quality of the pre-constructed spanning tree. Most existing spanning tree algorithms build trees in a top-down and centralized fashion and do not consider the available bandwidth in the network, and may create trees with poor routing performance. In this research, we propose a novel spanning tree, named Bandwidth-Aware Routing Tree (BART), that is constructed completely in a bottom-up fashion and with available bandwidth in mind. Simulation results show that compared with other spanning trees, BART has much fewer conflicting hulls and higher path throughput, thus leads to better routing performance in 3-D underwater sensor networks.