Abstract
Underwater sensor networks (UWSN) have two major challenges: limited bandwidth and high propagation delay. In-network processing can improve the performance of UWSN data flow in the network, leading to better channel utilization. In order to do this, processing nodes have to be deployed to perform local processing, such as compression, mining, and feature extraction on the collected data. In this paper, we first develop an optimization framework based on Integer Linear Programming (ILP) for processing node deployment. We then solve the problems of processing and surface-level gateway node deployment and investigate their trade-offs. The advantage of processing node deployment on end-to-end delay and network lifetime is highlighted. Our results show that deploying processing nodes can minimize the number of surface-level gateways in addition to improving the performance of the network. Simulations are used to validate our work and show the advantages of the proposed architecture.