Network lifetime-aware data collection in Underwater Sensor Networks for delay-tolerant applications

Abstract

Development of energy-efficient data collection and routing schemes for Underwater Wireless Sensor Networks (UWSNs) is a challenging issue due to the peculiarities of the underlying physical layer technology. Since the recharging or replacement of sensor nodes is almost impossible after deployment, the critical issue of network lifetime maximization must be considered right from the beginning of designing the routing schemes. We propose a mobile sink (MS)-based data collection scheme that can extend network lifetime, taking into account power-constrained sensor nodes, partitioned networks with geographically distant data collection points and periodic monitoring applications with delay-tolerance. Lifetime extension is achieved by mitigating the ‘sink neighbourhood problem’ and by deferring the data transmissions until the MS is at the most favourable location for data transfer. Unlike the models available for terrestrial WSNs, we consider non-zero travel time of the MS between data collection points, thus making our model more realistic for UWSNs, both connected and partitioned. The performance of the proposed mobility-assisted data collection scheme is thoroughly investigated using both analytical and simulation models. The analytical results are compared to those of existing models to assess their effectiveness and to investigate the trade-offs. Results show that, with a network size of 60 nodes, the network lifetime achieved by the proposed model is 188% higher than that of static sink model and 91% higher than that of mobile sink model (MSM). The proposed maximum lifetime routing scheme is implemented in the network simulation platform OMNET++, for validating the analytical results as well as for evaluating other performance metrics that are not tractable via analytical methods. Both analytical and simulation results demonstrate the superiority of the proposed model in capturing realistic network conditions and providing useful performance indicators prior to network deployment.