Energy-Aware Node Selection In A Distributed Sensor Network; Conference paper

摘要

This work develops a resource management strategy for a wireless sensor network of bearings-only sensors. Specifically, the resource manager determines which nodes actively sense and communicate during each snapshot in order to achieve a tolerable level of geolocalization accuracy while attempting to maximize the effective lifetime of the network. This work compares three energy-related metrics. The traditional metric that summarizes the energy usage over a single snapshot consists of the first metric. The other two metrics represent the current lifetimes of the currently active node set and the next active node set. These metrics can achieve load balancing of the nodes without resorting to computationally demanding non-myopic optimization. For any of the three metrics, the activation decision is performed in a decentralized manner over the active set of nodes. Each active node transmits just far enough to reach all the active nodes for information sharing and the potentially active nodes for information handoff. In determining the active set, partial network knowledge is considered. The partial network approach assumes that a node only knows the location of itself, the previous active set, and neighboring nodes. Simulations demonstrate the advantage of the current lifetime metrics over the more traditional energy based metric.