| Wireless Sensor Networks (WSNs) are an emerging technology, recently finding extensive application in scientific and military surveillance. Typical WSNs comprise hundreds of tiny, severely energy constrained (battery operated) sensor nodes that are statically deployed to collaborate and accomplish common application tasks. The effectiveness of a WSN depends on its efficiency in using the limited energy supply, and thus, energy conservation is a forefront research area for sensor network technology.; WSNs are largely characterized by short-range multi-hop radio communications, which provides abundant scope for routing-level energy conservation schemes for sensor networks. WSNs are often densely deployed for robustness against node and link failures, and only a sub-set of nodes is required to be operating at any instant to satisfy the application needs. This dissertation work aims to exploit such connection redundancies in sensor deployments, using network level adaptations, and achieve energy efficiency in two different network mechanisms: (1) Reducing the energy expenditure of radio communications---This research develops Fuzzy Diffusion, an energy-adaptive data forwarding scheme that reduces the total amount of data transmissions in dense, high traffic WSNs (conservative routing), which proportionally decreases the energy costs of radio communications. The results show that fuzzy diffusion achieves 12-25% increase in simulated network lifetime, without degrading the user data supply. (2) Reducing the overall energy expenditure of sensor nodes, especially during idle periods---For the current sensor technology, the node energy consumption rate during idle state is in the same order of magnitude as during active radio communications. "Sleep" is an effective strategy to reduce the overall energy consumption of a WSN, since it not only reduces the idle time operation of nodes, but also reduces the total network radio transmissions (sleeping nodes do not participate in data exchanges). Sleep becomes extremely important for low traffic sensor applications (e.g. threat surveillance), where the sensing nodes are idle most of the time and the idle-mode energy consumption would dominate the overall energy expenditure.; This dissertation develops Topology and Energy Adaptive, Non-synchronous (TEAN) sleep, a network-level sleep coordination scheme for WSNs that achieves long durations of continuous node sleep, while ensuring consistent network coverage and connectivity for reliable surveillance. Further, the performance of TEAN-sleep is precisely analyzed using a family of sensor-oriented a-metrics, a unified representation of energy conservation and connection reliability. Simulations, under a variety of traffic scenarios, predict 70% average improvement in the performance of TEAN-sleep coordinated networks as compared to the sleep-deprived networks, while ensuring close to 100% network connectivity. |
