Adaptive Strategies for Energy Conservation in Wireless Sensor Networks

In the last years wireless sensor networks (WSNs) have emerged as an enabling technology for a wide range of real-life applications. Among them we can mention environmental monitoring, security/surveillance and industrial control. In a typical scenario, a number of tiny sensor nodes sample a quantity of interest in a sensing field. Each node can acquire physical data (e.g., temperature, humidity, and so on), process them locally and send the results to one or more collection points (usually referred to as sinks or base stations) through wireless communication. As sensor nodes are battery powered, their energy budget is very limited. Thus, they need mechanisms to reduce energy consumption. In addition, since they are designed to run unattended, they have to provide self-organizing and self-configuring features, in order to adapt to the operating conditions without any external intervention. In this thesis we address the issues of both energy efficiency and auto-configuration in WSNs. Specifically, we design and evaluate mechanisms for reducing the energy consumption due to both radio and sensors (i.e. transducers). From one side, we develop a sleep/wakeup protocol which is fully distributed and can adapt to the current operating conditions autonomously. From the other side, we devise an algorithm for dynamically tuning the frequency of acquisitions in order to match the actual demands of the monitored signal. As a result, both solutions can not only adapt to the actual needs, but they can also achieve a more efficient usage of energy resources, resulting in an overall low energy consumption. Finally, we analyze the problem of energy-efficient data collection where mobile elements are exploited for data gathering in WSNs. We define protocols for energy-efficient data collection, and we carry out a performance evaluation in order to obtain the best operating parameters to use in specific scenarios.