Energy efficiency in wireless sensor networks

Wireless sensor networks are one of the most promising technologies and are believed to be one of the emerging technologies that would change the world! Because sensor nodes are battery powered and have limited resources, energy efficiency is one of the most critical problems in wireless sensor networks. In this dissertation, we study how to efficiently conserve energy consumption and maximize network lifetime for the coverage problem and the virtual backbone construction. The coverage problem is one of the important issues in wireless sensor networks. It reflects the quality of service that can be provided by a particular sensor network. This dissertation aims on the energy efficient coverage problem. Since the number of deployed sensors is much higher than the optimum required, one possible way to efficiently conserve energy is to organize sensors into a maximum number of set covers that are activated successively. We prove that this problem is NP-complete and propose several different algorithms, both centralized and localized. In this dissertation, we also consider the bandwidth limitation for the coverage problem as well as introduce the relation between the coverage and the domatic number problems. In wireless sensor networks, there is no fixed or pre-defined infrastructure. To simplify the routing and network topology maintenance as well as to reduce the energy consumption in the communication, it is very desirable to find a small set of sensor nodes serving as the virtual backbone. It is possible to approximate a virtual backbone by constructing a Connected Dominating Set (CDS) of a graph representing a network. The problem of finding a minimum CDS is NP-hard. In this dissertation, we propose several algorithms for different network scenarios. If all nodes have the same transmission ranges, we model a network as a unit disk graph and propose a (5.8 + In 4)-centralized algorithm, which can be implemented as a distributed algorithm. For the asymmetric networks, we model them as disk graphs and introduce a strongly CDS problem in disk graphs. Based on this novel model, we propose several constant algorithms.