Research On The Energy-efficient Wireless Sensor Networks

Wireless sensor networks (WSNs) consist of large numbers of sensor nodes distributed in a certain area, organizing themselves into an Ad Hoc-like wireless network. The task of WSNs is to collect physical information and transmit it to one or several sinks. The resources in sensor nodes are usually limited. The nodes are always powered by batteries, which are impractical to be replaced or recharged. So the energy efficiency is very important in WSNs. This thesis focuses on the key technologies in WSNs to improve energy efficiency, dealing with three subjects: Media Access Control (MAC), mobile WSNs (mWSNs) and information compression based on correlations.Wireless communication is the main cost of energy in WSNs. MAC protocols control the communication module and influence the energy efficiency of WSNs significantly. In this thesis MAC protocols in WSNs are introduced and two new MAC protocols are presented: SS-MAC (Synchronized Sensor MAC) and AS-MAC (Adaptive Schedule MAC). To solve the problem of boundary nodes in S-MAC (Sensor MAC), that is, the problem that the boundary nodes between clusters in S-MAC consume more energy than the inner nodes, SS-MAC is presented to incorporate clusters with the criterion of node numbers of clusters. With SS-MAC, the global synchronization of schedules is achieved and the boundary nodes are eliminated. The energy efficiency in SS-MAC and the connectivity of the networks are improved. In AS-MAC, the influence of listen/sleep cycles in sensor nodes to the energy efficiency, packet latency and throughput is analyzed. A new MAC protocol is presented, introducing the adaptation of the cycles in MAC layer to the traffic, which improves the performances of networks.Former WSNs are almost static. But now, WSNs with mobile nodes are to be applied, which bring forth challenges to network design. The performance of mWSNs is analyzed in three aspects: MAC, routing and clock synchronization. In MAC for mWSNs, some algorithms based on S-MAC are presented first, to improve the performance of WSNs with some mobile nodes. A new MAC protocol, MAMAC (Mobility-Adaptive MAC), is also introduced for the scenario where there are more nodes move. In routing protocols for mWSNs, the related works are introduced and the main idea in routing design is presented. About synchronization, an algorithm called Diffusion-based method is analyzed in mWSNs and two algorithms are presented to speed up the convergence of clock.There is correlation of data in adjacent sensor nodes in WSNs. Utilizing the correlation to compress the sensor data is a main idea in WSNs' design and research to improve the energy efficiency. Data aggregation and distributed source coding (DSC) are two schemes to realize the idea, and they are introduced in this thesis. The network architecture influences the compression efficiency significantly with these schemes. In data aggregation, it is network layer, that is, the routing scheme which affects the efficiency mainly. In DSC, it is in application layer, deciding which nodes are to transmit the side information, and which nodes are to compress according the correlation. We focus on the research on network architecture with DSC. A practical dynamic clustering scheme is presented, with which the network is divided to clusters according to the topology. Nodes only compress their data with the correlation in the cluster. This scheme is more practical, and still with good compression efficiency.Lastly the future works are proposed, and the underwater WSNs (UWSNs) and the video sensor networks (VSNs) are simply introduced. The cross-layer design is introduced, too.