Supporting data collection in wireless sensor networks : from energy conserving to perpetual living

• 主要作者: Xiang, Liu
• 其他作者: Luo, Jun
• 格式: Theses and Dissertations
• 語言: English
• 出版: 2013
• 主題: DRNTU::Engineering
• 在線閱讀: http://hdl.handle.net/10356/52662
• 機構: Nanyang Technological University

Wireless sensor networks (WSNs) with their networked sensor nodes, though powerful in monitoring the physical environment by integrating the information in large scale, are deficient in energy aspect due to the resource constrained design. Aiming at energy conservation in WSNs or lifetime elongation, energy-aware protocols have been inten- sively studied by the community. Among those proposals, data collection, arguably the most fundamental function of a WSN, attracts most of the research efforts. Whereas most approaches bargain for every piece of energy in this scenario, another trend of research seeks for recharging the networked nodes via external power resources. In this thesis, we first follow the traditional way by developing an energy efficient data collection scheme, then we take a step further by supporting the network through wire- less power transfer. In this way, we strive to make WSNs from energy conserving to perpetual living. In the first part of this thesis, we devote ourselves to the design of an energy efficient data collection mechanism. Attempting to obtain full recovery of the field data, we make use of the newly developed compressed sensing and matrix completion to devise a dual-level compressed data aggregation scheme. We innovate in designing proper sparse bases with diffusion wavelets to enable high-dimension data recovery on arbitrary network topologies. We also investigate the joint routing and aggregation problem with a view to minimizing the total energy consumption. We propose both mixed integer programming and heuristic algorithms to solve this NP-complete problem. With encouraging results, we confirm that our aggregation is energy efficient and preserves data fidelity. In the second part of this thesis, we explore the possibility of supporting WSNs perpetually. Inspired by the emerging technique of wireless power transfer that promises wireless charging in mid-range distance, we envision a novel perpetual wireless sensor network (PWSN) paradigm where power is injected into the network and distributed to each node through multi-hop transfer. In this case, a WSN can survive perpetually if its energy dissipation can be replenished by power injection. Noting the significant difference between power flow and data flow, we reconsider the classic maximum flow and min-cost flow problems concerning only power or jointly with data. We analyze the problem structures and develop efficient solutions. Based on the extensive results, we gain inspiring insights for practical implementation of PWSN.