Research On Clustering Routing And Its Robustness Evaluation Of Wireless Sensor Networks

Wireless sensor network (WSN) is a low-power and self-organized network. The nodes are tiny designed and battery power supplied, which makes it hard to supply continuously sufficient energy. What's more, some inconstant factors would provide great affection to network performance. Consequently, how to enhance the energy efficient and reliability of WSN are the key questions. Focusing on these, these dissertation researches on energy efficiency and reliability of WSN during data transmitting in network layer, and builds an energy-efficient WSN based on cluster, after that, presents a high reliability routing algorithm. At the same time, the dissertation discusses the question of robustness evaluation of routing algorithm, and presents a scientific method of evaluation. On the other hand, the dissertation presents a localization algorithm for the sake of measuring the geography position of data source or event spring during routing. The details are as following:1. The technology of area partition in WSN routing is studied, and a method of virtual grid partition based on power control of two base stations is presented, which partitions the large areas of target scope into a series of virtual grid cells, where each cell is determined by multi-nodes and assigned with a unique virtual ID.2. Based on the virtual grid cells, this thesis presents a Parallel Distributed Self-Organization Clustering Protocol (PDSOCP) which builds a layer structure of cluster. It selects cluster head according to residual energy of nodes and its relative localization in each grid cells, which make the selected cluster head more reasonable. What's more, the cluster heads create routing adaptively by its grid ID. The presented research shows that the algorithm yields an improvement in saving energy and prolonging system lifetime, as well as achieving better performance.3. The routing in PDSOCP is based on the unitary mode in essentially whose reliability of data transmitting is lower. This paper presents a Grid-based High Reliability Routing (GHRR) which is designed specifically for robust data delivery in face of unreliable nodes and fallible wireless links. GHRR builds a gradient field based on virtual ID of grid cells, forwards data along a band of interleaved mesh from each source to the receiver. The GHRR performance has been evaluated in reliability, link overhead, energy dissipation etc. through both theoretic analysis and extensive simulation. Compared with other similar work, the performance of GHRR is more favorable.4. The geography position of data source or event spring plays an important role in many applications of WSN, and it would be nonsensical for these data without position information. Although PDSOCP algorithm has already mentioned it, the localization error is not small. Fortunately, Centroid Localization Algorithm based on Power Control (CLAPC) is presented in this dissertation which enhances the localization precision.5. For the sake of evaluating the adaptability of routing algorithm to dynamic environment, the dissertation provides some beneficial exploration on routing robustness evaluation. It involves analyzing the characteristics of WSN routing robustness, creating robustness evaluation model based on cloud theory, presenting corresponding methods and steps, validating its feasibility by some examples.