| Summary: | 碩士 === 國立交通大學 === 應用數學系所 === 106 === In wireless sensor networks, hole detection and hole healing are two major problems. Coverage holes may arise in a Region of Interests (ROI) due to the running out of batteries of sensor nodes or due to the damage of sensor nodes. Li et al. [4] proposed a tree-based coverage hole detection method and a tree-based coverage hole healing method that can effectively detect and heal coverage holes. Li et al. [4] used trees to describe the structures of coverage holes and they proposed the concept that holes that are too small can be ignored. To fulfill this concept, Li et al. [4] first obtained a Delaunay triangulation of the sensor nodes. Then, they used the radiuses of inscribed empty circles of Delaunay triangles to estimate the sizes of local coverage holes, and ignored all local coverage holes with radiuses no larger than a given threshold. In this thesis, we use the idea of [4] to estimate local coverage holes. In particular, we propose a hole-detection algorithm that does not require the assumption that the initial deployment of the sensor nodes completely covers of the boundary of the ROI, and propose a hole-healing algorithm that does not require merging local coverage holes. Do notice that reference [4] requires the above assumption and requires merging local coverage holes. Our hole-healing algorithm supports the concept that holes that are too small can be ignored; if needed, it can achieve 100% coverage of the ROI by adjusting the threshold.
|
|---|
