The Study of Fractal Geometry in Wireless Sensor Networks

• Main Authors: Zheng-Da Xie, 謝政達
• Other Authors: Chi-Chang Chen
• Format: Others
• Language: zh-TW
• Published: 2016
• Online Access: http://ndltd.ncl.edu.tw/handle/au7y59

碩士 === 義守大學 === 資訊工程學系 === 104 === There are a growing number of studies focused on the fractal geometry in recent years. Fractal geometry is a disciplinary research of non-integer dimension graphics, and all of them have a kind of nesting or recursive structure. Previous studies found that curves with seemingly messy irregularity in the world are actually chaos with order. Scientists have found that there are rules to produce them. This paper attempts to use the characteristics of nesting or recursive structure of fractal geometry to the routing paths of wireless sensor networks (WSNs). We selected two space filling curves, Node-Gosper Curve and Moore curve, as the goal of our study. Node-Gosper Curve is based on node-replacement curve and its fractal dimension is two. The order one of Node-Gosper Curve consists of seven segments, and when its order grows, it can fill a similar hexagonal area. To be a routing path of a mobile sink in a WSN, we can adjust the order of Node-Gosper Curve in accordance with the transmission range, size of sensing area, parameters, etc. It ensures that all sensors can be visited in the sensing area, and thus helps sensor’s self-localization, collecting and transmitting information, etc. Most of the space-filling curves have no self-loop. Thus, the distance between the start point and the end point is very large, which causes a mobile beacon node needs to spend extra path from the end point returning to the start point. However, Moore curve have the self-loop characteristic, its start point and the end point are very close. In this thesis, Moore curve was applied to the mobile beacon path planning in which the mobile beacon follows the path to travel entire sensing area and stops at the center of each square cluster to collected information from the nodes with sensing events. With the self-loop characteristic of Moore curve, we expected the Moore curve can more quickly get to each sensor to collect information than other space-filling curves, and therefore the Moore curve can reduce sensor’s transmission delay.