Localization Study Using DV-Hop/Min-Max Approach with Increased Boundary for Wireless Sensor Networks

• Main Authors: Kai-Da Lee, 李凱達
• Other Authors: 李俊賢
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/43jm58

碩士 === 國立臺北科技大學 === 電機工程系研究所 === 101 === In application of wireless sensor network (WSN), analysis could be only conducted when the locations of sensor nodes have been calculated. Therefore, it is critical for WSN to be able to localize sensor nodes within a reasonable cost. Existing WSN localization algorithms can be categorized into two types: Range-based and Range-free. Although the range-based algorithm has better accuracy than the Range-free algorithm, but it is relatively costly, however, the Range-free method has problems such as excessive localization deviation and computation. Therefore, by basing on the Distance Vector-hop/Lateration (DV-hop/Lateration), this paper suggests a hybrid localization approach that combined both Range-based and Range-free algorithms. The proposed approach is split into distance estimation and position estimation according to the structure. First, the paper uses DV-hop, along with the fact that there was a blind node within a hop count for the Received Signal Strength Indicator (RSSI), to calculate the distance between the blind node and each anchor node in the distance estimated. This study also detected the number of anchor nodes in the environment to determine if grey correlation method should be used to calculate the average hop-size (HopSize), so as to reduce the estimated error in distance without increasing too much computation load. Then, based on the Min-Max method, we consider increased distance to the estimation of the blind node coordinates in the position estimated. This approach will leave higher percentage in the over lapping boundaries established around the anchor nodes to increase the localization accuracy. This could avoid complex calculation using Lateration and Least Square methods for position estimations, which reduces the amount of calculation of the localization system. According to simulation results, the proposed approach could effectively reduce localization error and operation of WSN.