The Localization In Wireless Sensor Networks

Localization is a fundamental problem in the wireless sensor networks. A precise localization result is not only useful in the applications of wireless sensor networks like area monitoring, but also important for the research of wireless sensor networks like routing. As for the limitation of sensory nodes in the aspects of battery power, computation ability and hardware devices etc., only a few of nodes can get the localization information of themselves. All the other nodes have to be localized by a certain localization algorithm in a style of cooperation. For the variance of environment, a localization algorithm always has to fight against the interference from environment during its course of obtaining a precise localization result, which is the main goal of all localization algorithms.Supported by CAS Knowledge Innovation Program, this dissertation focuses on the localization technologies in wireless sensor networks. The major research work is as below:1) An adaptive fuzzy theory based localization algorithm for wireless sensor networks: Recent localization algorithms can be mainly divided into two classes: range-based algorithms and range-free algorithms. As the former kind of algorithms mainly uses range information including unpredictable environment interference, the localization results often tend to be unstable. Correspondingly, since the latter kind of algorithms often takes the connectivity information into consideration, the localization results usually become imprecise. To fill the gap between range-based algorithms and range-free algorithms, in this paper we propose an adaptive localization algorithm based on fuzzy theory (fuzzy localization, FL). By fuzzing the ranging information, FL Algorithm reduces the impaction of environment. By selecting fuzzy parameter adaptively, it gains better stability and precise than traditional method. The simulation results show that our scheme gains 31% and 6% more accurate and stable than DV-Hop and Spring in sparse networks, respectively.2) A fuzzy theory based localization algorithm for mobile wireless sensor networks: Current localization algorithm for mobile wireless sensor networks are mostly base on Monte Carlo method. Such Monte Carlo based algorithms usually require heavy communications or lead to an imprecise localization results. In this paper we propose a localization algorithm based on fuzzy theory for mobile wireless sensor networks (fuzzy Monte Carlo localization, FMCL). By fuzzing the observed data, a much more precise localization result can be achieved as an input of Monte Carlo algorithms while only a few of extra communications are introduced. Compared with MCL and MSL algorithms, our scheme is about 20% and 13% more accurate in the localization results.3) Simulation platform for localization algrithms of wireless sensor networks: Compared with the large numbers of localization algrithms, there are only a few of unified platform for them. This status quo encourages us to design a unified platform to evaluate the performance of different localization algorithms. In this work, we design and implement a simulation platform with simple user interface and unified programming interface for localization algorithms of wireless sensor networks, which is very useful for the research and development of localization algorithms.