Research On The Localization Mechanism For Wireless Sensor Network

Rapid developments in MEMS(Micro-Electro-Mechanism System), SoC(System on Chip) and Wireless Communications have advanced the birth of Wireless Sensor Network (WSN). Each sensor node in WSN individually has the ability to make a realtime surveillance, and collaboratively achieves complex informations sensing, gathering and processing, pioneering humans to enter the time of the"network is sensor". As a novel research field in automatization, communication engineering and computer science, WSN provides the researchers a lot of challengeable issues. The issue of localization mechanism is one of these challenging tasks.The localization mechanism in WSN is divided into two parts, target localization or tracking and sensors'self-localization. This dissertation briefly introduces the backgrounds, the prospects and the actualities on the research of target tracking. The emphasis of this paper is on the basic conceptions and computing methods, i.e. trilateration, triangulation and maximum likelihood estimation in sensors'self-localization. Furthermore, according to the categorizing standard that whether the localization algorithms are based on the range or not, the performances of some classical localization algorithms in the range-based localization and the range-free localization have been analysed and summaried in detail.Based on the suggestions of range-free localization, a novel localization algorithm, Approximate Point-In-Triangulation test (APIT) has been proposed in this dissertation. A detailed description how to make an uncertain node localized and the performance analysis of the localization have been presented. To increase the performance of the proposed APIT algorithm, an Improved APIT (IAPIT) algorithm has been implemented by introducing the iterative approach and adding the counter. In addition, the performance analysis of two proposed algorithms, APIT and IAPIT, has been made by computer simulation. The simulation results show that the performances in estimation error and node coverage of IAPIT are much better than those of APIT.