Research On MDS-Based Localization Algorithms For Wireless Sensor Networks

With its unique advantages and potential, the technology of wireless sensor networks has been identified as one of the most important technologies in the 21st century. As its supporting technologies such as wireless communication, micro-fabrication and integration, and embedded microprocessor have greatly advanced, the application of wireless sensor networks has penetrated into every aspect of human lives. Among all the related technologies, node localization is the most fundamental and key issue in the implementation and management of sensor networks. The detected information by a node is often meaningless without its knowledge of position. Therefore, to determine the location of an event or a node is one of the underlying functions of wireless sensor networks, which plays an essential role in the effectiveness of the whole network.This paper first briefly introduces the features of wireless sensor networks, such as its large scale, one-time deployment and limited power, which must be taken into account when doing relevant research or design. Then, the paper poses the problem of node localization, carefully studies current localization algorithms and summarizes existing range-based and range-free localization algorithms. Moreover, this paper analyzes the drawbacks of the traditional multidimensional scaling localization algorithm and proposes two improved MDS-based centralized localization algorithms. The first improved algorithm replaces the previous eigen-decomposition with congruent-decomposition, simplifying the steps of matrix decomposition and enhancing localization accuracy while maintaining a computational complexity of O(n3). This algorithm is suitable in scenarios where high localization accuracy is required. The second improved algorithm, using forced positive-definite Cholesky decomposition, can reduce the computational complexity to O(n2) with limited error, which fits well where real-time response is the primary concern. This paper also proposes a distributed localization algorithm which applies the jigsaw puzzle segmentation. Simulations using MATLAB show that all the proposed algorithms exhibit good performances and can realize node localization in corresponding scenarios.