Robust localisation algorithm for large scale 3D wireless sensor networks

Abstract

Nodes positioning has recently been of great interest in wireless networks owing to its crucial role in many applications. In wireless sensor networks (WSNs), the task of localising sensor nodes with unknown position is important for efficient network configuration and operation. This challenge has stimulated research of various localisation algorithms. In this paper we propose robust localisation algorithm for large scale three-dimensional (3D) WSNs based on multidimensional scaling (MDS). Our approach has two main improvements over classical MDS algorithm. Firstly, it uses heuristic approach for distance matrix calculation, and secondly, it applies Levenberg-Marquardt (LM) method for absolute map refinement using received signal strength (RSS) measurements. Furthermore, the performance of the proposed approach is compared to other 3D WSN localisation techniques and it is shown that the proposed approach outperforms other techniques for 3D localisation.