Two-stage weighted centroid localization for large-scale wireless sensor networks in ambient intelligence environment

Abstract

Wireless sensor networks (WSNs) are a necessary technology for the development of ambient intelligence (AmI) applications. One of the major concerns of designing WSNs for AmI applications is the localization of the sensor nodes. This paper proposes a new range-free localization algorithm for large-scale WSNs using received signal strength (RSS) measurements. Different from most RSS-based range-free localization in WSNs, the proposed two-stage weighted centroid localization (TS-WCL) algorithm utilizes the RSS measurements between pairs of anchors, except for the RSS measurements between the unknown node and anchor nodes. In the first stage of TS-WCL, the unknown node constructs a virtual neighboring anchor list by doing half-symmetric lens presence tests. A half-symmetric lens presence test determines a residence area of the unknown node using the locations of two neighboring anchors and the corresponding RSS measurements. The centroid of the estimated residence area is then regarded as a virtual neighboring anchor of the unknown node. In the second stage, a weighted centroid formula is used to estimate the location of the unknown node based on the virtual neighboring anchor list. Simulation results show that the proposed method has better performance compared to other three typical range-free localization methods.