A two-phase ranging algorithm for sensor localization in structural health monitoring

Chunfeng Wan,Lei Zhao,Youliang Ding,Songtao Xue

doi:10.1177/1687814016685964

Abstract

Information about the positions of the sensors in sensor networks is very important, and the deployment of more and more sensors is increasing the need for automatic sensor localization. This article therefore describes a novel two-phase ranging algorithm that first obtains rough estimate of the distance to a sensor’s position using time difference of arrival or time of arrival methods and then obtains a high-resolution estimate based on the rough one using a phase-based ranging scheme. This algorithm can easily resolve the otherwise intractable integer ambiguity that often appears in localization systems, and experimental results show that it can greatly decrease the ranging error in a decentralized distance-based localization system having transmitter beacons and receivers in the nodes. Related problems such as signal filtering and multipath effect are also discussed. This algorithm can make the deployment of large numbers of sensors very simple and the determination of their positions so accurate that it would be feasible to use dense networks of sensors to monitor the structural integrity of large structures.

Highlights

Because even earthquake-resistant structures have been destroyed by earthquakes, such as Kobe earthquake in 1995, the ability of structures to withstand earthquakes needs to be checked again using rational design methods
The EDgrid inspired by the Network for Earthquake Engineering Simulation (NEES) project is the cyber infrastructure to support full-scale experiments conducted at the E-Defense facility.[1,2,3]
Because the line of sight between a sensor and a reference point cannot always be obtained in or around a civil engineering structure[8] and a small error in angle estimation can result in a very large localization error, in the work presented here, we focused on distance-based localization in which lateration is used to locate a node (Figure 1)

Summary

Introduction

Because even earthquake-resistant structures have been destroyed by earthquakes, such as Kobe earthquake in 1995, the ability of structures to withstand earthquakes needs to be checked again using rational design methods. These factors result in a small sensing area, in which case, beacons should be densely deployed in the network, which will be very inconvenient for a large-scale civil engineering monitoring system With these considerations, this article describes a two-phase ranging algorithm improving the localization accuracy of a sensor location system using the audible sound. In order to solve the integer ambiguity problem, the wavelength of the sound cannot be too small, and it should be large enough to tolerate the ranging error in the first step In this point of view, a lower frequency sound is required to be used to get a good partial cycle estimate in the second step. All the nodes are located using the range-based localization scheme

Discussion

Experiments

Conclusion