Elsevier

Pervasive and Mobile Computing

Volume 30, August 2016, Pages 45-57
Pervasive and Mobile Computing

Enhancing WiFi-fingerprinting accuracy using RSS calibration in dual-band environments

https://doi.org/10.1016/j.pmcj.2015.10.017Get rights and content

Abstract

Wi-Fi radio signals are commonly used to localize mobile users indoors. The use of a dual-band channel (2.4 GHz and 5 GHz) in mobile devices, however, significantly affects the accuracy of localization, because scanning all the channels requires a relatively long delay of approximately 5 seconds. During this scan time, a user may move tens of meters, depending on his or her walking speed. In this paper, we propose the self-calibration of Wi-Fi signals in a dual-band channel to solve the received signal strength (RSS) delay problem in Wi-Fi-based indoor localization. We first investigate the causes of RSS delay by analyzing Wi-Fi driver implementation and observing users’ walking behaviors. The proposed system comprises four components: a delay detector, a speed calculator, a search window manager, and a window selector. The delay detector estimates the delay in Wi-Fi scanning at each access point. The speed calculator estimates a user’s walking speed using an accelerometer. The search window manager quantifies the size of the reference fingerprints from the radio map based on the update delay and movement speed. The window selector revises the signals using the RSS compensation model. The experiment results in two buildings show that the proposed system greatly improves the accuracy of indoor localization.

Introduction

Indoor positioning techniques have actively been studied using radio signals, such as Wi-Fi, ZigBee, Bluetooth, and 3G/LTE networks. In particular, the IEEE 802.11 Wi-Fi technology is commonly used for indoor positioning because of its high popularity and penetration rate  [1], [2]. Recently, commercial mobile devices have begun to support dual-band radio frequencies (i.e., 2.4 and 5 GHz), which provide a wider range of available access points (APs) for indoor localization. However, AP scanning in a dual-band environment requires prolonged scan delays of an additional 32 channels in 5 GHz [3], [4]. We term this issue the received signal strength (RSS) delay problem. In fact, the scan time of the APs in the low band is different from that in the high band, and the RSS packet of the APs in the low band is received earlier than that in the high band  [3], [4]. This phenomenon means that a single fingerprint consists of signal strengths of APs measured at different locations when a user is moving.

The RSS delay problem is a critical issue in indoor localization because the integrity of RSS value plays an important role in calculating location. By the time the scanned result is transmitted for calculating the location, some of the RSS data is already outdated due to the scan time. For example, the RSS of AP A is received at location X, but the RSS of AP B is actually scanned at Y, a different location. This scan delay in RSS has a direct impact on the accuracy of device positioning because the error significantly increases as the user moves, especially when the user is moving quickly. The delay problem is inherently caused by the use of dual-band in WiFi environments, thus the problem exists in all types of mobile systems such as Android, iOS, Windows mobile, Tizen, etc. In order to deal with the RSS delay problem, an adequate scheme is therefore required for Wi-Fi-based indoor localization.

In this paper, we propose a self-calibration scheme that addresses the RSS delay problem in indoor localization. We investigate the fundamental causes of the RSS delay problem, and we subsequently design a system to handle the problems by analyzing Wi-Fi driver implementation and the characteristics of user movement. The proposed system comprises four components: a speed calculator, a delay detector, a search window manager, and a window selector. The system first estimates the initial position of a user using Wi-Fi fingerprints. The speed calculator estimates a user’s walking speed using an accelerometer  [5], [6]. The delay detector then estimates the updated delays of each AP in the scanned fingerprints. The search window manager uses the updated delay data and the movement speed to determine the size of the search window in the radio map. Finally, the window selector revises the RSS using the selected RSS in the radio map. After self-calibration is completed for all the APs, the system estimates the location of the device. The main contribution of our work is to solve the RSS delay problem in dual-band environments, which has not been addressed in the literature.

This paper is organized as follows. In Section  2, we describe the RSS delay problem by explaining the mechanism of Wi-Fi scanning and the scan time in both single- and dual-band environments. Section  3 provides a description of the system we developed to resolve the RSS delay problem in indoor localization. In Section  4, we evaluate our approach with experiments at two different sites and discuss the performance results. Section  5 offers a discussion of related work, and Section  6 concludes the paper.

Section snippets

Preliminary study

In this section, we describe the scanning mechanism of the IEEE 802.11 Wi-Fi technology and address the following questions:

  • How long is the scan time delay in a dual-band environment?

  • What is the main factor of this delay?

  • How much error is caused by the RSS delay in indoor localization?

Self-calibrating RSS delays

Our preliminary studies indicate that the RSS delay in a dual-band environment significantly decreases the accuracy of indoor localization. In this section, we propose a self-calibration system that addresses the RSS delay problem.

Evaluation

We conducted the evaluation in four parts. The first and second parts validate each component of the proposed system, that is, the battery capacity estimation and the application power profiling. In this section, we describe the accuracy of the prediction of the available battery time of applications and evaluate the overhead of the proposed system.

Related work

Wi-Fi-based tracking. Diverse research has been conducted to improve the accuracy of Wi-Fi-based fingerprinting. The traditional method uses Kalman and particle filters to reduce signal noise  [14]. Kjærgaard et al.  [15] proposed a calibration algorithm that minimizes RSS variances across different mobile devices. To overcome the RSS variance problem, Kim et al.  [16] proposed the use of peaks in RSS signal. Fang et al.  [17] provided a model to mitigate the impact of RSS fluctuation.

Conclusion

This paper proposes a self-calibration scheme for the RSS delay in dual-band environments. The proposed method solves the RSS delay by investigating the scanning mechanism in Wi-Fi drivers and user movement speeds. Our compensation model reduced the error distance by approximately 7.3 m in 95% of cases compared to the uncalibrated results. This improvement is meaningful for room-level localization as well as for general indoor navigation.

To the best of our knowledge, our study is the first to

Acknowledgment

This work was supported by a grant from the National Research Foundation of Korea (NRF), funded by the Korean government, Ministry of Education, Science and Technology under Grant (NRF-2014R1A2A1A11049979).

References (31)

  • P. Bahl, V. Padmanabhan, RADAR: An in-building RF-based user location and tracking system, in: Proc. 19th IEEE...
  • P. Bahl et al.

    Enhancements to the RADAR user location and tracking system. Technical Report MSR-TR-2000-12

    (2000)
  • IEEE standard, 802-11.1997, IEEE Standard for Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY)...
  • IEEE Std. 802.11a, Wireless LAN Medium Access Control (MAC) and Physical Layer (PHY) Specifications: High-speed...
  • F. Li, C. Zhao, G. Ding, J. Gong, C. Liu, F. Zhao, A reliable and accurate indoor localization method using phone...
  • R.W. Levi, T. Judd, Dead reckoning navigational system using accelerometer to measure foot impacts, US Patent, December...
  • D. Murray, M. Dixon, T. Koziniec, Scanning delays in 802.11 networks, in: Proc. of the 2007 International Conference on...
  • Android Developer, Wi-Fi Manager API on Android Developers....
  • Android Developer, ScanResult on Android Developers....
  • Tizen Native API, Wi-Fi Manager on Tizen Developer....
  • Y.-T. Li, G. Chen, M.-T. Sun, An indoor collaborative pedestrian dead reckoning system, in: Proc. IEEE 42nd Int. Conf....
  • S.H. Shin

    Adaptive step length estimation algorithm using low-cost MEMS inertial sensors

  • P. Jaccard

    The distribution of the flora in the alpine zone

    New Phytol.

    (1912)
  • D. Liu, Y. Xiong, J. Ma, Exploit Kalman filter to improve fingerprint-based indoor localization, in: Proc. 2011 IEEE...
  • M.B. Kjærgaard

    Automatic mitigation of sensor variations for signal strength based location systems

  • Cited by (9)

    View all citing articles on Scopus
    View full text