Toward Scalable and Robust Indoor Tracking: Design, Implementation, and Evaluation

Feiyu Jin, Kai Liu, Hao Zhang, Joseph Kee Yin Ng, Songtao Guo, Victor C.S. Lee, Sang H. Son

Research output: Contribution to journalArticlepeer-review

15 Scopus citations

Abstract

Although indoor localization has been studied over a decade, it is still challenging to enable many IoT applications, such as activity tracking and monitoring in smart home and customer navigation and trajectory mining in smart shopping mall, which typically require meter-level localization accuracy in a highly dynamic and large-scale indoor environment. Therefore, this article aims at designing and implementing an adaptive and scalable indoor tracking system in a cost-effective way. First, we propose a zero site-survey overhead (ZSSO) algorithm to enhance the system scalability. It integrates the step information and map constraints to infer user's positions based on the particle filter and supports the auto labeling of scanned Wi-Fi signal for constructing the fingerprint database without the extra site-survey overhead. Further, we propose an iterative-weight-update (IWU) strategy for ZSSO to enhance system robustness and make it more adaptive to the dynamic changing of environments. Specifically, a two-step clustering mechanism is proposed to delete outliers in the fingerprint database and alleviate the mismatch between the auto-tagged coordinates and the corresponding signal features. Then, an iterative fingerprint update mechanism is designed to continuously evaluate the Wi-Fi fingerprint localization results during online tracking, which will further refine the fingerprint database. Finally, we implement the indoor tracking system in real-world environments and conduct a comprehensive performance evaluation. The field testing results conclusively demonstrate the scalability and effectiveness of the proposed algorithms.

Original languageEnglish
Article number8897653
Pages (from-to)1192-1204
Number of pages13
JournalIEEE Internet of Things Journal
Volume7
Issue number2
DOIs
StatePublished - Feb 2020

Bibliographical note

Publisher Copyright:
© 2014 IEEE.

Keywords

  • Algorithm design
  • Wi-Fi fingerprint
  • indoor localization
  • performance evaluation
  • trajectory tracking

Fingerprint

Dive into the research topics of 'Toward Scalable and Robust Indoor Tracking: Design, Implementation, and Evaluation'. Together they form a unique fingerprint.

Cite this