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

Feiyu Jin, Kai Liu*, Hao Zhang, Joseph K Y NG, Songtao Guo, Victor C.S. Lee, Sang H. Son

*Corresponding author for this work

Research output: Contribution to journalJournal articlepeer-review

15 Citations (Scopus)

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
Publication statusPublished - Feb 2020

Scopus Subject Areas

  • Signal Processing
  • Information Systems
  • Hardware and Architecture
  • Computer Science Applications
  • Computer Networks and Communications

User-Defined Keywords

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

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