[Paper Review] Pedestrian Positioning Using WiFi Fingerprints and a Foot-mounted Inertial Sensor

Yang Gu, Caifa Zhou+2 more|arXiv (Cornell University)|Apr 11, 2017

Indoor and Outdoor Localization Technologies6 references4 citations

TL;DR

This paper proposes a particle filter-based fusion method combining foot-mounted inertial sensors and WiFi fingerprinting for indoor pedestrian positioning, eliminating the need for offline site surveys and reducing computational cost compared to Gaussian process methods. The approach limits drift in inertial positioning by weighting particles based on signal strength consistency, achieving accurate, real-time localization without prior knowledge of AP locations or RSS distributions.

ABSTRACT

Foot-mounted inertial positioning (FMIP) and fingerprinting based WiFi indoor positioning (FWIP) are two promising solutions for indoor positioning. However, FMIP suffers from accumulative positioning errors in the long term while FWIP involves a very labor-intensive offline training phase. A new approach combining the two solutions is proposed in this paper, which can limit the error growth in FMIP and is free of any offline site survey phase. This approach is realized in the framework of a particle filter, where each particle denotes a potential trajectory of the user and is weighted according to its consistency in signal strength space. Compared with the traditional Gaussian process based approaches, the proposed one has less computational cost and is free from any prior information in the position domain, such as the positions of access points, received signal strengths at certain positions and so on. An experiment is carried out to demonstrate the performance of the proposed approach compared to the traditional Gaussian process based approach.

Motivation & Objective

• To address the long-term drift issue in foot-mounted inertial positioning (FMIP) without relying on expensive offline site surveys.
• To eliminate the need for prior knowledge of access point locations and received signal strength (RSS) distributions in WiFi fingerprinting.
• To develop a computationally efficient indoor positioning system that maintains accuracy over extended periods.
• To integrate inertial and fingerprinting data in a unified framework that avoids Gaussian process-based modeling.

Proposed method

• A particle filter framework is used, where each particle represents a candidate user trajectory.
• Particles are weighted based on their consistency with observed WiFi signal strengths in the fingerprinting space.
• The method avoids modeling RSS as a function of position using Gaussian processes, reducing computational complexity.
• The system operates without requiring prior knowledge of AP locations or RSS values at specific points.
• The fusion of inertial measurements and real-time WiFi fingerprinting enables continuous, drift-compensated tracking.
• The particle filter dynamically updates particle weights using signal strength matching, enabling robust localization in dynamic environments.

Experimental results

Research questions

• RQ1 Can a particle filter-based fusion of foot-mounted inertial data and WiFi fingerprinting achieve accurate indoor positioning without offline site surveys?
• RQ2 How does the proposed method compare to Gaussian process-based approaches in terms of computational cost and accuracy?
• RQ3 To what extent can the particle filter framework limit the accumulation of inertial positioning errors?
• RQ4 Can the system operate effectively without prior knowledge of access point locations or RSS distributions?

Key findings

• The proposed method achieves accurate indoor pedestrian positioning without requiring any offline site survey or prior knowledge of AP locations.
• The particle filter-based approach has significantly lower computational cost than traditional Gaussian process-based methods.
• The system effectively limits long-term drift in foot-mounted inertial positioning through real-time WiFi fingerprinting correction.
• Experimental results demonstrate that the proposed method outperforms Gaussian process-based approaches in terms of localization accuracy and efficiency.