Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Continual Person Identification using Footstep-Induced Floor Vibrations on Heterogeneous Floor Structures

Yiwen Dong, Hae Young Noh

TL;DR

The paper addresses privacy-preserving, continual online person identification in indoor environments by leveraging footstep-induced floor vibrations on heterogeneous structures. It develops a variability analysis that decomposes sources into gait-related footstep variability and structure-related wave propagation variability, and introduces a physics-guided, linear feature transformation to suppress the dominant structural variation. This transformed feature space is paired with a non-parametric Bayesian online learner (Dirichlet Process Mixture Model) to detect and incorporate newcomers on the fly. Field experiments across wood and concrete structures with 20 participants show a 70% reduction in feature variability and 90% online identification accuracy for up to 10 individuals starting from data from a single person, highlighting a scalable, non-intrusive method for smart-building personalization. The approach has potential for robust, privacy-friendly continual identification in public facilities with sparse sensor deployments.

Abstract

Person identification is important for smart buildings to provide personalized services such as health monitoring, activity tracking, and personnel management. However, previous person identification relies on pre-collected data from everyone, which is impractical in many buildings and public facilities in which visitors are typically expected. This calls for a continual person identification system that gradually learns people's identities on the fly. Existing studies use cameras to achieve this goal, but they require direct line-of-sight and also have raised privacy concerns in public. Other modalities such as wearables and pressure mats are limited by the requirement of device-carrying or dense deployment. Thus, prior studies introduced footstep-induced structural vibration sensing, which is non-intrusive and perceived as more privacy-friendly. However, this approach has a significant challenge: the high variability of vibration data due to structural heterogeneity and human gait variations, which makes online person identification algorithms perform poorly. In this paper, we characterize the variability in footstep-induced structural vibration data for accurate online person identification. To achieve this, we quantify and decompose different sources of variability and then design a feature transformation function to reduce the variability within each person's data to make different people's data more separable. We evaluate our approach through field experiments with 20 people. The results show a 70% variability reduction and a 90% accuracy for online person identification.

Continual Person Identification using Footstep-Induced Floor Vibrations on Heterogeneous Floor Structures

TL;DR

The paper addresses privacy-preserving, continual online person identification in indoor environments by leveraging footstep-induced floor vibrations on heterogeneous structures. It develops a variability analysis that decomposes sources into gait-related footstep variability and structure-related wave propagation variability, and introduces a physics-guided, linear feature transformation to suppress the dominant structural variation. This transformed feature space is paired with a non-parametric Bayesian online learner (Dirichlet Process Mixture Model) to detect and incorporate newcomers on the fly. Field experiments across wood and concrete structures with 20 participants show a 70% reduction in feature variability and 90% online identification accuracy for up to 10 individuals starting from data from a single person, highlighting a scalable, non-intrusive method for smart-building personalization. The approach has potential for robust, privacy-friendly continual identification in public facilities with sparse sensor deployments.

Abstract

Person identification is important for smart buildings to provide personalized services such as health monitoring, activity tracking, and personnel management. However, previous person identification relies on pre-collected data from everyone, which is impractical in many buildings and public facilities in which visitors are typically expected. This calls for a continual person identification system that gradually learns people's identities on the fly. Existing studies use cameras to achieve this goal, but they require direct line-of-sight and also have raised privacy concerns in public. Other modalities such as wearables and pressure mats are limited by the requirement of device-carrying or dense deployment. Thus, prior studies introduced footstep-induced structural vibration sensing, which is non-intrusive and perceived as more privacy-friendly. However, this approach has a significant challenge: the high variability of vibration data due to structural heterogeneity and human gait variations, which makes online person identification algorithms perform poorly. In this paper, we characterize the variability in footstep-induced structural vibration data for accurate online person identification. To achieve this, we quantify and decompose different sources of variability and then design a feature transformation function to reduce the variability within each person's data to make different people's data more separable. We evaluate our approach through field experiments with 20 people. The results show a 70% variability reduction and a 90% accuracy for online person identification.

Paper Structure

This paper contains 13 sections, 8 equations, 4 figures.

Table of Contents

  1. Introduction
  2. Variability Analysis in Footstep-Induced Structural Vibrations
  3. Variability Decomposition
  4. Variability Quantification
  5. Variability Reduction Via Optimization-based Feature Transformation
  6. Physics-Guided Transformation Function Design
  7. Formulation of the Optimization Problem
  8. Online Person Identification using Dirichlet Process
  9. Evaluation
  10. Experiment Setup
  11. Evaluation Results
  12. Conclusion
  13. References

Figures (4)

  • Figure 1: The same person induces different structural vibrations, visualized in time- and frequency- domain.
  • Figure 2: Variability decomposition using 2 types of impulses: 1) ball drops, 2) footsteps. Different colors indicate different excitation locations, numbered as 1-9.
  • Figure 3: Overall variability proportion for wood and concrete structures.
  • Figure 4: Experiment layout.