Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Why Commodity WiFi Sensors Fail at Multi-Person Gait Identification: A Systematic Analysis Using ESP32

Oliver Custance, Saad Khan, Simon Parkinson

TL;DR

The paper investigates whether hardware constraints or algorithmic limitations prevent reliable multi-person gait identification using WiFi CSI from commodity ESP32 devices. It systematically compares six signal separation methods across seven 1–10 person scenarios and two environments, introducing novel diagnostic metrics to quantify intra-subject variability, inter-subject separability, and performance degradation. Results show uniformly low accuracy (45–56%) with no significant method advantage, driven by substantial intra-subject variability, limited inter-subject distinguishability, and environmental unpredictability. The study concludes that hardware limitations set a hard ceiling for commodity WiFi sensing of multiple pedestrians, suggesting future progress requires upgraded sensing hardware or alternative modalities rather than marginal algorithmic refinements.

Abstract

WiFi Channel State Information (CSI) has shown promise for single-person gait identification, with numerous studies reporting high accuracy. However, multi-person identification remains largely unexplored, with the limited existing work relying on complex, expensive setups requiring modified firmware. A critical question remains unanswered: is poor multi-person performance an algorithmic limitation or a fundamental hardware constraint? We systematically evaluate six diverse signal separation methods (FastICA, SOBI, PCA, NMF, Wavelet, Tensor Decomposition) across seven scenarios with 1-10 people using commodity ESP32 WiFi sensors--a simple, low-cost, off-the-shelf solution. Through novel diagnostic metrics (intra-subject variability, inter-subject distinguishability, performance degradation rate), we reveal that all methods achieve similarly low accuracy (45-56\%, $σ$=3.74\%) with statistically insignificant differences (p $>$ 0.05). Even the best-performing method, NMF, achieves only 56\% accuracy. Our analysis reveals high intra-subject variability, low inter-subject distinguishability, and severe performance degradation as person count increases, indicating that commodity ESP32 sensors cannot provide sufficient signal quality for reliable multi-person separation.

Why Commodity WiFi Sensors Fail at Multi-Person Gait Identification: A Systematic Analysis Using ESP32

TL;DR

The paper investigates whether hardware constraints or algorithmic limitations prevent reliable multi-person gait identification using WiFi CSI from commodity ESP32 devices. It systematically compares six signal separation methods across seven 1–10 person scenarios and two environments, introducing novel diagnostic metrics to quantify intra-subject variability, inter-subject separability, and performance degradation. Results show uniformly low accuracy (45–56%) with no significant method advantage, driven by substantial intra-subject variability, limited inter-subject distinguishability, and environmental unpredictability. The study concludes that hardware limitations set a hard ceiling for commodity WiFi sensing of multiple pedestrians, suggesting future progress requires upgraded sensing hardware or alternative modalities rather than marginal algorithmic refinements.

Abstract

WiFi Channel State Information (CSI) has shown promise for single-person gait identification, with numerous studies reporting high accuracy. However, multi-person identification remains largely unexplored, with the limited existing work relying on complex, expensive setups requiring modified firmware. A critical question remains unanswered: is poor multi-person performance an algorithmic limitation or a fundamental hardware constraint? We systematically evaluate six diverse signal separation methods (FastICA, SOBI, PCA, NMF, Wavelet, Tensor Decomposition) across seven scenarios with 1-10 people using commodity ESP32 WiFi sensors--a simple, low-cost, off-the-shelf solution. Through novel diagnostic metrics (intra-subject variability, inter-subject distinguishability, performance degradation rate), we reveal that all methods achieve similarly low accuracy (45-56\%, =3.74\%) with statistically insignificant differences (p 0.05). Even the best-performing method, NMF, achieves only 56\% accuracy. Our analysis reveals high intra-subject variability, low inter-subject distinguishability, and severe performance degradation as person count increases, indicating that commodity ESP32 sensors cannot provide sufficient signal quality for reliable multi-person separation.
Paper Structure (22 sections, 9 equations, 3 figures, 1 table)

This paper contains 22 sections, 9 equations, 3 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methodology
  3. Data Collection
  4. Preprocessing
  5. Person Count Estimation
  6. Signal Separation Methods
  7. FastICA
  8. SOBI
  9. PCA
  10. NMF
  11. Wavelet Transform
  12. Tensor Decomposition
  13. Feature Extraction
  14. Classification
  15. Experimental Results
  16. ...and 7 more sections

Figures (3)

  • Figure 1: System pipeline overview showing: (1) CSI data collection using ESP32 sensors, (2) preprocessing with filtering and normalization, (3) person count estimation via eigenvalue decomposition, (4) signal separation using six methods (FastICA, SOBI, PCA, NMF, Wavelet, Tensor), (5) feature extraction (24 features), (6) SVM classification.
  • Figure 2: Lab-Classroom performance gap showing environmental robustness across methods.
  • Figure 3: Failure mode distribution showing misclassification dominates across all methods.