Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AI-Enhanced Real-Time Wi-Fi Sensing Through Single Transceiver Pair

Yuxuan Liu, Chiya Zhang, Yifeng Yuan, Chunlong He, Weizheng Zhang, Gaojie Chen

TL;DR

This work addresses the challenge of realizing high-precision Wi-Fi sensing with minimal hardware by establishing a theoretical basis for AI-driven gains. It identifies prior information and temporal correlation as the two core sources of AI-enhanced performance under hardware constraints, and validates this through a real-time system built from a single transceiver pair. The authors implement an end-to-end CSI-based pipeline for simultaneous indoor localization and 3D human pose estimation, achieving ~0.61 m localization and ~0.22 m pose error at 42+ fps on commodity hardware, with temporal correlation providing consistent accuracy gains. The results underscore the practical potential of AI-enhanced Wi-Fi sensing for scalable deployments and real-time visualization in indoor environments, while highlighting the need for pre-training to maximize temporal benefits.

Abstract

The advancement of next-generation Wi-Fi technology heavily relies on sensing capabilities, which play a pivotal role in enabling sophisticated applications. In response to the growing demand for large-scale deployments, contemporary Wi-Fi sensing systems strive to achieve high-precision perception while maintaining minimal bandwidth consumption and antenna count requirements. Remarkably, various AI-driven perception technologies have demonstrated the ability to surpass the traditional resolution limitations imposed by radar theory. However, the theoretical underpinnings of this phenomenon have not been thoroughly investigated in existing research. In this study, we found that under hardware-constrained conditions, the performance gains brought by AI to Wi-Fi sensing systems primarily originate from two aspects: prior information and temporal correlation. Prior information enables the AI to generate plausible details based on vague input, while temporal correlation helps reduce the upper bound of sensing error. Building on these insights, we developed a real-time, AI-based Wi-Fi sensing and visualization system using a single transceiver pair, and designed experiments focusing on human pose estimation and indoor localization. The system operates in real time on commodity hardware, and experimental results confirm our theoretical findings.

AI-Enhanced Real-Time Wi-Fi Sensing Through Single Transceiver Pair

TL;DR

This work addresses the challenge of realizing high-precision Wi-Fi sensing with minimal hardware by establishing a theoretical basis for AI-driven gains. It identifies prior information and temporal correlation as the two core sources of AI-enhanced performance under hardware constraints, and validates this through a real-time system built from a single transceiver pair. The authors implement an end-to-end CSI-based pipeline for simultaneous indoor localization and 3D human pose estimation, achieving ~0.61 m localization and ~0.22 m pose error at 42+ fps on commodity hardware, with temporal correlation providing consistent accuracy gains. The results underscore the practical potential of AI-enhanced Wi-Fi sensing for scalable deployments and real-time visualization in indoor environments, while highlighting the need for pre-training to maximize temporal benefits.

Abstract

The advancement of next-generation Wi-Fi technology heavily relies on sensing capabilities, which play a pivotal role in enabling sophisticated applications. In response to the growing demand for large-scale deployments, contemporary Wi-Fi sensing systems strive to achieve high-precision perception while maintaining minimal bandwidth consumption and antenna count requirements. Remarkably, various AI-driven perception technologies have demonstrated the ability to surpass the traditional resolution limitations imposed by radar theory. However, the theoretical underpinnings of this phenomenon have not been thoroughly investigated in existing research. In this study, we found that under hardware-constrained conditions, the performance gains brought by AI to Wi-Fi sensing systems primarily originate from two aspects: prior information and temporal correlation. Prior information enables the AI to generate plausible details based on vague input, while temporal correlation helps reduce the upper bound of sensing error. Building on these insights, we developed a real-time, AI-based Wi-Fi sensing and visualization system using a single transceiver pair, and designed experiments focusing on human pose estimation and indoor localization. The system operates in real time on commodity hardware, and experimental results confirm our theoretical findings.

Paper Structure

This paper contains 18 sections, 48 equations, 10 figures, 1 table.

Table of Contents

  1. Introduction
  2. Theoretical Analysis of the Performance Gains
  3. Spatial Resolution Analysis Based on Radar Theory
  4. Performance Gain Brought by Prior Information
  5. Performance Gain Brought by Temporal Correlation
  6. AI-based Real-time Wi-Fi Sensing and Visualization System
  7. System Overview
  8. CSI Data
  9. CSI Preprocessing
  10. Network Design
  11. Loss Function
  12. Real-Time Data Processing Design
  13. Experimental Setup and Results
  14. CSI Collection
  15. 3D Human Pose Collection
  16. ...and 3 more sections

Figures (10)

  • Figure 1: Angle-of-arrival estimation in a typical Wi-Fi sensing system.
  • Figure 2: An example of the collected CSI. The area within the red box exhibits clear sinusoidal characteristics.
  • Figure 3: CSI based Wi-Fi sensing process.
  • Figure 4: Schematic diagram of the AI-based real-time Wi-Fi sensing system.
  • Figure 5: Illustration of the network structure, where we incorporate a hyperparameter in the stacked ConvGRU layers to control the maximum processable sequence length. The depth head and 2D pose estimation head work jointly to achieve 3D human pose estimation.
  • ...and 5 more figures