WiPIHT: A WiFi-Based Position-Independent Passive Indoor Human Tracking System

Area of Science:

• Human-computer interaction
• Wireless sensing
• Indoor localization

Background:

• Traditional indoor localization methods like vision-based tracking, infrared, and acoustics face environmental limitations or require specialized equipment.
• Current WiFi-based human sensing methods often rely on WiFi fingerprints, demanding extensive training, fixed environments, or precise transmitter/receiver (TX/RX) positioning, leading to instability with position changes.

Purpose of the Study:

• To propose WiPIHT, a novel system for stable indoor human activity trajectory tracking and reconstruction using commercial WiFi devices.
• To overcome the dependency on known TX-RX positions in existing WiFi-based trajectory reconstruction methods.
• To enable passive, real-time, and accurate tracking without requiring users to carry devices or attach locators.

Main Methods:

• Utilizes an innovative Channel State Information (CSI) channel analysis method.
• Extracts location-independent real-time movement speed features by analyzing the autocorrelation function of CSI.
• Incorporates Fresnel zone and motion velocity direction decomposition for movement direction change patterns independent of TX-RX positions.

Main Results:

• WiPIHT accurately reconstructs human activity trajectory shapes without prior knowledge of TX or human initial positions.
• Demonstrates significant advantages in tracking accuracy, real-time performance, equipment simplicity, and cost compared to existing methods.
• Achieves stable sensing performance even when TX-RX positions are uncertain or change.

Conclusions:

• WiPIHT offers a robust and practical solution for indoor human activity trajectory reconstruction using readily available WiFi infrastructure.
• The system's ability to perform location-independent analysis makes it adaptable to dynamic environments.
• Presents a cost-effective and efficient alternative for passive, real-time human sensing and activity recognition.