Footstep localization and force estimation through structural vibrations using the FEEL Algorithm

Area of Science:

• Smart Environments
• Human-Computer Interaction
• Sensor Fusion

Background:

• Indoor localization is crucial for smart environments but faces challenges like privacy concerns and device requirements.
• Existing methods often rely on time-of-flight variations, necessitating complex infrastructure or wearable devices.
• Structural vibrations from footsteps offer a privacy-preserving, device-free alternative for indoor positioning.

Purpose of the Study:

• To present the Force Estimation and Event Localization (FEEL) algorithm for improved indoor person localization.
• To demonstrate FEEL's effectiveness using lower sampling rates and fewer sensors compared to time-of-flight methods.
• To enhance FEEL's force estimation and localization accuracy for practical applications.

Main Methods:

• Developed and improved the Force Estimation and Event Localization (FEEL) algorithm.
• Utilized structural vibrations generated by footsteps for localization.
• Analyzed 1100 footsteps to evaluate localization accuracy and estimate ground force reactions (GRF).

Main Results:

• Achieved 98.6% localization accuracy with the improved FEEL algorithm.
• FEEL demonstrated superior performance with lower sampling rates and fewer sensors.
• Accurately estimated participant body-weight-ratios (BWR) from ground force reactions.

Conclusions:

• The improved FEEL algorithm offers a highly accurate and practical solution for indoor person localization.
• FEEL effectively overcomes privacy and device-related challenges associated with other localization technologies.
• The algorithm's ability to estimate body-weight-ratios adds further utility for smart environment applications.