Multi-resident identification using device-free IR and RF fingerprinting.

Area of Science:

• Engineering
• Computer Science
• Gerontology

Background:

• Remote health and mobility monitoring is crucial for seniors aging in place.
• Passively monitoring individuals in multi-resident homes presents significant challenges.
• Existing methods often require wearable devices, limiting passive application.

Purpose of the Study:

• To develop a novel, passive, device-free method for individual identification in homes.
• To leverage radio frequency (RF) transceivers and infrared (IR) sensors for this identification.
• To assess the accuracy and robustness of the proposed identification technique.

Main Methods:

• Utilized wall-mounted RF transceivers and IR sensors for passive sensing.
• Employed radio frequency fingerprinting techniques based on received signal strength (RSS) disruptions.
• Integrated IR sensors for timing information and employed Gaussian Mixture Models for classification.

Main Results:

• Achieved over 98% classification accuracy in distinguishing between female and male subjects.
• Attained over 83% classification accuracy in distinguishing between two male subjects.
• Demonstrated high accuracy (96% and 82% respectively) even with significantly reduced training data.

Conclusions:

• The proposed RF and IR sensor-based system offers a highly accurate, device-free solution for individual identification.
• This technology has strong potential for enhancing remote monitoring and safety for seniors aging in place.
• The method proves effective even with limited training data, suggesting practical applicability.