Human gait identification using UWB radar micro-Doppler signature.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Radar Technology

Background:

• Human gait analysis is crucial for identification and health monitoring.
• Existing methods may lack non-intrusiveness or continuous monitoring capabilities.
• Radar-based sensing offers a promising non-contact approach for gait analysis.

Purpose of the Study:

• To introduce and evaluate impulse radio ultra-wideband (IR-UWB) radar for person identification based on human gait.
• To develop a robust signal processing framework for extracting gait features from radar spectrograms.
• To enhance the reliability of radar-based person identification through uncertainty quantification.

Main Methods:

• Development of a signal processing framework to extract features from micro-Doppler gait signatures.
• Utilization of Principal Component Analysis (PCA) for feature extraction.
• Application of a cross-entropy-based approach for uncertainty quantification in classification.
• Implementation of a Support Vector Machine (SVM) classifier for person identification.

Main Results:

• Achieved a high accuracy rate of 97.1% for person identification using IR-UWB radar and gait analysis.
• Demonstrated the ability to discern subtle variations in gait patterns for effective individual recognition.
• Validated the proposed algorithm's performance on a dataset of 14 individuals over multiple walking trials.

Conclusions:

• IR-UWB radar, coupled with advanced signal processing, provides an effective method for non-contact person identification via gait analysis.
• The developed algorithm, including uncertainty quantification, enhances the reliability of gait-based identification systems.
• This technology holds significant potential for integration into smart home and home care systems, improving safety and personalized assistance.