Utilization of Micro-Doppler Radar to Classify Gait Patterns of Young and Elderly Adults: An Approach Using a Long...

Area of Science:

• Biomedical Engineering
• Gerontology
• Signal Processing

Background:

• Falls are a significant risk for elderly individuals, leading to injury and reduced mobility.
• Early detection of fall risk through gait analysis is crucial for preventative interventions.
• Existing gait analysis methods may lack the accuracy or real-time capabilities for daily monitoring.

Purpose of the Study:

• To develop and validate a highly accurate micro-Doppler radar (MDR)-based gait classification method for early fall risk detection in the elderly.
• To differentiate gait patterns between young and elderly adults using MDR data.
• To establish a foundation for a daily monitoring system for elderly fall prevention.

Main Methods:

• Utilized micro-Doppler radar (MDR) to capture leg motion during walking in 300 participants.
• Extracted time-series velocity data from MDR spectrograms (time-velocity distribution).
• Employed a long short-term memory (LSTM) recurrent neural network for gait classification.

Main Results:

• Achieved a classification accuracy of 94.9% in distinguishing between young and elderly gaits.
• The MDR-based LSTM method significantly outperformed previous velocity-parameter-based classification techniques.
• Demonstrated the effectiveness of MDR spectrogram analysis for detailed gait characterization.

Conclusions:

• The proposed MDR-based gait classification method offers high accuracy for differentiating elderly gait patterns.
• This technology shows promise for developing a non-invasive, daily monitoring system for early fall risk detection in older adults.
• Further research can refine the system for real-world clinical application and personalized fall prevention strategies.