Gait characteristic analysis and identification based on the iPhone's accelerometer and gyrometer

Area of Science:

• Biometrics
• Human-Computer Interaction
• Signal Processing

Background:

• Gait identification offers a non-intrusive method for remote human recognition.
• Existing methods may require specialized equipment or controlled environments.

Purpose of the Study:

• To propose and validate a novel gait identification approach using readily available smartphone sensors.
• To extract and analyze key gait characteristic parameters for improved identification accuracy.

Main Methods:

• Collected gait datasets using the triaxial accelerometer and gyrometer embedded in an iPhone.
• Processed sensor data to extract parameters: gait frequency, symmetry, dynamic range, and similarity coefficients.
• Developed a weighted voting scheme based on extracted gait parameters for identification.

Main Results:

• Validated the proposed gait identification scheme through four experiments with 40 datasets from 10 subjects.
• Demonstrated the effectiveness of using smartphone sensor data for gait characteristic analysis.
• The weighted voting scheme showed reliable performance in distinguishing individuals based on gait.

Conclusions:

• Smartphone-based gait analysis is a feasible and promising method for human identification.
• The proposed approach offers a practical solution for remote and non-intrusive individual recognition.
• Further research can explore advanced feature extraction and machine learning models for enhanced accuracy.