Research on the Human Motion Recognition Method Based on Wearable.

Area of Science:

• Biomedical Engineering
• Human-Computer Interaction
• Wearable Technology

Background:

• Accurate human dynamic behavior analysis is crucial for addressing limitations in movement diversity and adaptability.
• Wearable devices offer a promising avenue for real-time human behavior monitoring.

Purpose of the Study:

• To propose and validate a wearable device-based method for recognizing human dynamic behaviors.
• To assess the accuracy and efficiency of the proposed method in distinguishing between common human movements.

Main Methods:

• Utilized a six-axis sensor to collect acceleration and angular velocity data from wearable devices.
• Employed a human movement data acquisition platform, DMP attitude solution algorithm, and threshold algorithm for data processing.
• Collected movement data (standing, walking, jumping) from ten volunteers wearing sensors on multiple body parts.

Main Results:

• Achieved high recognition accuracies: 98.33% for standing, 96.67% for walking, and 94.60% for jumping.
• Attained an overall average recognition rate of 96.53%.
• Demonstrated improved accuracy, simplified algorithms, and efficient resource utilization compared to similar methods.

Conclusions:

• The proposed wearable sensor-based method effectively recognizes human dynamic behaviors with high accuracy.
• This approach offers a novel perspective for human dynamic behavior recognition, applicable to daily living assistance and health management.
• The method's efficiency in computation and improved accuracy promote wider adoption of wearable technology.