Estimation of lower limb torque: a novel hybrid method based on continuous wavelet transform and deep learning...

Area of Science:

• Biomechanics
• Biomedical Engineering
• Machine Learning

Background:

• Biomechanical analysis of human lower limbs is crucial for movement assessment, injury prevention, and rehabilitation.
• Traditional methods like optical motion capture are expensive and complex.
• There is a need for accessible and user-friendly biomechanical analysis tools.

Purpose of the Study:

• To develop a cost-effective and user-friendly system for real-time lower limb joint torque estimation.
• To integrate inertial measurement units (IMUs) with a novel deep learning framework.
• To enhance the accuracy and efficiency of biomechanical analysis.

Main Methods:

• Utilized inertial measurement units (IMUs) for data acquisition.
• Employed a deep learning framework combining Continuous Wavelet Transform (CWT) for time-frequency analysis.
• Integrated Multi-Head Self-Attention (MHSA), Bidirectional Long Short-Term Memory (Bi-LSTM), and 1D Convolutional Residual Network (1D Conv ResNet).

Main Results:

• Achieved high accuracy in joint torque estimation with RMSE of 0.16 N·m/kg, R² of 0.91, and Pearson correlation of 0.95.
• Demonstrated superior computational efficiency compared to existing models.
• Attained a single-cycle inference time of 152.6 ms, enabling real-time application.

Conclusions:

• The proposed IMU-based deep learning framework offers an accurate, efficient, and cost-effective solution for lower limb joint torque estimation.
• This method facilitates portable biomechanical monitoring systems for diverse applications.
• The approach addresses limitations of traditional gait analysis techniques.