Estimating Rotational Acceleration in Shoulder and Elbow Joints Using a Transformer Algorithm and a Fusion of...

Area of Science:

• Biomechanics
• Signal Processing
• Machine Learning

Background:

• Accurate estimation of joint rotational acceleration is crucial for understanding upper-limb biomechanics.
• Existing methods for mechanomyography (MMG) signal extraction and acceleration prediction have limitations.

Purpose of the Study:

• To develop and evaluate a novel framework for estimating rotational acceleration in elbow and shoulder joints.
• To compare the performance of a variational mode decomposition (VMD) algorithm against empirical mode decomposition (EMD) for MMG signal isolation.
• To assess the efficacy of transformer models and signal fusion in improving acceleration estimation accuracy.

Main Methods:

• Proposed a mechanomyography (MMG) signal isolation technique using variational mode decomposition (VMD).
• Employed transformer models for estimating rotational acceleration.
• Investigated the impact of fusing multiple biosignals on estimation performance.

Main Results:

• The VMD algorithm demonstrated superior performance in MMG signal extraction compared to EMD.
• Transformer models provided more precise joint acceleration estimates (average R² ≈ 0.96) than traditional time series models.
• Signal fusion enhanced estimation performance, yielding average R² increases of 0.041-0.053 over MMG alone.

Conclusions:

• The combination of VMD for signal isolation, transformer models for prediction, and signal fusion offers a robust framework for rotational acceleration estimation.
• This framework shows significant potential for applications in upper-limb biomechanical analysis.
• Further research is recommended to explore its applicability in diverse musculoskeletal contexts.