ULTRA-MoCap: A Multimodal IMU and sEMG Dataset for Upper Body Joint Kinematics Analysis

Area of Science:

• Biomechanics
• Wearable Sensing
• Human-Computer Interaction

Background:

• Multi-modal wearable sensors offer non-intrusive human upper limb movement prediction for rehabilitation and sports analysis.
• Existing datasets lack comprehensive multi-modal data (IMU, sEMG) and high-fidelity upper limb kinematics.
• Current datasets struggle to capture isolated single-joint elbow and coordinated multi-joint shoulder movements.

Purpose of the Study:

• Introduce the Upper Limb Tracking with Multi-modal Capture (ULTRA-MoCap) dataset.
• Address limitations in existing datasets for upper limb movement analysis.
• Provide a comprehensive resource for research in biomechanics and wearable sensing.

Main Methods:

• Integrated multi-modal sensing including Inertial Measurement Units (IMUs) and surface electromyography (sEMG).
• Collected data from IMUs on hand, wrist, and forearm, and sEMG/IMU sensors on biceps, triceps, and deltoid.
• Utilized a multi-degree-of-freedom musculoskeletal model for high-fidelity upper limb kinematics.

Main Results:

• Developed the ULTRA-MoCap dataset integrating IMU, sEMG, and detailed kinematics.
• Collected data from thirteen subjects performing five distinct upper limb movements at varying speeds.
• Established a robust data collection and processing pipeline for multi-modal sensor data.

Conclusions:

• The ULTRA-MoCap dataset advances research in human upper limb movement prediction.
• This dataset supports diverse applications including rehabilitation, sports analysis, and assistive device control.
• ULTRA-MoCap provides a valuable resource for biomechanics, wearable sensing, and motion analysis research.