A Data-Driven Framework for Postural Assessment in Combat Skill Instruction Using Embedded Inertial Micro-Sensing

In the domain of martial arts pedagogy and athletic conditioning, accurately capturing biomechanical movement patterns is vital for individualized feedback and injury prevention. With the evolution of micro-electromechanical systems (MEMS), wearable sensor arrays have become increasingly compact, ef...

Full description

Saved in:
Bibliographic Details
Published in:2025 5th International Conference on Artificial Intelligence, Big Data and Algorithms (CAIBDA) pp. 840 - 843
Main Authors: Zeng, Qin, Liang, Jie, Song, Guodong
Format: Conference Proceeding
Language:English
Published: IEEE 20.06.2025
Subjects:
Biomechanics
Data Fusion Algorithms
Injuries
Injury Risk Prediction
Kinematics
Martial Arts Posture Analysis
MEMS Inertial Sensors
Microelectromechanical systems
Prediction algorithms
Predictive models
Safety
Tracking
Training
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In the domain of martial arts pedagogy and athletic conditioning, accurately capturing biomechanical movement patterns is vital for individualized feedback and injury prevention. With the evolution of micro-electromechanical systems (MEMS), wearable sensor arrays have become increasingly compact, efficient, and suitable for long-duration, unobtrusive motion tracking. In this study, a wearable lower-limb monitoring system is introduced, which integrates tri-axial inertial sensing modules comprising accelerometric, gyroscopic, and magnetometric units. To enhance the reliability of multi-modal kinematic signals, an adaptive fusion framework is constructed, incorporating quaternion-based orientation estimation and magnetic field calibration via ellipsoid compensation techniques. This system enables the extraction of joint kinematic parameters and supports the development of a comprehensive dataset tailored to martial arts biomechanics. Building upon this dataset, a regression-based predictive model is trained to quantify technical execution quality and estimate susceptibility to musculoskeletal strain. The resulting system facilitates decision-making in instructional contexts by offering interpretable, data-driven insights that inform skill refinement and safety interventions.
DOI:10.1109/CAIBDA65784.2025.11183104