Multisensor Smart Glove With Unsupervised Learning Model for Real-Time Wrist Motion Analysis in Golf Swing Biomechanics

Abstract

This study presents a novel approach for golf swing biomechanics through deep learning-assisted to enhance the sensing, processing, interpretation, and assessment of swing quality. The proposed methodology introduces a smart golf glove (SGG) system, incorporating a deep neural network and the Internet of Things (IoT) capabilities. The key contributions of this study include 1) a comprehensive design framework encompassing both hardware and software aspects of the SGG system, as well as 2) an algorithm of phase and events to segment swing signals as inputs for deep learning models. Moreover, 3) an unsupervised learning model (ULM) architecture is introduced to address the challenges of dataset limitation and effort for data labeling. Additionally, 4) an IoT-assisted SGG platform is proposed, enabling remote monitoring and management. Experimental results indicate that the trained ULM model achieved an average accuracy of 92.4%. The findings highlight the effectiveness of the proposed SGG system in accurately detecting abnormal motion from beginner players. The integration of artificial intelligence (AI) and IoT-based platforms with targeted videos-based self-coaching capabilities represents a significant advancement in golf swing biomechanics analysis.