Abstract
Abstract In this study, a digital human model was constructed using virtual reality (VR) technology to simulate sports training to reduce injuries caused by improper training. A distance likelihood-based probabilistic model was developed to capture and reconstruct sports training movement data by incorporating computer techniques to improve the quality of data capture. In addition, the study created an adaptive graph convolutional neural network to recognize the same movement under different viewpoints. Through experiments in taekwondo double flying kick training, it was demonstrated that the virtual simulation model could significantly improve the training effect, and the experimental group’s performance of high horizontal kick and double flying kick enhanced by 21.73% and 85.21%, respectively. Applying VR technology to sports training can not only avoid injuries but also significantly improve the training effect.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
