Application of IoT-enabled computing technology for designing sports technical action characteristic model
Wenjuan Li
DOI: https://doi.org/10.1007/s00500-023-08966-4
IF: 3.732
2023-07-17
Soft Computing
Abstract:Internet of Things (IoT) is frequently employed to improve human lives by monitoring various physical and mental activities. IoT may significantly improve the performance and advancement of athletes by collecting, analyzing, and tracking consumer data. Training and teaching are widely used in sports to exploit the benefits of various technologies such as image processing, computer vision, big data, and advanced artificial intelligence. In sports, the traditional technical movement analysis using video and high-speed photography is used to standardize action, which is essential for improving performance of the athletes. This paper uses IoT technology to build a feature model for technical actions in sports so that the manifestation of information can be more diversified and then collate, store, transmit, and process the multimedia data. It uses image processing technology to quantitatively analyze sports technical action to build an action model. It combines the forward kinematics of the human body based on the root bone using deep learning-enabled human bone recognition algorithm. Next, it uses the obtained information to construct the structural model of human body, and athletes’ sports data are fed into it to complete the quantitative processing of action characteristics and pattern analysis. Finally, the application impact of the analysis model for striking force by the legs on sports athletes reveals that the maximum strike force of athletes is 486 N after reaching the target for 13 s, and then, it starts to fall. In sports training, the accuracy of each leg joint varies according to athletes’ speed, wherein the contribution of the hip was 20.5%, the knee was 20.5%, and the ankle was 71.1%, respectively. As a result, the method presented in this study for forward kinematics of the human body provides a robust and fast technique for determining the position and attitude of rigid body chains. This approach eliminates repetitive computations and increases computing performance by focusing on the trunk bone first, recursive calculations, and transformation matrices. These developments lead to this paper’s higher performance when compared to earlier research by making it a noteworthy addition to human body kinematics.
computer science, artificial intelligence, interdisciplinary applications