IoT Motion Tracking System for Workout Performance Evaluation: A Case Study on Dumbbell
Shilong Sun, Tengyi Peng, Haodong Huang, Yufan Wang, Xiao Zhang, Yu Zhou
Abstract
An intelligent sports training system based on Internet of Things (IoT) technology is proposed to build a low-cost, easy-to-use home exercise guidance solution, which can provide reliable exercise guidance when gymnasiums are inaccessible for users. The proposed intelligent system includes an inertial measurement microelectromechanical system with Bluetooth low-energy data transmission technology, a smart dumbbell with an acceleration sensor, an application on the smartphone terminal, and a computing central server in the clouds. Two-loop Kalman filters, dynamic motion segmentation method, and neural network are developed to demonstrate and evaluate the user’s dumbbell exercise modes. Six dumbbell exercise postures and 10 exercise cycles for eight participants are collected for system validation in the experimental study. The experimental results demonstrate that the proposed system can effectively and accurately segment multiple types of dumbbell movements (98.9% accuracy), recognize movements with high reliability (98.3% accuracy), and distinguish standard and non-standard movements (89% accuracy). Finally, this system with an intelligent algorithm software and hardware can be expanded to other similar types of sporting excises.