Prediction of athlete movements using wearable sensors for sports person health monitoring application

Autor: Yantao Lou, Tian Gao, Mamoun Alazab, C.B. Sivaparthipan
Rok vydání: 2021
Předmět:
Zdroj: Journal of Intelligent & Fuzzy Systems. :1-12
ISSN: 1875-8967
1064-1246
DOI: 10.3233/jifs-219160
Popis: Improvement in the data gathering to track the practise environments of the sports performance. Among these, the Internet of Things (IoT) technology with smartphones is increasingly evolving to help people with their health problems. In the world of athletics, wearable devices can provide real-time data to track athletes’ heart rhythms and help athletic activities. The players’ pulse rates change at various positions as they play sport and track their heartbeat, allowing them to understand their fitness and improve a person’s health. Therefore, the study proposes a wearable sensor-based athletic movement prediction (WS-AMP) model. The model uses the deep learning algorithm to effectively classify motions usually extracted from the interactive motion panels and determine how feasible it is to perform wearable sensor data classification. On 523 athletes with nine athletic motions, data on optical motion capture have been obtained. The research performs the deep neural network model’s training and validation, incorporating the convolutional neural network. The experimental study performs the prediction analysis and comparison with existing machine learning models. The experimental above analysis of wearable sensor-based IoT health monitoring of Sport person movements prediction are Abnormal Conditions ratio is 86.65%, Spectrum analysis of heart rate ratio is 87.12%, the Error rate of body maintenance ratio is 83.51%, Mental acuity ratio is 87.10% and finally overall accuracy, and F1 score ratio is 93.80%.
Databáze: OpenAIRE
Externí odkaz: