In college sports, health monitoring and real-time early warning are crucial for the prevention of sports injuries and sudden illnesses, but traditional monitoring means are limited by real-time and multi-dimensional analysis capabilities. To this end, this study designs a real-time health warning model based on wearable devices, which realizes dynamic parsing of physiological timing signals and multimodal data synergistic sensing to accurately identify sports risks by integrating a flexible sensor network with a hybrid long short-term memory-graph convolutional network (LSTM-GNN) algorithm. The model adopts a lightweight edge computing architecture to reduce response latency and combines with migration learning to enhance cross-scene adaptability. Experiments show that its warning accuracy reaches 95.2%, which can be adapted to real-time monitoring and personalized sports guidance in sports classrooms, providing innovative theoretical and technological support for the safety and security of college students’ sports health and intelligent application of wearable devices.
ZhongCLouWWangC. Neural network-based modelling for risk evaluation and early warning for large-scale sports events. Mathematics2022; 10: 3228.
2.
MahatoKSahaTDingS, et al.Hybrid multimodal wearable sensors for comprehensive health monitoring. Nat Electron2024; 7: 735–750.
3.
ReisFJAlaitiRKVallioCS, et al.Artificial intelligence and machine-learning approaches in sports: concepts, applications, challenges, and future perspectives. Braz J Phys Ther2024; 2024: 101083.
4.
WundersitzDWJosmanCGuptaR, et al.Classification of team sport activities using a single wearable tracking device. J Biomech2015; 48: 3975–3981.
5.
WilliamsonB. Algorithmic skin: health-tracking technologies, personal analytics and the biopedagogies of digitized health and physical education. Sport Educ Soc2015; 20: 133–151.
6.
LindbergRSeoJLaineTH. Enhancing physical education with exergames and wearable technology. IEEE Trans Learn Technol2016; 9: 328–341.
7.
ChangC-YLaiC-LHwangG-J. Trends and research issues of mobile learning studies in nursing education: a review of academic publications from 1971 to 2016. Comput Educ2018; 116: 28–48.
8.
StradzeAEPushkinaVNFedorovaEY, et al.Using wearable devices to stimulate students motor of physical activity and consequence physcological responce. PyR2020; 8: 121022.
9.
AhmedMDLeungWCW. Using wearable devices to enhance quality of physical education for students. Strategies2021; 34: 54–56.
10.
LiangFShanZWangD. Effect monitoring of physical education curriculum based on wearable sensor in intelligent environment. J Phys: Conf Ser2021; 2: 022001.
11.
AlmusawiHADurugboCMBugawaAM. Innovation in physical education: teachers’ perspectives on readiness for wearable technology integration. Comput Educ2021; 167: 104185.
12.
HsuY-LYangS-CChangH-C, et al.Human daily and sport activity recognition using a wearable inertial sensor network. IEEE Access2018; 6: 31715–31728.
13.
FernándezJBornnLCervoneD. A framework for the fine-grained evaluation of the instantaneous expected value of soccer possessions. Mach Learn2021; 110: 1389–1427.
14.
StrainTWijndaeleKDempseyPC, et al.Wearable-device-measured physical activity and future health risk. Nat Med2020; 26: 1385–1391.
15.
HaghiMThurowKStollR. Wearable devices in medical internet of things: scientific research and commercially available devices. Healthc Inform Res2017; 23: 4–15.
16.
NahavandiDAlizadehsaniRKhosraviA, et al.Application of artificial intelligence in wearable devices: opportunities and challenges. Comput Methods Progr Biomed2022; 213: 106541.
17.
KreissSBertoniLAlahiA. Openpifpaf: composite fields for semantic keypoint detection and spatio-temporal association. IEEE Trans Intell Transport Syst2021; 23: 13498–13511.
18.
LambrichtNEnglebertANguyenAP, et al.Impact of running clothes on accuracy of smartphone-based 2D joint kinematic assessment during treadmill running using OpenPifPaf. Sensors2025; 25: 934.
19.
QuanHBonariniA. Pyhapt: a python-based human activity pose tracking data processing framework. Software Impacts2022; 13: 100305.
20.
LambrichtNEnglebertAPitanceL, et al.Quantifying performance and joint kinematics in functional tasks crucial for anterior cruciate ligament rehabilitation using smartphone video and pose detection. Knee2025; 52: 171–178.
21.
YuYSiXHuC, et al.A review of recurrent neural networks: LSTM cells and network architectures. Neural Comput2019; 31: 1235–1270.
22.
SherstinskyA. Fundamentals of recurrent neural network (RNN) and long short-term memory (LSTM) network. Phys Nonlinear Phenom2020; 404: 132306.
23.
Al-SelwiSMHassanMFAbdulkadirSJ, et al.RNN-LSTM: from applications to modeling techniques and beyond—systematic review. J King Saud Univ Comput Inform Sci2024; 2024: 102068.
24.
ZhangSTongHXuJ, et al.Graph convolutional networks: a comprehensive review. Comput Soc Netw2019; 6: 1–23.
25.
BhattiUATangHWuG, et al.Deep learning with graph convolutional networks: an overview and latest applications in computational intelligence. Int J Intell Syst2023; 2023: 8342104.
26.
HongDGaoLYaoJ, et al.Graph convolutional networks for hyperspectral image classification. IEEE Trans Geosci Rem Sens2020; 59: 5966–5978.
27.
LiuCLiXLiQ, et al.Robot recognizing humans intention and interacting with humans based on a multi-task model combining ST-LSTM-GCNs model and YOLO model. Neurocomputing2021; 430: 174–184.
28.
WangXWangXYinX, et al.Distributed LSTM-GCNs-based spatial–temporal indoor temperature prediction in multizone buildings. IEEE Trans Ind Inf2023; 20: 482–491.
29.
YangLMiaoZLiT, et al.LSTM-GCNs based multidimensional parameter relationship analysis and prediction framework for system level experimental bench. Ann Nucl Energy2025; 210: 110890.
30.
WangYLiuJQianG. Hierarchical FFT-LSTM-GCNs based model for nuclear power plant fault diagnosis considering spatio-temporal features fusion. Prog Nucl Energy2024; 171: 105178.
31.
HanXGongS. LST-GCNs: long Short-Term Memory embedded graph convolution network for traffic flow forecasting. Electronics2022; 11: 2230.
32.
ShiYWangYQuY, et al.Integrated LSTM-GCNs stock prices movement prediction based on knowledge-incorporated graphs construction. Int J Mach Learn Cybern2024; 15: 161–176.
