Real-Time Gym Activity Recognition: A Fog-Enabled IoT Solution with Cloud Integration

Fog and cloud computing, providing more complex environments for cognitive fitness analytics, have gained importance in the development of Internet of Things (IoT) applications. The present work introduces a cloud-fog system to recognize activities in gymnasium environments, using the ensemble learning approach based on the Random Forest classifier and XGBoost classifier with Voting classifiers for better accuracy. The system utilizes IoT sensors in a smart, integrated manner to safely and effectively call the information, thereby optimizing Cloud-Fog architecture for authenticating physical activity detection as fast as possible. Experimental results demonstrate the capabilities of the system for identifying 101 real-time physical activities within a virtual gym. The reduced response time of the proposed cloud-fog hybrid as compared to traditional cloud-only systems proves success in enabling IoT with fog for better health analytics. The study also sheds light on directions that can be explored for furthering research in the domain of intelligent fitness analytics by fusing cloud, fog as well as IoT edge devices.