FitSense: A Closed IoT Ecosystem for Gym Environments with RFID Tracking and ML-Driven Diet and Exercise Recommendations

Authors: Rudramurthy V C, J Madhav Pai, Gaurang Bharadwaj, Somya Das

Abstract: Modern gym environments suffer from fragmented tracking ecosystems. Users typically rely on personal wearables to track workouts and separate mobile applications to log diets, while gym owners lack granular data on machine utilization. This paper introduces FitSense, a holistic closed IoT ecosystem designed to bridge this gap. By retrofitting standard gym machines with an ESP32-based IoT module featuring an RFID-RC522 reader and an MPU-9250/6500 inertial measurement unit (IMU), we enable seamless session tracking. Users tap their RFID cards to start and stop workouts on any machine. Instead ofcomplex motor instrumentation, the IMU calculates steps and speed based on step vibrations and the user's pre-configured stride length. This hardware ecosystem is tightly integrated with a centralized dietary and calorie tracker. A Machine Learning engine utilizes the combined history ofphysical exertion and nutritional intake to recommend personalized diet adjustments and future exercise routines. This creates a unified platform where fitness and nutrition are tracked, analyzed, and optimized in a single ecosystem.