IoT-Enabled Adaptive Back-Posture Monitoring System Using Sensor Fusion and Personalized Thresholds

This work presents an IoT-enabled wearable back-posture monitoring system using an IMU sensor to detect sitting posture deviations in real time. Two prototypes were developed: an Arduino-based threshold detector and an ESP32 IoT-connected system with customizable alert thresholds and historical data logging. Validation with 80 participants demonstrated up to 90% detection accuracy with reduced false alerts. An adaptive machine-learning framework is proposed to personalize posture thresholds based on user behaviour. The system demonstrates a practical, low-cost approach to posture awareness and musculoskeletal health support.