HEART RATE MONITORING SYSTEM

Heart rate is a key physiological indicator for assess ing cardiovascular load, training intensity, and recovery status in sportspersons. Conventional monitoring techniques such as manual pulse checks or basic fitness bands are inadequate during high-intensity sports because they lack continuous, accurate, and intelligent analysis capabilities. This report presents the design and implementation of an AI based heart rate monitoring system for sports training that combines wearable sensors, an ESP32 based embedded platform, and machine learning algorithms to analyse cardiac activity in real time. The system acquires physiological signals from ECG and PPG based heart rate sensors, together with auxiliary parameters such as temperature. The raw data is pre-processed using filtering, normalization, and segmentation before extracting time-domain, frequency-domain, and nonlinear features. Multiple supervised learning models—Decision Tree, Random Forest, Support Vector Machine (SVM), Logistic Regression, Na¨ıve Bayes, K-Nearest Neighbour (KNN), and Weighted KNN—are trained to classify heart rhythm patterns and risk levels. Experimental evaluation shows that Decision Tree and Weighted KNN achieve an accuracy of approximately 97.78%, while SVM reaches about 96.67%, demonstrating the suitability of AI for arrhythmia detection in sports scenarios. The proposed system categorizes heart rate into training zones, detects abnormal conditions such as tachycardia or bradycardia, and generates alerts when values deviate from safe thresholds. The overall solution is portable, low power, and scalable for multi sport applications, providing coaches and athletes with actionable insights for safe and optimized training sessions.