DEVELOPMENT OF AN AI-SUPPORTED TRAINING METHODOLOGY FOR TRIATHLETES

Аннотация

The purpose of this study was to develop a scientifically grounded methodology
for managing triathletes’ training in which artificial intelligence (AI) is used as a decision-support
tool for the coach. The study followed a design-oriented methodological approach. Scientific
literature on endurance monitoring, training load, heart rate variability, wearable technology, and
machine learning was analysed; data requirements were defined; an algorithm for daily readiness
assessment and weekly load adjustment was developed; and safety and pedagogical-control
criteria were formulated. The resulting methodology integrates swimming, cycling, and running
data with heart rate, heart rate variability, sleep, perceived exertion, and subjective well-being. A
closed loop is proposed: data acquisition, quality control, state assessment, recommendation,
coach decision, and feedback. AI classifies the athlete’s state, identifies atypical changes, and
proposes adjustments to volume, intensity, and recovery. Final authority, however, remains with
the coach and, where appropriate, a medical professional. The practical value of the methodology
lies in its potential use by sports schools, clubs, and individually coached athletes. A controlled
trial using real athlete data is required to determine its effectiveness.