INTELLIGENT SYSTEMS IN SPORTS TRAINING MANAGEMENT: AN ANALYSIS OF THE EFFECTIVENESS OF USING ARTIFICIAL INTELLIGENCE TECHNOLOGIES TO EVALUATE LOW START TECHNIQUE

Abstract

The article describes aspects of the implementation of an intelligent system for analyzing the technique of performing a low start in sprinting based on video processing.  The innovation of the work lies in the comparative analysis of two approaches to processing motor actions: a hybrid system based on biomechanical features and classical machine learning methods, and a neural network architecture with a bidirectional LSTM network. The economic significance of the study is determined by a significant increase in the efficiency of sports organizations. The implementation of the system makes it possible to optimize the time spent on analyzing the technique of performing exercises, which leads to an increase in the productivity of the coaching staff. Automating the process of evaluating the starting technique reduces the processing time of the video, which allows the coach to pay more attention to individual work with athletes. The methodological basis of the research includes the use of MediaPipe PoseLandmarker technology to extract key body points, the development of algorithms for converting an athlete's posture into measurable biomechanical parameters, as well as the creation of a neural network model to evaluate the technical elements of the start. The practical significance of the work is determined by the system of structured reports in Russian and JSON format for managers defined by the 1C:Enterprise system.