The purpose of this study is to provide a theoretical substantiation and systematization of innovative approaches to intelligent training process management based on the integration of artificial intelligence (AI) technologies and wearable sensor systems. The relevance of the work is driven by the necessity of transitioning from subjective, empirical methods of coaching control to data-driven sports management within the "in the wild" concept (monitoring in natural competitive environments). As a result of the research, a comparative analysis of three generations of data collection systems was conducted, demonstrating the advantage of soft epidermal sensors over traditional optical systems due to the minimization of soft tissue artifacts and the possibility of multimodal monitoring (combining biomechanics and sweat biochemistry). It has been established that deep learning algorithms (CNN, LSTM, and their hybrids) serve as the fundamental tool for transforming "raw" data into managerial decisions, as they are capable of recognizing micro-changes in an athlete's technique with a level of precision invisible to the human eye. A particular novelty of the study is the development of an analytical matrix of predictive markers that links cumulative ground reaction force and kinematic asymmetry to the proactive forecasting of non-contact injury risks. It was revealed that the "black box" problem remains a limiting factor for the implementation of AI in professional sports, requiring the development of Explainable Artificial Intelligence (XAI) to increase coaching staff's trust in automated recommendations. The practical significance lies in the possibility of implementing decision support systems (DSS), which allow for the transformation of sports organizations' operations, ensuring a balance between maximizing athletic performance and preserving the health of athletes as the club's key assets.
Glek et al. (Thu,) studied this question.