Train4U - Mobile Sport Diagnostic Expert System for User-Adaptive Training

As training plans of athletes are static and are not adapted dynamically towards current conditions (e.g. lack of sleep, hunger, weather, and stress), this contribution depicts an expert system to close this gap. Based on an extensive literature research, relevant parameters that influence sport performance, and explicit effects in the form of rules are detected. An empirical study that we conducted with more than 100 athletes revealed significance levels of certain parameters and further rules. Taking into account this data, a prototype Android application using a rule-based expert system is presented. While some parameters such as weather and location are determined automatically, the remaining parameters are collected by self-assessment that users have to do before and after the workout. All data (also including sensor data during workout) is stored in a database. In future, machine learning technologies will be included in order to derive new rules based on user statistics.