Decision support in AmI sport environments

In this paper, we comparatively analyze a set of decision making methodologies that have been successfully applied to improve athletes' training in real-time. The decision-making processes are based on the environmental data as well as on the athletes' biometrics in a scenario with runners in cross country circuits. Data acquisition is performed by means of a WSN which measures temperature and circuits slopes, whereas heart rate is monitored for each athlete. In all cases, the goal was to select the best tracks in the circuit to control the heart rate of the athlete within a given interval. Two broad types of decision-making procedures have been studied: (i) optimal classifiers, and (ii) dynamic programming. Results show a notable performance increase of these methods over heuristics, as well as the importance of environmental sensor data.