Wearables and Machine Learning for Improving Runners' Motivation from an Affective Perspective

Area of Science:

• Sports Science
• Affective Computing
• Wearable Technology

Background:

• Wearable technology is increasingly used in sports for performance enhancement and injury prevention.
• Emotional states are crucial in athletic training but are often not monitored or utilized in real-time.
• Existing solutions lack real-time emotion monitoring for personalized athlete support.

Purpose of the Study:

• To develop a wearable system capable of real-time emotion recognition for runners.
• To integrate emotion detection into a system that enhances athletic training quality and safety.
• To leverage detected emotions for personalized motivation through music during training.

Main Methods:

• Utilized electrodermal activity (EDA) as a physiological measure for emotion detection.
• Developed machine learning models to analyze EDA data and infer runners' emotions.
• Integrated the wearable device and models into the DJ-Running mobile application for real-time feedback.

Main Results:

• Successfully developed a wearable system and machine learning models to deduce runners' emotions during training.
• Demonstrated the feasibility of using electrodermal activity for emotion recognition in a running context.
• Enabled real-time emotion-driven music selection to enhance runner motivation.

Conclusions:

• Real-time emotion monitoring via wearables can significantly improve athletic training.
• The DJ-Running system effectively uses physiological data to personalize the training experience.
• This approach offers a novel method for enhancing athlete motivation and well-being through affective computing.