Monitoring physical exercise is becoming an increasingly relevant topic, with new fitness tracking devices and applications being released very often. While these options are highly popular among fitness communities, they are either very limited (i.e.: just counting steps or workout time) or require a lot of effort and dedication to log your activities (i.e.: manually inserting exercise type, repetitions, duration).
This project aims to fix these issues and create new possibilities, by developing software for smartphones or wearable devices that will automatically track physical activities, detecting specific workouts, repetitions and even correctness (e.g. posture, speed, frequency, extension, etc.), becoming a digital personal trainer.
To achieve this goal, the software will depend on sensor fusion and machine learning algorithms, which will go through a short learning period to correctly identify specific exercises.
The versatility of this concept makes it attractive for very diverse scenarios, ranging from rehabilitation activities for elderly people to high performance exercises for professional athletes.
Author: Tiago Ferreira
Type: MSc thesis
Partner: Instituto Superior de Engenharia do Porto