Developing a web-based full body exertion game in Godot using ML-based skeletal tracking

Sammanfattning: With the modernization of our society it has become more common to live a sedentary lifestyle. Nowadays a large percentage of people are required to sit for prolonged periods of time during office hours. This thesis presents the development and evaluation of an two-dimensional platformer exercise game, called Cave Copt, which was developed for Liopep. Liopep is an offshoot research program at Linköping University that aims to reduce periods of sedentary work through the use of gamification concepts. The objective of the game is to gather resources by navigating a helicopter through a large cave system. The helicopter can be controlled by using the player's physical movement as input for the game. This is done by using a machine learning algorithm called Pose, that is based on Google's MediaPipe framework, which can provide human pose tracking data. In order to answer the research questions, the game was played 100 times by the developerwhile relevant data was saved. Results show a slight upwards trend of player movement with each session played. The results also show that certain game mechanics, including level design and scoring systems, can increase the amount of motion experienced by the player.