Playing a 3D Tunnel Game Using Reinforcement Learning

Abstract

Tunnel games are a type of 3D video game in which the player moves through a tunnel and tries to avoid obstacles by rotating around the axis of the tunnel. These games often involve fast-paced gameplay and require quick reflexes and spatial awareness to navigate through the tunnel successfully. The aim of this thesis is to explore the representation of a tunnel game in a discrete manner and to compare various reinforcement learning algorithms in this context. The objective is to evaluate the performance of these algorithms in a game setting and identify potential strengths and limitations. The results of this study may offer insights on the application of discrete tabular methods in the development of AI agents for other continuous games.