A Dynamic Fitness Game Content Generation System Based on Machine Learning

Abstract

The purpose of this study is to explore the applications and design considerations for incorporating machine learning systems into fitness games, and to further examine the feasibility of using artificial intelligence systems to enhance player engagement. In the experimental phase, we used three AI systems to drive the dynamic progress of the game and enrich the changing and challenging elements of the game content. Benefiting from the guidable and generative nature of machine learning AIs, the dynamic progress of the fitness game can be reasonably balanced under the influence of both human player and AI. While the game elements can be generated through the co-creation of human creativity and machine derivation. The experimental results show that artificial intelligence technology can bring a new solution space combined of virtual coach and empowering fitness game systems. Through the advantages of balanced competition and co-created game characters, an AI empowered fitness game can effectively enhance player engagement under reasonable cost.