EmoElites – Emotion meets quality diversity in game design

The digital gaming world, is constantly pushing its boundaries, striving to create increasingly realistic games with each new generation. This FYP, EmoElites, focuses on combining emotions and game design methods to explore how the environment and its elements impact players’ emotions during gameplay.

Taking into account human psychology, key factors in the ambiance of the virtual environment that affect the psyche of the individual were identified. These include  elements such as curvature, illumination, and the balance between obstacles and open space. The objective of this project was to allow users to experience various combinations of these features, with the aim of observing the emotions evoked by different combinations. The users were presented with procedurally generated levels, each distinct from the previous one. Moreover, users had the ability to make real-time adjustments to features such as illumination and curvature, influencing the intensity and bentness of the levels based on their chosen values.

To achieve these goals, the project used neuroevolution, a combination of neural networks and evolutionary computation. By leveraging the learning capabilities of neural networks and the search abilities of evolutionary computation, the project aimed to discover the optimal solution. As the neural networks evolved and were fine-tuned for improved performance and user preference, they generated the most suitable levels based on the collected data. Procedural generation was used to make this data visible to the users by generating levels with specific features, eliminating the need for manual creation of the entire level and its contents.

Figure 1. Visual demonstration of neuroevolution

Figure 2. Diagram outlining how all the different components work together within the system

Student: Dylan Zerafa
Supervisor: Dr Vanessa Camilleri