Research Article
Artificial Intelligence Based Procedural Content Generation in Serious Games for Health: The Case of Childhood Obesity
@INPROCEEDINGS{10.1007/978-3-031-32029-3_19, author={Eleftherios Kalafatis and Konstantinos Mitsis and Konstantia Zarkogianni and Maria Athanasiou and Antonis Voutetakis and Nicolas Nicolaides and Evi Chatzidaki and Nektaria Polychronaki and Vassia Chioti and Panagiota Pervanidou and Konstantinos Perakis and Danae Antonopoulou and Efi Papachristou and Christina Kanaka-Gantenbein and Konstantina S. Nikita}, title={Artificial Intelligence Based Procedural Content Generation in Serious Games for Health: The Case of Childhood Obesity}, proceedings={Wireless Mobile Communication and Healthcare. 11th EAI International Conference, MobiHealth 2022, Virtual Event, November 30 -- December 2, 2022, Proceedings}, proceedings_a={MOBIHEALTH}, year={2023}, month={5}, keywords={serious game adaptive procedural content generation genetic algorithm health sensors childhood obesity}, doi={10.1007/978-3-031-32029-3_19} }- Eleftherios Kalafatis
Konstantinos Mitsis
Konstantia Zarkogianni
Maria Athanasiou
Antonis Voutetakis
Nicolas Nicolaides
Evi Chatzidaki
Nektaria Polychronaki
Vassia Chioti
Panagiota Pervanidou
Konstantinos Perakis
Danae Antonopoulou
Efi Papachristou
Christina Kanaka-Gantenbein
Konstantina S. Nikita
Year: 2023
Artificial Intelligence Based Procedural Content Generation in Serious Games for Health: The Case of Childhood Obesity
MOBIHEALTH
Springer
DOI: 10.1007/978-3-031-32029-3_19
Abstract
This paper presents a novel Procedural Content Generation (PCG) method aiming at achieving personalization and adaptation in serious games (SG) for health. The PCG method is based on a genetic algorithm (GA) and provides individualized content in the form of tailored messages and SG missions, taking into consideration data collected from health-related sensors and user interaction with the SG. The PCG method has been integrated into the ENDORSE platform, which harnesses the power of artificial intelligence (AI), m-health and gamification mechanisms, towards implementing a multicomponent (diet, physical activity, educational, behavioral) intervention for the management of childhood obesity. Within the use of the ENDORSE platform, a pre-pilot study has been conducted, involving the recruitment of 20 obese children that interacted with the platform for a period of twelve weeks. The obtained results, provide a preliminary justification of PCG’s effectiveness in terms of generating individualized content with sufficient relevance and usefulness. Additionally, a statistically significant correlation has been revealed between the content provided by the proposed PCG technique and lifestyle-related sensing data, highlighting the potential of the PCG’s capabilities in identifying and addressing the needs of a specific user.
