Proposal of a Game with Dynamic Difficulty Adjustment from Physiological Signals in the Context of an Exergame

Abstract

A sedentary lifestyle is considered one of the main risk factors for mortality according to the World Health Organization, due to its association with diseases such as diabetes and hypertension. This work presents an exergame to promote physical activity (PA), and allows the storage of performance data and heart rate (HR) obtained from the user’s interaction with it. For application development, Visual Studio Code, Python programming language and Pygame library were used; this application recreates an invasion theme scenario in which the user must perform PA to save the world. Two data collection processes were carried out: 18 people with a mean age of 22.5 years participated in the first process, generating a data set of 90 records, which were used to generate the characterization of a heuristic algorithm and the training of an artificial neural network (ANN); Subsequently, once the automatic parameter adjustment algorithms were established in the game, the second data collection was carried out in order to generate data sets regarding HR progression and performance, in which 8 people with an average age of 25.3 years participated. The eight people were randomly divided into two groups, and interacted with the heuristic and ANN-based configuration game for three days a week for a month. Additionally, surveys were conducted to find out information about motivation throughout the process. The results indicate that the game with an algorithm based on ANN has a greater impact on the progress and motivation of the users.