Dynamic Adaptative Surveillance Training in a Virtual Environment Using Real-Time Cognitive Load and Performance

Abstract

Dynamic difficulty adjustment of serious games is a field of research seeking to assist participants attain an ideal learning state. To achieve this, the difficulty of a game is adjusted in real-time to approach the capabilities of the player. Many dynamic difficulty adjustment systems only measure a few variables, often solely player performance, and adjust a limited number of in-game aspects. Little research has sought to ascertain if measuring a combination of cognitive load with measures of performance in real-time leads to a more effective dynamic difficulty adjustment system. Building on research which defined ‘mental efficiency’, we conducted a novel experiment to address this gap. This is achieved by comparing two versions of a surveillance training serious game; one a linear approach and the other with our unique cognitive load and performance-based dynamic difficulty adjustment system. Our experiment included n=52 participants (26 per treatment group). The experiment demonstrates that our approach achieved similar performance outcomes with lower cognitive load, in less time than the linear difficulty approach. These results indicate that our system enhances learning capacity and may prove beneficial for future serious games.