An ambient agent model for reading companion robot

Reading is essentially a problem-solving task. Based on what is read, like problem solving, it requires effort, planning, self-monitoring, strategy selection, and reflection. Also, as readers are trying to solve difficult problems, reading materials become more complex, thus demands more effort and challenges cognition. To address this issue, companion robots can be deployed to assist readers in solving difficult reading tasks by making reading process more enjoyable and meaningful. These robots require an ambient agent model, monitoring of a reader’s cognitive demand as it could consist of more complex tasks and dynamic interactions between human and environment. Current cognitive load models are not developed in a form to have reasoning qualities and not integrated into companion robots. Thus, this study has been conducted to develop an ambient agent model of cognitive load and reading performance to be integrated into a reading companion robot. The research activities were based on Design Science Research Process, Agent-Based Modelling, and Ambient Agent Framework. The proposed model was evaluated through a series of verification and validation approaches. The verification process includes equilibria evaluation and automated trace analysis approaches to ensure the model exhibits realistic behaviours and in accordance to related empirical data and literature. On the other hand, validation process that involved human experiment proved that a reading companion robot was able to reduce cognitive load during demanding reading tasks. Moreover, experiments results indicated that the integration of an ambient agent model into a reading companion robot enabled the robot to be perceived as a social, intelligent, useful, and motivational digital side-kick. The study contribution makes it feasible for new endeavours that aim at designing ambient applications based on human’s physical and cognitive process as an ambient agent model of cognitive load and reading performance was developed. Furthermore, it also helps in designing more realistic reading companion robots in the future.