学术文献库

Socially assistive robots provide physical and mental assistance for humans via cognitive human-machine interactions. These robots should sustain long-term engaging interactions with humans in a similar way humans interact with each other. According to the theory of mind, in their interactions humans develop cognitive models of each other in order to estimate their unobservable state-of-mind, predict their behavior, and act accordingly. Based on the theory of mind, we propose mathematical cognitive models of humans, which enable machines to understand cognitive procedures of humans in general and as distinct individuals. In particular, a network representation that is formulated based on a proposed extended version of fuzzy cognitive maps is introduced. The resulting models are identified and validated using (1) computer-based simulations designed according to a general data set of human's intuitive reasoning and literature and (2) real-life personalised experiments with 15 human participants. The results of the experiments show that the proposed cognitive models can excellently be personalised to different participants and precisely estimate and predict their current and future state-of-mind and expected behaviors.