摘要：In this paper, we propose a research approach that is used in cognitive science to investigate more complex coordination mechanisms between three or more people. That is, we propose an approach that uses position data to quantitatively analyze group behavior and to link these characteristics with the cognition of anticipating others' behaviors. It is important for coordination to anticipate others' behaviors and to adapt one's own body movement to others based on anticipation. We highlight previous studies on problem solving and learning in cognitive science which have investigated interaction processes from verbal protocols during task implementation and indicated the importance of understanding others' perspectives. Additionally, recent cognitive models of estimating others' intentions and anticipating others' behaviors during interactions using non-verbal information such as eye movement, posture, and gesture, have been investigated. Considering these previous studies, we focus on group behavior and propose to apply the new approach mentioned above to discuss a mechanism of more complex coordination. We also refer to some studies of biological group behaviors in biology, artificial life, and sports science, and demonstrate a potential issue that such papers did not focus on the cognition related to coordinative group behaviors. This paper illustrates an example of discussing interactions with others, to which the new approach is applied. Our previous study here analyzed children's group behavior during nursery activities using position data and linked these characteristics with the cognitive development of anticipating others' behaviors based on spontaneous sociality. However, it is difficult to investigate some details of group behavior due to the limitation of field measurement, for example, the accuracy of a child's anticipation and whether a child moved based upon anticipation. In future work, it is important to analyze controlled group behavior and to indicate accuracy of individuals' anticipation from movement data to solve these problems.
