Force-Based Modeling of a Resilient Helping Role in Coordinated Behavior of a Triad

The aim of this study was to understand complex and dynamic coordination, which refers to the nonverbal interaction of more than three members. This would be applied to smooth and effective role-sharing, as in multiplayer games with AI characters. Our previous study indicated the importance of resilient help using a coordinated drawing task involving a triad. In the experiment, each participant operated a reel to change the thread tension and moved a pen connected to the three threads to draw an equilateral triangle. Three heterogeneous roles were shared: the pulling and relaxing roles moving the pen as if pulling it closer to the hand and supporting its smooth movement, respectively. However, these roles alone cannot draw a triangle’s side because of the task specifications. The adjusting role must moderately intervene in the two roles to correct the pen trajectory, which can deviate. The multiple regression model using pen position and tension data revealed that the third role was related to the high task performance of the pen’s quick movement while minimizing deviation. To constructively understand this resilient helping role, this study conducted multi-agent simulation using equations of motion. Here, the tension in the adjusting role was changed as optimally as possible so that the resultant force of the three roles would coincide on a side. The results showed that the proposed model partially replicates the pen trajectory. Overall, this role might anticipate the forces of the other roles based on the performance information, that can be perceived.