.png){kind=icon}
.png){kind=icon}
Return
Strategic Persona for Social Robots: Modeling Behavioral Tendencies in the Assertiveness-Cooperativeness Space
.png){kind=icon}
PRE
AI
J
S
DOI：10.1007/s12369-026-01411-0.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
As human-robot interaction becomes more social and emotional, an agent’s ability to perform the social functions of emotion is increasingly important. Existing emotion-based agents, however, are still constrained by rigid links between internal state and outward behavior, making it difficult to model the human ability to strategically adopt an appropriate persona in response to social context and an interaction partner. To address this limitation, we propose an emotion-based decision-making framework that integrally considers the emotional states of both the agent and the user. The framework is theoretically motivated by psychological work on emotion, behavioral tendencies, and social signaling, and it maps the respective emotional states of the agent and the user from a three-dimensional Pleasure-Arousal-Dominance space into a two-dimensional Assertiveness-Cooperativeness strategic behavior space. On the basis of the resulting strategic position, the agent dynamically selects one of five behavioral tendencies—competing, collaborating, avoiding, accommodating, or compromising—to generate contextually appropriate actions. Rather than remaining limited to a one-to-one correspondence between internal emotion and external behavior, the framework introduces a relational coordination layer in which the agent’s own affective state and its inferred estimate of the partner’s affective state jointly determine behavior. We evaluated the framework through a controlled video-based HRI user study. Compared with a conventional self-emotion-based baseline, the proposed framework yielded more favorable user ratings on several dimensions, particularly perceived social appropriateness and emotional understanding. The framework therefore contributes a structured basis for designing socially intelligent agents that can flexibly adopt a context-aware strategic persona.
Keywords:
Affective computing
Social intelligence
Decision-making
Dual concern theory
Social robotics
Persona
Journal
IF:
3.7
Papers:
1.4K
Citations:
5.6K
Organization
Cited Papers