Interaction strategy for robotic apology based on human orientation toward service

In a human–robot symbiotic society, it is crucial to have seamless communication between human and robot. However, there are still numerous obstacles to overcome to develop a interaction system that can flexibly interact in various situations. A practical approach to implementing interaction systems socially is to optimize interaction strategies by limiting and selecting tasks and situations. This research focused on the behavior of 'apology,' which is unique to humans and expected to be challenging for robots to replace. Effective interaction strategies were examined, with emphasis on the user's orientation toward the service and the essential elements for a robot to apologize. An apology process was implemented based on specific situations, capable of responding to both utilitarian and relational orientations, along with appropriate multimodal behavior control and speech expressions that consider the user's relationship with the robot. The system with these elements was evaluated for validity and effectiveness in the apology process.