TICK: A Knowledge Processing Infrastructure for Cognitive Trust in Human-Robot Interaction

Area of Science:

• Robotics
• Artificial Intelligence
• Cognitive Science

Background:

• Assessing trust in human-robot interaction (HRI) is crucial for social and assistive robotics.
• Existing systems lack a unified cognitive infrastructure for shared knowledge and trust evaluation.
• The need for domain-agnostic solutions that promote trust in HRI is growing.

Purpose of the Study:

• To introduce and evaluate the Trust-Inferring Infrastructure for Cognitive Knowledge (TICK) for HRI domains.
• To develop a domain-agnostic ontology for shared knowledge encompassing behavior, beliefs, intentions, and situational awareness.
• To enable participants to take actions that build trust with other parties in HRI.

Main Methods:

• Developed a layered, domain-agnostic ontology for shared knowledge.
• Introduced the TICK framework with three constructs: performance, process, and purpose.
• Integrated a multimodal sensory module for enhanced robot reasoning and validated through real-world scenarios and lab experiments.

Main Results:

• TICK enhances robot capabilities in understanding human intentions, situational awareness, and personalized behavior.
• Robot trustworthiness evolved based on performance, with trust automatically assessed via proximity and recommendation adherence.
• Experiments demonstrated the scalability and flexibility of TICK in a kitchen domain handover task.

Conclusions:

• The TICK framework provides a robust cognitive infrastructure for assessing and building trust in HRI.
• The multimodal sensory integration and layered ontology allow for adaptable and effective human-robot collaboration.
• The study validates TICK's potential for diverse HRI applications requiring trust and shared understanding.