Intelligent Agent Transparency in Human-Agent Teaming for Multi-UxV Management

Area of Science:

• Human-computer interaction
• Human-robot interaction
• Human-agent teaming

Background:

• Investigated operator performance, trust, and workload in multirobot management.
• Participants acted as unmanned vehicle (UxV) operators, guided by an intelligent agent (IA).
• The IA provided plan recommendations for mission completion.

Purpose of the Study:

• To examine the impact of varying levels of agent transparency on operator performance, trust, and workload.
• To assess the effectiveness of intelligent agent recommendations in a multirobot management context.

Main Methods:

• A within-subjects design with three levels of agent transparency.
• Eight missions per block, with IA errors introduced.
• Collected data on operator performance, trust, workload, and usability.

Main Results:

• Operator performance, trust, and perceived usability significantly increased with higher transparency levels.
• No significant increase in subjective or objective workload was observed.
• Response times remained consistent across different transparency levels.

Conclusions:

• Transparency enhances operator performance and trust without imposing additional workload or response time costs.
• Findings support the implementation of intelligent agents in military settings for heterogeneous unmanned vehicle (UxV) team design.