Client-Server Approach for Managing Visual Attention, Integrated in a Cognitive Architecture for a Social Robot

Area of Science:

• Robotics
• Artificial Intelligence
• Cognitive Architectures

Background:

• Managing visual attention is crucial for social robot interaction and task performance.
• Existing visual attention mechanisms often lack seamless integration with higher-level cognitive processes.

Purpose of the Study:

• To propose and validate a novel client/server system for managing visual attention in social robots.
• To integrate this system with a cognitive architecture using a distributed knowledge graph.
• To quantitatively compare the proposed system against traditional visual attention mechanisms.

Main Methods:

• A client/server architecture was developed for visual attention management.
• A distributed knowledge graph was employed as a common knowledge representation for perceptual needs.
• The system was implemented on ROS and tested on a social robot.
• Performance was evaluated in RoboCup @ Home and SciROc robotic competitions.

Main Results:

• The proposed system demonstrated effective management of visual attention in social robots.
• Quantitative comparisons showed advantages over traditional visual attention mechanisms.
• Successful application in challenging robotic competition environments was achieved.

Conclusions:

• The novel visual attention system offers a robust and integrated approach for social robots.
• The client/server architecture and knowledge graph facilitate enhanced robot perception and cognitive control.
• The system's performance in competitions validates its practical applicability and superiority over conventional methods.