Incorporating Commonsense Knowledge to Enhance Robot Perception

Robots have been significantly employed across various avenues in recent years. However, their application has been largely limited to controlled environments where variables are few and predictable. To address this challenge, we propose Robo-CSK-Organizer, a novel system that enhances robotic perception by integrating commonsense knowledge (CSK) for improved object organization, classification, and decision-making. By combining ConceptNet for semantic reasoning, DETIC for object identification, and BLIP for contextual analysis, Robo-CSK-Organizer achieves superior ambiguity resolution, task adaptation, and explainability compared to models without CSK. Testing in a real-world robotics setting demonstrates notable gains in transparency, user trust, and error handling, making this approach valuable for advancing AI transparency and the development of versatile robotic applications in automation and engineering. Future directions of this work are also comprehensively discussed. Note to Practitioners - This paper introduces the Robo-CSK-Organizer, a system designed to enhance robotic decision-making through the integration of commonsense knowledge (CSK). Developed with the goal of improving both the efficiency and transparency of robots in task execution, this work addresses the critical need for advanced object organization and classification capabilities in multipurpose robotics, especially in domestic household settings. By leveraging a classical commonsense knowledge base alongside cutting-edge object detection and context discernment technologies, Robo-CSK-Organizer demonstrates significant improvements in ambiguity resolution, placement consistency, and explainable AI (XAI), highlighted by its superior performance over a baseline model using ChatGPT for comparison. Practitioners engaged in design and implementation of robotic systems will find Robo-CSK-Organizer truly valuable for its adequate use of commonsense reasoning (inherent in humans) due to which it is adaptable to numerous real-world applications.