Robot action planning by commonsense knowledge in human-robot collaborative tasks

Robotics and artificial intelligence (AI) span the broad realm of mechatronics in general. Humans and robots can collaborate with each other to enhance various tasks, e.g. in the vehicle industry. Facets of AI such as commonsense knowledge can play a significant role here. In this paper, we propose an approach for human-robot collaboration such that it leverages commonsense knowledge to develop models for the simulation of assembly tasks in real-world applications. We consider modeling based on relevant attributes, e.g. weight and stability of parts. The proposed approach thereby entails human-robot interaction, mathematical modeling, both semantics and pragmatics in commonsense knowledge, as well as challenges specific to the application domain. We describe our approach, focusing on the mathematical modeling, and conduct our experimental evaluation using simulation tasks. Experimental results indicate that the proposed approach yields better outcomes than humans or robots working alone, which is in line with other such claims in the field of human-robot collaboration. This work sets the stage for real robot applications based on the results of our simulation tasks.