Safe and Explainable AI Techniques to Human-Robot Collaboration

Abstract

Human-robot collaboration (HRC) has become a revolutionary paradigm of industrial, healthcare, and assistive robotics. To work well, robots are not only expected to work efficiently but also operate in a safe environment with human beings and ensure that their activities are transparently explained. Safe AI will provide the robots with collision prevention, compliance with physical limitations, and workspace consideration, whereas the Explainable AI (XAI) will improve the interpretability, confidence, and human control of the collaborative space. In this article, the design, implementation and evaluation of the HRC systems has been investigated based on the use of safe and explainable AI. We deconstruct models of intent modeling, risk-conscious control, and decision description, examine experimental data in the fields of manufacturing, healthcare, and assistive robotics, and interpolate the results of efficiency, safety, and human trust. The findings show that safety aware algorithms are much more effective when used in conjunction with interpretable models and lead to much fewer errors, greater efficiency in completing the tasks, and human confidence in autonomous systems.