Background: Collaborative robots have emerged as a solution for complex and precision-driven tasks. However, there is no comprehensive risk assessment model for collaborative robotic work in the Republic of Korea. Thus, we aim to develop a risk assessment model tailored to the unique characteristics of human-robot collaboration. Methods: In this study, we examine the risk assessment factors in three key fields: worker, robotic systems, and work environment. The items adopted in the existing models of HFACS, USUS, and human–robot trust were initially selected to extract items for the risk assessment model. The analytic hierarchy process methodology was applied to refine and prioritize the evaluation items. Score scales were constructed for the six checklists. A five-level score ranging from one to five was given to each question in the checklist, and the average score was used to quantitatively evaluate each area. Results: The six evaluation checklists coalesced to constitute a comprehensive risk assessment model was adjusted for collaborative robot operations. Implementing the proposed model in six robot-operated workplaces yielded consistent results. Companies involved in previous robot-related accidents exhibited deficiencies in these risks than accident-free robot workplaces. Thus, comprehensive risk assessments encompassing the necessary factors are crucial to prevent accidents in robot-related work environments. Conclusion The proposed risk assessment model can offer a robust foundation for guiding the future development of diverse risk assessment models for collaborative robot environments and facilitate the safe coexistence of humans and robots in the Industry 4.0 era.

Background: With the increasing demand for industrial robots and the “noncontact” trend, it is an appropriate point in time to examine whether risk assessments conducted for robot operations are performed effectively to identify and eliminate the risks of injury or harm to operators. This study discusses why robot accidents resulting in harm to operators occur repetitively despite implementing control measures and proposes corrective actions for risk assessments. Methods: This study collected 369 operator-injured robot accidents in Korea over the last decade and reconstructed them into the mechanism of injury, work being undertaken, and bodily location of the injury. Then, through the techniques of Systematic Cause Analysis Technique (SCAT) and Root Cause Analysis (RCA), this study analyzed the root and direct causes of robot accidents that had occurred. Causes identified included physical hazards and complex combinations of hazards, such as psychological, organizational, and systematic errors. The requirements of risk assessments regarding robot operations were examined, and three case studies of robot-involved tasks were investigated. The three assessments presented were: camera module processing, electrical discharge machining, and a panel-flipping robot installation. Results: After conducting RCA and comparing the three assessments, it was found that two-thirds of injury-occurring from robot accidents, causative factors included psychological and personal traits of robot operators. However, there were no evaluations of the identifications of personal aspects in the three assessment cases. Conclusion Therefore, it was concluded that personal factors of operators, which had been overlooked in risk assessments so far, need to be included in future risk assessments on robot operations.