Human Error Identification and Risk Prioritization in Human–Robot Collaboration in Manufacturing

Abstract

ABSTRACT Human error recognition and subsequent prioritization are the most important tasks in the human–robot reliability analysis. This study aims to address the issue of human error in human–robot collaboration (HRC) by developing a model for identifying and assessing risks. First, the key tasks performed by operators during HRC were identified using the hierarchical task analysis, and a cognitive model was built based on information processing theory. This model breaks down the collaboration process into stages and identifies potential human errors at each step. Next, failure modes and effects analysis and evidence reasoning were applied to quantify the risk levels of these errors. Finally, the risks associated with human errors were measured, ranked, and compared with existing studies, and recommendations were made. The findings showed that the leading causes of safety risks in HRC are fatigue, illegal operations, error operations, misjudgments, and misperception. The perception stage of the process was found to carry the highest risk level, which means operators are more likely to make errors during the perception stage than during decision or execution, largely due to factors such as fatigue, distraction, and misperception. These results provide important theoretical support for improving safety in HRC and offer practical suggestions for refining risk management strategies in HRC systems.