A Study on the Association Between Tower Crane Operator Fatigue State and Collision Risk Under Human–Machine Interaction

To investigate the relationship between operator fatigue and collision risk under human–machine interaction (HMI) in intelligent tower crane operations, and to reveal the mitigating effects of HMI on fatigue-induced collision risks, a comprehensive data acquisition approach integrating eye-tracking signals, risk indicators, and fatigue scale assessments was proposed and validated through scenario-based experiments. First, two experimental scenarios—traditional mechanical operation and HMI operation—were established. Based on a review of existing studies, representative eye-movement metrics and fatigue scale indicators were selected. Subsequently, operator fatigue states were classified into three levels: low fatigue, moderate fatigue, and high fatigue. A total of 28 participants were recruited to complete fatigue assessments and subsequently perform tower crane lifting tasks under both experimental scenarios. Finally, collision risk under different scenarios was quantitatively evaluated using the safety distance between the crane hook and the rigger, as well as the frequency of collision alarms. The results indicate that, under traditional mechanical operation, increasing fatigue levels were associated with a significant reduction in safety distance between the crane hook and the rigger, accompanied by a marked increase in collision alarm occurrences, resulting in a relatively high overall collision risk. In contrast, under the HMI operation scenario, participants demonstrated superior operational control at equivalent fatigue levels. Specifically, under moderate fatigue, collision risk was reduced from low risk to no risk, while under high fatigue, collision risk decreased from high risk to low risk. These results indicate that, under laboratory-simulated conditions, human–machine interaction can mitigate, to a certain extent, the increasing trend of collision risk when operators perform tower crane lifting operations under fatigue. These findings provide a scientific basis for further optimization of intelligent tower crane operational modes and the development of enhanced safety management strategies.