How do communication errors between crane operators and signalers occur? An analytical approach integrating a cognitive model and a dynamic Bayesian network

Communication errors (CEs) between crane operators and signalers are confirmed as leading causes of tower crane accidents. Prior studies focus on technological enhancements to reduce CEs, lacking a systematic discussion on their root causes and formation mechanisms. This study integrates a cognitive model and a dynamic Bayesian network (DBN) to qualitatively and quantitatively analyze CEs. The cognitive model was developed by integrating a Delphi study, interactive communication model, information processing model, information, decision, and action in crew context model, and Phoenix method. These help identify communication phases, cognitive activities, factors influencing communication, and their underlying mechanisms. DBN was developed by integrating computational models and multiple information sources to capture temporal evolution of individual cognition and CE probability. Demonstrated by case studies and sensitivity analysis, this study clarifies how CEs arise from a cognitive perspective, quantifies their occurrence trends under different construction conditions, and offers valuable practical implications for mitigating CEs.