基于TRACEr的管制员差错风险量化分析

Quantitative analysis method of ATCO’s error risks based on TRACEr

为准确辨识管制员(ATCO)的人为差错风险,量化预测不同认知领域的差错发生概率,根据认知差错回溯与预测分析(TRACEr)中的认知功能框架定义管制差错类型,分析心理差错机制的致因因素,将其作为根节点建立空中交通管理(ATM)不安全事件的贝叶斯网络;引入加权距离计算网络节点的条件概率,分析我国近10年的空管人误数据,以挖掘差错风险因素。研究结果表明:管制员的人为差错产生于各个认知领域,由此导致的不安全事件发生概率为0.0407;警觉性失效、未充分学习、忽略危险、沟通误解和信息传输不清晰是主要风险因素,产生判断差错和信息传输差错的概率较大。

In order to accurately identify human error risks of ATCO and quantitatively predict their probability in different cognitive domains,error types of ATCO were defined in accordance with cognitive functional framework of TRACEr,and causing factors for psychological error mechanism were analyzed before a Bayesian network of unsafe incidents in Air Traffic Management(ATM)was established with such factors as root nodes.Then,conditional probabilities of network nodes were calculated by introducing weighted distance,and human error data of unsafe incidents in China during last decade were analyzed to investigate risk factors.The results show that ATCO’s errors could be generated in any cognitive domain,which leads to a probability of 0.0407 of unsafe incidents.Among all factors,alertness failure,insufficient learning,neglecting danger,miscommunication and unclear convey of information are major ones,which results in a higher probability of judgement error and information transmission error.