Structure Analysis and Prevention Measures of Traffic Accidents

本研究提出交通事故的结构分析，以评估可能的交通安全措施。交通事故的特点是驾驶员/导航员/飞行员周围的交通状况和道路/航道/机场配置等诸多因素共同促成了交通事故的发生。这一特点表明应用故障树分析（FTA）来明确交通事故的结构是有用的。由于调查和报告制度的建立，过去的交通事故记录丰富。考虑到这些特点，本文提出了一种基于事故记录构建交通事故故障树的方法。该方法分为两个阶段，从事故记录中提取部分傅立叶变换，并将提取的部分傅立叶变换叠加形成统一的傅立叶变换。为了找出事故的主要原因，评估安全措施的效果，需要进行重要性分析。然而，传统的重要性分析方法并不一定适用于此类交通事故分析。虽然传统的技术假设因果事件的发生是明确判断的，但在ft中存在许多与人有关的因果事件，这些事件是否发生难以明确判断。为了解决这一评价困难，引入模糊集理论的概念，发展了一种新的技术。这些技术被安排成一个程序，并通过一些选定的船舶碰撞和机场交通事故的具体记录进行了演示。然后将所得的时域波动系数应用于安全评价所提出的安全措施的有效性。通过案例研究，获得了许多有用的信息，包括防止几个因果事件同时发生的重要性。研究结果为制定有效的交通系统安全措施指明了方向。

This study proposes a structure analysis of traffic accidents for evaluation of possible traffic safety measures. Traffic accident is characterized that many factors in traffic condition and road/fairway/airfield configuration surrounding driver/navigator/pilot jointly contribute the occurrence of traffic accident. This characteristics suggests the usefulness of applying Fault Tree Analysis (FTA) to making clear the structure of traffic accidents. There is a rich stock of past traffic accident records because of the surveying and reporting systems.Taking these characteristics into consideration, this study proposed a method to construct a Fault Tree (FT) of traffic accidents from accident records. This method consists of two stages, extracting a partial FT from an accident record and superimposing the extracted partial FTs to form a unified FT.For finding dominant causes of accidents and for evaluating the effect of safety measures, an importance analysis should be conducted. However, the conventional technique of importance analysis cannot necessarily be applied directory on such a FT of traffic accidents. Though the conventional technique assumes that the occurrence of causal event is clearly judged, many human-related causal events, which is difficult to be clearly judged whether it occurred or not, exist in the FT.For resolving this difficulty in evaluation, a new technique has been developed by introducing the concept of fuzzy set theory.These techniques were arranged into a procedure, and was demonstrated in the concrete by using several selected records of ship collision and airfield traffic incident. Then the FTs obtained were applied to a safety evaluation how effective the proposed safety measures are. Through the case study, many useful information was obtained including the importance for preventing simultaneous occurrence of several causal events. The results suggest us a direction in developing effective safety measures of transportation systems.