Quantification of vehicle collision risk in smart way traffic environment to decrease traffic accidents

针对智能交通系统涉及的7起针对智能道路的交通事故风险，提出了一种智能道路引入后事故风险量化的方法，并从数值上阐明了事故的减少效果，选取了具有代表性的事故实例来描述一个基于人们具有最大可接受风险标准的模型。定义了关于碰撞车辆的不安全关系、不安全关系发生概率和不安全行为概率的三个识别结果。通过建立它们之间的关系，定量地说明了驾驶员的识别、判断和行动之间的关系。在此基础上，给出了支持该模型的证据，证明了该模型在理论上推导出的事故识别误差与事故判断误差之比与现有事故数据是等价的，并将该方法应用于其他事故。假设了一种对危险车辆发出警告的安全装置，并对事故的减少效果进行了定量评估。考虑到无安全装置警告的故障，得出了减少事故的必要的无警告程度。事实证明，高度精密的设备几乎无法实现工程上的能力，这是对严重依赖设备的驾驶员的要求，而对独立于设备的驾驶员的要求是放松的可靠性设备。结果还表明，安全设备可能能够告诉完全依赖设备的司机危险，但不应该保证安全。

This study focused attention on the risk of seven traffic accidents targeted for the smart way concerned with Intelligent Transport Systems and proposed a methodology to quantify the accident risk after the introduction of the smart-way, and then clarified reduction effects of accidents numerically.Representative example of the accidents was selected to describe a model which is based on a supposition that people has maximum standard of the acceptable risk. Three recognition results about an unsafe relation with a collision vehicle, occurrence probability of the unsafe relation and an unsafe action probability are defined. By formulating their relation, a relation between the recognition, the judgment and the action by a driver is shown quantitatively. Then, supporting evidence for this model is given by showing that a ratio derived by this model theoretically between the recognition errors and judgment errors in accidents was equivalent to existing accident data.This methodology was applied to the other accidents. A safety device, which gives a warning of a dangerous vehicle, was assumed and a reduction effect of accidents was evaluated quantitatively. Considering a failure lacking a warning by a safety device, necessary level of lacking a warning to reduce accidents was derived. It turns out that the extremely highly precise device, near to realization inability for engineering, is demanded from the driver who depends heavily on a device and relaxed reliability device is demanded from the driver who is independent on a device. And also a result was shown that a safety device might be able to tell the driver who depends completely on a device about danger but should not guarantee security.