基于驾驶员视觉感知机理的隧道交通标志动态视认行为研究

当前隧道交通事故频发且严重程度很高，提高隧道交通标志视认性是解决其安全问题的重要的措施之一。视觉心理学和多智能体仿真方法为驾驶员隧道交通标志视认行为的研究提供了很好的思路。 本项目首先提取影响驾驶员对隧道交通标志视认的关键因素，基于模拟驾驶实验数据对其进行量化，获得模型研究的前提条件边界。 接下来，针对信息驱动的bottom-up和视觉任务驱动的top-down两个视认过程模型的框架特征，把驾驶员对隧道交通标志的感知、记忆、思维和决策统一起来，形成能在各自特定条件下准确反映驾驶员隧道交通标志动态视认行为的认知模型。 然后探讨bottom-up和top-down两种模型需要协同作用才能解释的边界条件及不同条件下两种模型的协同机制，并在此基础上开发多智能体仿真平台，在实证基础上进行归纳总结和一般性扩展，提出由交通标志视认行为模型过渡到实用设计指南和标准的方法。

Currently, the tunnel traffic accident is frequent and serious. One way of making it safer is to improve tunnel traffic sign visibility.Visual psychology and multi-agent simulation are thoughtful approaches for research on drivers' tunnel traffic sign visibility. First of all, this project is going to extract critical factors that influence driver's perception of traffic sign visibility and quantify these factors based on driver simulation experiment data in order to obtain prerequisite boundary condition of the model. Furthermore, as to the information drived bottom-up and visual task drived top-down visual recognition framework, we will unify driver's perception, memory, thought and decision making process of tunnel traffic signs, forming accurate driver's tunnel dynamic visual recognition model under specific condition. Then we plan to discuss the boundary condition that requires synergistic effect of both bottom-up and top-down approaches and the synergistic mechanism of these two models under different conditions. Based on this, multi-agent simulation platform could be developed. Classic tunnel traffic sign dynamic visibility optimization scheme can be summarized and extended based upon empirical research. The methodology of transition from traffic sign visibility behavior theory to the practical design guidelines and standards should be proposed.