基于视觉特性的驾驶人预瞄时间建模

针对固定时长的预瞄时间无法真实反映驾驶人预瞄行为的问题，首先对自由流条件下的驾驶人视觉特性进行研究，得到直行路段驾驶人注视点多分布在车道中央、弯道路段驾驶人注视点多分布在内侧边缘线附近且视线近似与内侧边缘线相切的结论；在此基础上，探究直行和弯道路段驾驶人预瞄时间的求解方法，并借助驾驶模拟器进行试验，获得了17位驾驶人的预瞄时间。由于预瞄时间的原始数据分布离散程度高，难以直观地体现其随道路线形的变化规律，因而对预瞄时间随道路线形的分布特征进行研究，利用分组求中位数的方法进行数据处理，最终构建预瞄时间关于道路线形的分段线性函数和非线性指数函数。最后，通过数值仿真研究不同类型的预瞄时间（固定时长的预瞄时间、分段线性函数预瞄时间及指数函数预瞄时间）对人-车-路闭环系统动力学特性的影响。研究结果表明：驾驶人预瞄特性会对车辆的轨迹跟随和车道保持绩效产生重要影响；在闭环系统中使用随道路线形变化的预瞄时间可以显著提高车辆的轨迹跟随绩效，确保车辆在指定车道内行驶；驾驶人对道路环境的感知具有自适应性，通过其视觉特性可准确感知环境变化，并据此调整决策目标及其操纵特性。

Modelling of Driver's Preview Time Based on Visual Characteristics

Constant preview time cannot reflect a driver's preview behavior comprehensively. Owing to this fact, drivers' visual characteristics while driving were investigated. It was discovered that driver's fixation points mostly distributed in the middle of the lane on straight lane sections, and mostly clustered near the inner road edge on curved sections. Besides, drivers' sight line is almost tangential to the inner road edge. Based on this discovery, a method to obtain drivers' preview time on straight-line and curved sections was developed, and driving simulator tests were conducted for 17 drivers to obtain their preview time. While the original preview time has quite a discrete distribution, which does not suggest a relationship between preview time and road alignment, the distribution features of preview time with respect to road alignment were investigated. A grouped median was then carried out to process the original preview time, and a piecewise linear function and an exponential nonlinear function of preview time in terms of road radius were obtained. In the end, the influence of different types of preview time (constant preview time, piecewise preview time, and exponential preview time) on the dynamics of the human-vehicle-road closed-loop system was studied by numerical simulation. It was discovered that driver's preview characteristics had a great impact on vehicle's track following and lane keeping performance. While driver's preview time changed with road alignment, vehicle's track following and lane keeping can be greatly improved. The research revealed that driver is adaptive to the change of road environment. From driver's visual characteristics, a change of road environment can be accurately detected, and the decision-making goals and handling characteristics thereof can be adjusted accordingly.