基于驾驶人视觉感知的低等级公路行车速度预测

发布时间：2018-02-28 04:26

本文关键词： 低等级公路视觉车道模型行车速度视觉路侧环境 GABP神经网络　出处：《同济大学学报(自然科学版)》2017年03期 　论文类型：期刊论文

【摘要】：低等级公路环境下的行车速度变化特征比较复杂,已有的基于几何线形指标的车速模型不足以全面地表述这种特征.通过大量实车实验发现,低等级公路环境下驾驶人所采用的行车速度和其视觉感知到道路条件之间存在很好的相关性.驾驶人感知的道路条件分为视觉车道信息和视觉路侧环境信息.采用Catmull-Rom样条曲线拟合的视觉车道模型能够反应驾驶人感知的低等级公路几何特征,根据视觉车道模型形状参数,使用遗传算法优化BP神经网络(GABP)建立了基于几何信息感知的几何车速模型.同时,基于路侧环境信息利用Logistic回归模型建立了路侧车速修正模型.将上述模型结合起来,形成了基于驾驶人视觉感知的低等级公路行车车速预测方法.由此方法计算的驾驶人行车速度与实测情况吻合性好,能够很好地描述驾驶人在低等级公路环境下通过对道路条件视觉认知而产生行车速度的行为特征,是低等级公路运行车速预测计算的一种有效方法,不仅可以作为道路安全评价的基础,也可以为基于行车速度的公路几何设计提供有力支持.
[Abstract]:The changing characteristics of driving speed in low-grade highway environment are quite complex, and the existing speed model based on geometric linear index is not enough to describe this feature comprehensively. It is found by a large number of real vehicle experiments that, There is a good correlation between the driving speed and the visual perception of the road condition in the low-grade highway environment. The road conditions perceived by the driver can be divided into visual lane information and visual roadside environment information. The visual lane model fitted with Catmull-Rom spline curve can reflect the geometrical characteristics of the low-grade highway perceived by the driver. According to the shape parameters of visual lane model, a geometric speed model based on geometric information perception is established by using genetic algorithm to optimize BP neural network (GABP). At the same time, Based on the road side environment information, the Logistic regression model is used to establish the road side speed correction model. A method for predicting the driving speed of low-grade highway based on the driver's visual perception is developed. The calculated driving speed of this method is in good agreement with the measured results. It is an effective method to predict and calculate the driving speed of low grade highway, which can describe the behavior characteristics of driving speed generated by the visual cognition of the road condition in the low grade highway environment, and it is an effective method to predict the running speed of the low grade highway. It can not only be used as the basis of road safety evaluation, but also can provide powerful support for highway geometric design based on driving speed.
【作者单位】：同济大学道路与交通工程教育部重点实验室;
【基金】：“十二五”国家科技支撑计划(2014BAG01B06) 国家自然科学基金(51078270)
【分类号】：U491.254

【相似文献】