基于时间递推与有师学习的交通路径动态诱导

发布时间：2018-04-13 01:30

本文选题：交通工程 + 路径动态诱导　；参考：《中国公路学报》2014年01期

【摘要】：为提高路况信息获取的预知性和路径引导的时效性,首先综合车辆运行过程中不同时段的路况差异和人因作用,运用时间递推原理,构建交通路径动态诱导整体结构模型,比较知识库累积经验路径与实时路况信息,为驾驶人的时间最短通行路径选择提供信息支撑;然后基于有师学习思路,对路径诱导逻辑内核结构进行解析,给出局部绕行和整体绕行2种算法实现的伪代码;最后通过对不同时刻出行和相同时刻出行的实例仿真分析,验证模型及程序的可靠性。研究结果表明:该交通路径动态诱导方法可辅助驾驶人及时对路况做出正确判断,减少因经验不足和突发事件造成的时间损失,适用于安装有GPS导航的车辆。
[Abstract]:In order to improve the predictability of road condition information acquisition and the timeliness of route guidance, firstly, by synthesizing the road condition difference and human factor action in different periods of time during vehicle operation, the whole structure model of dynamic guidance of traffic path is constructed by using the principle of time recursion.Comparing the accumulated empirical path with the real-time road condition information in the knowledge base to provide the information support for the driver's shortest passage path selection, and then based on the teacher-learning idea, analyze the path inducing logic kernel structure.The pseudo codes of local detour and global detour are given. Finally, the reliability of the model and program is verified by the simulation analysis of different travel at different times and the same travel at the same time.The results show that the method can help the driver to judge the road condition correctly and reduce the loss of time caused by the inexperience and unexpected events. It is suitable for the vehicle with GPS navigation.
【作者单位】：大连交通大学土木与安全工程学院;伦斯勒理工学院琼森工程中心;
【基金】：国家自然科学基金项目(50674052) 辽宁省自然科学基金项目(201202022) 大连市科技计划项目(2011E15SF118)
【分类号】：U491

【参考文献】