摘要
路段行程时间超前预报是动态交通诱导方案制定的基础,应用短时交通预测的方法可以获得将来某个时刻的路段行程时间数据,但已有研究成果,还存在适应性不强,计算量大,基础数据需求多等不足。应用谱分析及Karhunen-Loeve变换对随机序列的分解与重构功能,通过挖掘路段历史行程时间序列与当前检测行程时间序列的相似性特征进行序列重构,实现对后一时段路段行程时间的预测,结果显示,该方法具有良好的预测精度。
Road travel time prediction is a crucial foundation for dynamic traffic guidance and the short traffic forecasting method can be applied to get near future travel time of the road. But the previously developed models have some deficiency, such as bad adaptability, large amount of calculation needing and history data requirement. This study is to develop a model that can estimate the road travel time by using spectral analysis and Karhunen-Loeve transformation. Karhunen-Loeve transformation can decompose and restructure random sequences, utilize the similarity of current detection series and history series, and then realize estimation of the following travel time series by means of reconstruction. The case study suggests that the proposed method has a good performance.
出处
《华东交通大学学报》
2014年第6期1-6,共6页
Journal of East China Jiaotong University
基金
江西省科技厅科技计划项目(20123BBE50094)
江西省自然科学基金项目(20142BAB201015)
