论文部分内容阅读
针对非参数回归短时交通流量预测算法的状态向量选取问题,基于高速公路交通流量在空间上演变趋势明显的特点,提出交通流量预测的改进非参数回归算法。引入各上游断面车流到达当前断面的行程时间作为状态向量选取的依据,并根据各上游断面影响程度的不同,调整相似机制的计算方法。利用渝武高速公路微波检测器数据对该模型进行验证。结果表明,改进的非参数回归算法克服了固定状态向量定义不能满足同一断面不同交通状态的缺点,对各种交通状态具有更好的适应性,预测精度更高。
Aiming at the selection of state vector of non-parametric regression short-term traffic flow prediction algorithm, an improved non-parametric regression algorithm of traffic flow prediction is proposed based on the obvious spatial evolution trend of expressway traffic flow. Introduce the travel time of each upstream section traffic arriving at the current section as the basis for selecting the state vector and adjust the calculation method of similar mechanism according to the influence of each upstream section. The model was verified by using the data of microwave detector in Yuwu expressway. The results show that the improved nonparametric regression algorithm overcomes the shortcoming that the definition of fixed state vector can not meet the different traffic conditions of the same section, and has better adaptability to various traffic conditions and higher prediction accuracy.