论文部分内容阅读
根据基于支持向量回归机的交通状态短时预测方法建立了数学模型,考虑以交通检测器收集到所要预测时刻前几个时段及被测路段上下游前几时段的交通流量、车道占有率、平均线速度等交通参数为输入,以对应时段的平均线速度为输出.选取核函数,对支持向量回归机进行训练.应用训练完成的支持向量回归机,利用输入参数预测下时段的交通线速度.最后,以北京市北四环某路段的实时监测数据来对模型进行检测,预测结果表明了模型的有效性.
According to the traffic state short-term forecasting method based on support vector regression machine, a mathematical model is established. Considering the traffic flow collected by the traffic detector several moments before the predicted moment and the traffic flow before the upstream and downstream of the measured section, the occupancy of the lane, Line speed and other traffic parameters as input and output the average line speed of the corresponding time period.The kernel function is selected to train SVR.Through the training of support vector regression machine, Finally, the real-time monitoring data of a section of North Fourth Ring Road in Beijing is used to test the model. The prediction results show the effectiveness of the model.