论文部分内容阅读
针对HCM2010交织区速度预测模型存在的3个典型问题,分别提出了交织区速度预测分段模型和改进的回归模型。选取上海市快速路4个典型交织区作为研究对象,利用遗传算法分别对HCM2010模型、分段模型和改进的回归模型参数进行了标定。为进一步测试所提出的模型,利用仿真试验的方法,对所提出的改进回归模型进行了敏感性测试。结果表明:改进的回归模型预测精度最高,交织车流速度预测平均误差为9.52%,非交织车流速度预测平均误差仅为6.64%,且改进模型不会出现交织车流速度大于非交织车流速度这种明显违背现实的预测结果;新的3个交织区的验证结果表明,改进的回归模型交织车流速度预测误差为10.79%,非交织车流速度预测误差为10.45%,所提出的模型具有良好的适用性。
Aiming at the three typical problems in the HCM2010 intertwined zone velocity prediction model, the subsection speed model and the improved regression model are proposed respectively. Four typical intertwining zones of Shanghai Expressway were selected as research objects. Genetic algorithms were used to calibrate the HCM2010 model, the piecewise model and the improved regression model respectively. In order to further test the proposed model, the method of simulation experiment is used to test the sensitivity of the proposed improved regression model. The results show that the prediction accuracy of the improved regression model is the highest, the average prediction error of the interlaced traffic flow is 9.52%, the average prediction error of the non-interlaced traffic flow is only 6.64%, and the improved model does not appear obvious that the interwoven traffic speed is greater than the non-interlaced traffic speed The results of the new three intertwined zones show that the improved regression model has a predictable error of 10.79% and a non-interleaved flow velocity of 10.45%. The proposed model has good applicability.