基于仿真的隧道检测器优化布设及短时交通流预测

Optimized Layout of Detectors and Short-term Traffic Flow Forecasting of Tunnel Based on Simulation

隧道是城市交通的咽喉所在,为了有效引导隧道交通流,实现隧道交通状况的实时发布,从而引导驾驶员理性进出隧道,准确可靠的数据采集及实时交通流预测是关键所在。以VISSIM仿真为基础,建立研究隧道路网模型,以行程时间误差为依据,探讨了检测器在隧道中的空间合理布局,保证数据采集的可靠性。在准确的数据采集基础上,选取了自适应权重指数平滑法(AWES)以及基于径向基函数的神经网络模型(RBFN)进行实证研究,并采用数据融合技术以保障预测精度及误差的稳定性。并进一步以VISSIM仿真平台简略探讨了交通状态判别指标,以期更好服务于城市交通诱导系统。

Tunnel can be regarded as the throat of urban transportation. For guiding the traffic flow, making real-time release of traffic conditions in tunnel, and guiding the drivers driving into and out of tunnel rationally, accurate and reliable data collection and real-time traffic flow forecasting is the key. We established a tunnel model using VISSIM simulation software, and explored rational space distribution of tunnel detectors based on travel time error to ensure the reliability of data collection. On the basis of accurate data collection, we used the method of adaptive weight exponential smoothing （AWES） and the radial basis function neural network model （RBFN） to make empirical studies, and used data fusion technology to insure the prediction accuracy and error stability. Furthermore, we briefly analyzed the indexes of traffic state identification based on VISSIM simulation platform to serve the urban traffic guidance system better.