摘要
已有快速路入口匝道控制手段是以定时控制方法为主,虽然存在动态调整等方法,但缺乏预测机制,这主要是由于车流的动态性和随机性而难以进行定量分析,引入人工神经网络可对车流进行动态预测。分析了影响主线交通量的与匝道相关的因素,并在此基础上建立了神经网络预测模型,通过上海典型匝道(延安路—江苏路)一组实测数据对网络进行训练和预测,得到了满意的效果。
Existing control measures are dominated by pre-timed control; others such as ALINEA-control and DC-control in VISSIM are limited for lacking of predictive mechanism, which is mainly because it is hard to make quantified analysis for the dynamics and randomicity of the traffic flow. By drawing into the neural net can make dynamic prediction of the traffic volume, based on which the ramp-control can be ad- justed in time. This paper first analyze the related factors affecting the main line's traffic, then neural network expressway ramp control model is established on them. The verification of the actual observation data's training and testing on it shows a satisfactory results.
出处
《交通科技与经济》
2008年第6期101-103,共3页
Technology & Economy in Areas of Communications
关键词
BP神经网络
匝道控制
动态预测
随机性
BP neural network
ramp control
real-time forecast
randomicity