基于模糊神经网络的结合部匝道控制方法研究

针对高速公路与关联城市快速路（简称结合部）路段拥堵日益严重的现状，从匝道控制影响要素分析入手，基于模糊控制和神经网络思想，本文提出了以主线交通状态与期望状态差值和匝道交通状态为输入变量，以匝道调节率为输出变量的模糊控制方法. 同时针对结合部路网互通式立交设计的实际情况，分单匝道控制和双匝道控制两种情况进行了分析，提出了相应的匝道控制方法，并建立了5层模糊神经网络控制模型. 最后以北京京津塘高速公路与北京三环和四环关联城市快速路为案例，对建立的模型进行效果验证，结果证明了所建立方法的有效性.

Ramp Metering Research of Junction Based on Fuzzy Neural Network Model

On relieving the growing serious urban traffic congestions, the problem of designing fuzzy control strategies for periphery networks that include freeway and related urban expressway(abbreviations: junction) is considered starting with the analysis on impact factors about ramp metering control. The presented novel methodology based on the fuzzy control and neural networks modeling paradigm whose input variables are the differentials between mainline real condition and expected condition and the ramp condition and output variables is the ramp control metering law. Based on the strategy, combining with space structure analysis of interchange in real world, the control metering method for single ramp and coordinated double ramps control metering is presented respectively and the corresponding 5 layer fuzzy neural network (FNN) model is also built. A preliminary simulation-based investigation of the ramp metering control problem for periphery networks consists of G2 freeway, the 3th and 4th ring road using the methodology demonstrates the comparative efficiency.