摘要
论文在模糊神经网络内部引入递归环节,设计了入口匝道的动态响应调节算法,介绍了算法的模糊神经网络结构、隶属度函数、模糊规则。由于递归神经元有内部反馈连接,可以捕获系统的动态响应,能简化网络模型,网络各个参数具有明确的物理意义,可根据经验选择初始值,且其是一个动态映射网络,比普通模糊神经网络更适于描述动态系统。最后分别通过数值仿真试验和交通TSIS模拟实验,详细分析了入口匝道智能控制的效果,仿真结果表明论文设计的入口匝道模糊神经网络控制算法在控制效果上比常规定时、Alinea控制显示了较大的优势,在重要指标上优于定时控制策略和Alinea控制策略。
This paper combined the adjustment strength of fuzzy control and neural network in freeway ramps,and designed fuzzy neural network controller and specially the controller for freeway on-ramps.This methodology considered the optimization of fuzzy principle and chose a few input and output quantities.It used ant colony algorithm training coefficient of network,aiming to control freeway mainline traffic flow,and provided adjusted quantity for ramp controlling rate according to fuzzy principle.At the end,the paper evaluated the adjustment algorithm by simulation.The simulating examples show that this controller can adjust traffic flow to the expected condition rapidly,and stabilize traffic flow density better,and that compared with the traditional method ALINEA,the convergence is faster and the operation is more efficient with better system stability.
出处
《交通信息与安全》
2011年第3期60-64,共5页
Journal of Transport Information and Safety
关键词
交通控制
匝道调节算法
动态模糊神经网络
交通仿真
traffic control
ramp adjusted algorithm
dynamic fuzzy neural network
traffic stimulation