几类快速路交通流模型的迭代学习控制研究

发布时间：2018-03-24 16:04

本文选题：迭代学习控制　切入点：快速路交通流常微分模型　出处：《广西科技大学》2015年硕士论文

【摘要】：快速路作为城市交通路网中的重要组成部分,给人们的出行带来了极大的便利,快速路交通拥堵问题的频繁发生,使得快速路交通流控制成为了交通系统研究中重要的研究热点。迭代学习控制(Iterative Learning Control,ILC)是智能控制方法中的一个分支,建立在严格的数学描述基础上,不需要依靠精确的数学模型,非常适合快速路交通流这类具有非线性、重复性、难建模等特点的研究对象。近年来,快速路交通流的迭代学习控制受到了广泛关注。本文针对快速路交通流常微分模型,采用开环PD型迭代学习律设计了控制方法,通过严格的数学证明,对算法收敛性进行了分析,给出了迭代学习误差的收敛条件,再通过仿真分析验证了算法的有效性。针对快速路交通流动量模型,采用开环PD型迭代学习律设计了控制方法,并进行了收敛性分析,给出了收敛条件,通过和开环D型迭代学习控制律的仿真对比分析,验证了开环PD型迭代学习控制律具有更好的控制效果。然后利用模糊控制(Fuzzy Control,FC)算法对开环PD型迭代学习控制律的学习增益进行调节,设计了模糊开环PD型迭代学习控制器,通过和开环PD型迭代学习控制律的仿真对比分析,验证了利用模糊控制算法改进后的迭代学习控制律提高了系统输出误差的收敛速度,改善了控制性能。针对快速路交通流分布参数系统扩散模型,采用开环PD型迭代学习律进行控制,通过仿真分析,验证了开环PD型迭代学习控制律在快速路交通流分布参数系统扩散模型的交通流控制上的有效性。然后采用粒子群(Particle Swarm Optimization,PSO)算法对迭代学习控制律的迭代学习增益进行优化,设计了粒子群开环PD型迭代学习控制器,通过和开环PD型迭代学习控制律的仿真对比分析,验证了快速路交通流分布参数系统扩散模型在粒子群算法改进后的迭代学习控制律的作用下,系统误差具有更快的收敛速度,系统输出具有更好的跟踪性能。
[Abstract]:As an important part of urban traffic network, expressway brings great convenience for people to travel. Iterative learning control (iterative Learning control) is a branch of intelligent control method, which is based on strict mathematical description. No need to rely on accurate mathematical models, it is very suitable for the expressway traffic flow with the characteristics of nonlinear, repeatability, difficult modeling and so on. In recent years, The iterative learning control of expressway traffic flow has been paid more and more attention. In this paper, an open-loop PD iterative learning law is used to design the control method for the ordinary differential model of expressway traffic flow, which is proved by strict mathematics. The convergence of the algorithm is analyzed, the convergence condition of iterative learning error is given, and the validity of the algorithm is verified by simulation. For the expressway traffic flow model, an open-loop PD iterative learning law is used to design the control method. The convergence conditions are given, and the results are compared with those of open-loop D-type iterative learning control law. It is proved that the open-loop PD iterative learning control law has better control effect, then the fuzzy control fuzzy control FCalgorithm is used to adjust the learning gain of open-loop PD iterative learning control law, and the fuzzy open-loop PD iterative learning controller is designed. By comparing with the open-loop PD iterative learning control law, it is proved that the improved iterative learning control law using fuzzy control algorithm can improve the convergence speed of the output error of the system. The control performance is improved. The open-loop PD iterative learning law is used to control the diffusion model of expressway traffic flow distribution parameter system, and the simulation analysis is carried out. The effectiveness of the open-loop PD iterative learning control law in the traffic flow control of the diffusion model of the traffic flow distribution parameter system on the expressway is verified. Then, the iterative learning gain of the iterative learning control law is optimized by using particle swarm optimization algorithm (PSO). An open loop PD iterative learning controller based on particle swarm optimization is designed and compared with the open loop PD iterative learning control law. It is verified that the system diffusion model of expressway traffic flow distribution parameters has faster convergence speed and better tracking performance under the improved iterative learning control law of particle swarm optimization algorithm.
【学位授予单位】：广西科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491.54;TP18

【相似文献】