Research On Iterative Learning Control For Several Types Of Freeway Traffic Flow Models

As an important component of urban road network, freeway has brought great convenience to the way people travel. As freeway traffic congestion occurs frequently, the control of freeway traffic flow has become a hot research in research of transportation system. Iterative learning control(ILC), which based on strictly mathematical description,is one branch of the intelligent control methods. ILC doesn't depend on precise model. It's suitable to research subject with nonlinearity, repeatability and difficulty in modeling such as freeway traffic flow. In recent years, ILC for freeway traffic flow has received widespread attention.In this paper, based on freeway traffic flow ordinary differential model, open-loop PD-type iterative learning law is adopted to design the control method. Through the strictly mathematical proof, the convergence of the law is analyzed and the sufficient condition of convergence of iterative learning error is given. The effectiveness of the proposed law is confirmed by simulation analysis.Based on freeway traffic flow momentum model, open-loop PD-type iterative learning law is adopted to design the control method. The convergence of the law is analyzed and the sufficient condition of convergence of iterative learning error is given.By compare simulation analysis with open-loop D-type iterative learning law, it is confirmed that open-loop PD-type iterative learning control law has better effect on control performance. Furthermore, fuzzy control(FC) is used to adjust the gains of open-loop PD-type iterative learning control law and fuzzy open-loop PD-type iterative learning controller is designed. By compare simulation analysis with open-loop PD-type iterative learning law, it is confirmed that open-loop PD-type iterative learning control law which is improved by fuzzy control has an improvement on the convergence of the output error and control performance of the system.Based on freeway traffic flow distributed parameter system diffusion model,open-loop PD-type iterative learning law is adopted to control it. The effectiveness of open-loop PD-type iterative learning control for traffic flow on freeway traffic flow distributed parameter system diffusion model is confirmed by simulation analysis.Furthermore, particle swarm optimization(PSO) is used to optimize gains of iterative learning control and PSO open-loop PD-type iterative learning controller is designed. By compare simulation analysis with open-loop PD-type iterative learning law, it is confirmed that under the control of open-loop PD-type iterative learning control law which is improved by PSO, the output error of freeway traffic flow distributed parametersystem diffusion model has a faster convergence and the tracking performance is better.