Distributed Boundary Observer Design And Boundary Feedback Control For Urban Freeway Traffic Network

The urban freeway network is an important infrastructure of the urban transportation system,and it carries a large amount of urban traffic demand.However,the current speed of infrastructure construction is far from matching the growth rate of traffic demand,which leads to a slowdown of traffic speed.The intelligent transportation system can scientifically coordinate the existing infrastructure and the traffic demand,and make more efficient use of existing resources,which is an effective method to alleviate the congestion of urban freeway network.This dissertation will develop an intelligent transportation system for urban freeway network,including traffic states reconstruction and real-time control,in order to achieve real-time monitoring of the network and ensure stable traffic.Firstly,the modeling basis of the macroscopic traffic flow model is explored,and several classic macroscopic traffic flow models are further analyzed and compared.The Aw-Rascle-Zhang non-equilibrium traffic flow model is selected as the basis for modeling the urban freeway segments and transformed into Riemann coordinates for characteristic analysis.Based on the eigenvalue,the segments are divided into two categories,free regime and congestion regime,and then linearized and normalized.By analyzing the conservation relationship at the nodes between adjacent segments,the freeway segment model is extended to the urban freeway network system in the form of hyperbolic partial differential equation and corresponding boundary condition.Secondly,consider the large model characteristics of the urban freeway network,a distributed boundary observer is designed to reconstruct state estimation on the urban freeway network.Based on the consensus algorithm,a consensus-based static boundary observer was first designed to state correction with output feedback input.Then a consensus-based dynamic boundary observer was designed for state correction,in which is dynamically adjusted by the ordinary differential equation and then feedback to the input.The latter observer can achieve smaller steady-state errors.The sufficient conditions of detectability are provived for both two kinds of observers,respectively,and the stability is proved by Lyapunov technique.At last,two numerical simulations are designed to verify the stability of two kinds of observers,meanwhile the results are analyzed and compared.Thirdly,consider the urban freeway network system with bounded uncertainties in traffic demand,the error model is established with respect to desired steady state to facilitate the control system.At the same time,a boundary feedback controller is designed at the nodes between adjacent segments with integration of the on-ramp metering and mainstream variable speed limit to stabilize the traffic with spontaneously increased entropy.Further,the segment model is extended ot the network,corresponding feedback control law is deduced as well,and then yield the boundary condition.After that,the sufficient condition for the controller’s input to state stability is given,and proved by Lyapunov technique.At last,a numerical simulation is designed to verify the input-to-state stability of the controller,meanwhile the control effect is analyzed.Finally,a traffic simulation platform Aimsun,is used to conduct the traffic simulation experiment on the 4th ring road of Beijing.Part of the Beijing 4th ring road is built and and the parameters of traffic is deployed in Aimsun.The application program interface provided by Aimsun was used for secondary development to perform the traffic simulation experiment under the action of the developed boundary feedback controller.Finally,the experimental data is statistically processed,and the effect of the controller on urban freeway network is analyzed from both macroscopic and microscopic levels.