Model-Free Adaptive Traffic Signal Control For Urban Traffic Networks

Based on model-free adaptive control(MFAC)theory,this thesis conducts in-depth research in the field of urban traffic network signal control.The significance of these works is that urban traffic signal control no longer depends on the mathematical urban traffic model.Instead,urban traffic controller can be designed only by using the input/output data of the urban traffic systems.The main research contents and contributions of this thesis are summarized as follows:1.Aiming at the problem that there is a congested region in the urban traffic network which leads to unbalanced traffic distribution,an MFAC-based hierarchical perimeter control strategy for single-region urban traffic networks is proposed.At the periphery of the congested region,the signal settings are optimized to meter the number of vehicles entering the congested region;while inside the congested region,traffic flow distribution within the region is optimized to improve the traffic efficiency.The perimeter controller uses a centralized MFAC control method,while the inner controller uses a decentralized estimation decentralized MFAC method,to derive the detailed signal settings in the periphery of the congested region and inside the congested region.Finally,joint simulations using the VISSIM simulation platform and MATLAB software verify the feasibility and effectiveness of the proposed control strategy based on the real traffic demand and network topology of the traffic network of Weifang,Shandong Province,China.2.Aiming at the problem of traffic flow balance and coordination among regions in multi-regional urban traffic networks,a distributed model-free adaptive predictive urban traffic control strategy is proposed.Firstly,the MFAC data model of each region in the network is obtained by dynamically linearizing the nonlinear traffic dynamics of the region.Then the derived MFAC data model is used to design the distributed predictive controller for multi-region urban traffic networks.Finally,joint simulations using VISSIM simulation platform and MATLAB software verify the feasibility and effectiveness of the proposed control strategy based on the real traffic demand and network topology of the traffic network of Linfen,Shanxi Province,China.3.Due to the characteristics of similarity and repeatability of urban traffic flow,an decentralized urban traffic control based on the model-free adaptive iterative learning control(MFAILC)scheme is proposed.Firstly,the nonlinear dynamics of the intersections in the network are dynamically linearized along the iteration axis,and then transformed into the MFAILC data model.Afterwards,the MFAILC-based traffic controller is designed to minimize the differences of queue lengths among different approaching links of the intersection.Finally,Simulation study based on a simple four-intersection network verifies the feasibility and effectiveness of the proposed method.4.A network-wide urban traffic control strategy and a traffic data dropout compensation method based on MFAILC scheme are proposed.Firstly,the nonlinear dynamics of an urban traffic network are dynamically linearized along the iteration axis,and transformed into the MFAILC data model.Then,the MFAILC-based traffic controller is designed to minimize the number of vehicles in the traffic network.Afterwards,to deal with the problem of traffic data dropout in practical traffic networks,an MFAILC-based traffic data compensation method is proposed.Finally,Joint simulations using VISSIM simulation platform and MATLAB software verify the feasibility and effectiveness of the proposed control strategy based on the real traffic demand and network topology of the traffic network of Linfen,Shanxi Province,China.