Study On Traffic Signal Coordination Control Of Urban Trunk Roads

In recent years, with the growth of the economy, urban construction develops quickly and the amount of motor vehicles increases sharply. Besides, traffic jams, traffic accident, and environmental pollution problems become more and more serious. Urban roads become crowded day by day and have been treated as the bottleneck of economic development and urban construction.In the urban traffic control process, model based traditional methods cannot get an effective result since the traffic is a time-varying random system, which is also complicated and indefinite. So the intelligent control approaches are often used to study the urban traffic signals at present.To begin with, this thesis provides a method of single intersection signal control based on Fuzzy Neural Network Control by analyzing the current situation of the research on urban single intersection signal control. Density of the vehicles of the current and next phase can be calculated by detecting the queue length of the current and next phase, from which we can determine whether the phase should be changed or not. The performance of the controller is evaluated by the average vehicle delay of the intersection in each period. The Matlab simulation results are compared with the induction control approach. Comparison shows that our method can decrease the average delay of the vehicles at the intersection, which proves the effectiveness of our method. Besides, the structure of urban trunk roads is analyzed. Take the correlation of the adjacent intersections into account, a fuzzy neural network coordination control method of the adjacent intersections is proposed. This method is used for the trunk roads which are made up of two adjacent four-phase intersections. In this paper the adjacent traffic intersections are considered as a whole system. The first-level controls every single intersection and adjusts the ratio of green signal in different directions according to the period given by the coordination unit. Using the traffic data of each intersection and the Webster minimum delay cycle length equation, the second-level determines the signal cycle which is used consistently by the adjacent intersections in the next stage. Neural network fuzzy control method is also used where the vehicle density between the two intersections is the input of the neural network and vehicle speed between the two intersections is the output. The phase difference between two adjacent intersections can be determined by the ratio of the distance to the vehicle speed of the two adjacent intersections. The simulation program is coded by MATLAB. The results show that our method can decrease the average delay of the vehicles at the adjacent intersections effectively. Compared with the traditional induction control method, the performance of our approach is higher. So the purpose of improving the traffic system can be achieved successfully.