Research On Predictive Framework Of The Signal Control System

Short-term traffic flow forecasting is an important foundation for the traffic control system. In order to relieve the traffic pressure, reduce pollution and energy conservation, it can provide effective real-time traffic information for advanced information systems and advanced traffic management systems. With the title of Research on Predictive Framework of the Signal Control System, after a brief introduction of the urban road traffic signal control system, the control system framework for forecasting system was analyzed, and several key parameters were discussed and studied. To the regional control, the prediction model based on the model predictive control was introduced. Using the prediction model, the critical routes in the urban traffic network were recognized, in order to serve the new methods of the regional control.This paper first briefly describes the framework for traffic flow forecasting system, including all levels of the system. Secondly, the forecasting framework of the system was descripted with a number of parameters, including the traffic flow, the vehicle travel time, the vehicle turning movement proportion, platoon recognition, and each part included present situation and the basic theory and method. Thirdly, the vehicle turning movement proportion was focused on the importance of forecasting and implementation model. Based on the characteristic of the complex transport system, neural network prediction model was introduced, and the model predictive control was chose to optimize the results of the forecasting, which made a good prediction.Finally, the recognition model of the critical routes in the urban traffic network was studied. The traffic network was transformed into the directed group. Based on the model of forecasting the vehicle turning movement proportion, the turning relationship of the traffic flow was analyzed. Considered to the traffic flow in the area traffic network, the critical routes were recognized in the traffic network.