Research On Traffic State Identification Technology For Control Subareas Based On Traffic Flow Forecasting

With the development of science and technology and the improvement of people’s living standard, the problem of traffic congestion has becoming more and more popular. Traffic congestion will not only reduce the quality of people’s life, but also bring economic recession, the waste of resources, environmental pollution and other issues, therefore, to alleviate traffic congestion brook no delay. Intelligent traffic control system is an important way to realize control and management of the city road network, and regional traffic state information are the premise of intelligent traffic control system, so it is necessary to study regional traffic state recognition technology.According to the traffic’s characteristic of our country, this paper mainly focuses on the traffic sub area division technology, short-term traffic flow forecasting technology and sub region traffic state recognition technology by using automatic control theory, artificial intelligence theory and traffic engineering. The specific contents are as follows:(1) Put forward static key intersection selection technique and dynamic key intersection selection technology respectively. The selection of static key intersection mainly relies on principle of arterial road, saturation, flow, time law and experience of traffic management department. The selection of dynamic key intersection depends on critical degree, in order to calculation critical degree this paper puts forward a method based on analytic hierarchy process.(2) Put forward a traffic control sub region division method base on improved genetic algorithm. Firstly, this paper presents a calculation method of intersection relation based on fuzzy algorithm. Then, it establishes a traffic control sub region division model, and puts forward a method based on improved genetic algorithm to solve the model, this method takes both effect of key intersection and intersection relation into account.(3) Put forward a short-time traffic flow forecasting method based on variable weight combination model. The paper analyzed the main characteristics of traffic flow and the factors which influence traffic flow, and then it proposed a combination model based on kalman filtering prediction model and neural network model, the weight of each model is adjusted through the error of last step. In order to make the model more stable, this paper also introduces an inertia factor.(4) Put forward a regional traffic state identification method based on fuzzy c-means ensemble classifier. In order to avoid lag of traffic control, this paper fuses the current traffic flow data and predicted traffic flow data, and then present a recognition method based on fuzzy c-means algorithm. In order to solve FCM’s problem of isolates cluster center, the paper uses ensemble classifier to do sub region traffic state identification.