Study On Data Fusion Technology Of Multi-source Traffic Data Of Highway And System Development

With the sustained development of Chinese highway,a series of traffic problems such as traffic congestion,frequent accidents,environmental pollution and so on have appeared.The emergence of these problems makes the research has become an urgent subject about the main problems and the development direction of the highway in our country.However,due to the speed of the highway construction and the accumulation of management experience is not proportional to growth,so we need to put more effort to analyze and solve these problems.In order to study the traffic problems of highway,it is necessary to start with the study of traffic flow state,and to obtain the information of the traffic state and the distribution of vehicle operating law are the foundation and premise of the study.In this paper,Lian Huo highway-G30(Jiangsu section)was taken as the study object,and a series of research on the traffic state of the section was carried out.Firstly,this paper gave a detailed description of highway traffic detection technology,basic characteristics of highway section,the development of data fusion and technical characteristics and so on.Then chose two main detection means,the toll station and traffic survey station of the section,with two detection data as the foundation,proposed a method for time-space to match these two kinds of detection based on ArcGIS space,chose and improved a set of methods to extract the traffic information and preprocess the data based on the two kinds of data,and to detail the flow features of highway traffic.Secondly,combining advantages of two kinds of data and Wavelet neural network in prediction of short-time traffic flow,the paper constructed the Wavelet neural network prediction model based on two kinds of data and chose the weight distribution method as the contrast.Established the multi-source data fusion model based on highway on the basis of the relative error of the main three kinds of predictive fusion index.On the basis of the fusion prediction results,a Calman filter model was established to predict the traffic flow state of the road section on the basis of historical fusion results.Once again,the paper analysed the feature of the collected data and traffic state by single detector data.As a starting point,verified the multi-source data fusion prediction model and estimated the traffic flow state based on historical traffic information.Finally,this paper introduced the development and design of multi-source data comprehensive perception system taking the multi-source data fusion as the core,mainly from the system design,the main problems need to be realized of system,Web output display to do the main description.