| With the development of transportation industry in our country,various data access facilities and technologies were widely used in the expressway domain.Under such Big Data in expressway domain,the traffic management and operation departments of expressways are eager to share and analyze the multi-dimensional data.Therefore the quality of data analyses requirements of the traffic control department are stringent to control the expressway capacity.In recent years,short-term traffic flow prediction has been widely used in urban transportation as a core research topic in the field of intelligent transportation,providing data support for traffic guidance and decision-making for transportation management departments.The short-term traffic flow prediction on Big Data in expressway domain is focused with spatio-temporal characteristic to analyze the diversity and complexity of Big Data in expressway domain.The specific research contents are listed as follows:1.Against the low data quality of Big Data in expressway domain and the lack of data required for prediction,the data preprocessing is focused.To ensure the rationality and validity of the data,a data cleaning method for massive spatio-temporal data is proposed to judge,modify and verify the spatio-temporal attributes of raw data,and filter the data,according to the business rules.On the cleaned data,multi-dimensional statistical analyses are done.Combined with the spatial correlation of expressway,the dependent data sets required for prediction were constructed to provide effective data support for further short-term traffic flow prediction.2.Against the insufficient use of data spatio-temporal characteristics,and dynamic growth of data size,a short-term traffic prediction model for Big Data in expressway domain is proposed.The spatio-temporal correlation analysis of the cleaned and analyzed data is carried out.Combined with the non-parametric regression model,a short-term traffic flow prediction model based on spatio-temporal characteristics is constructed.The experimental results show that the proposed prediction model has higher accuracy and scalability than the traditional time series model,and can predict the traffic flow value of the entire locations in expressway network.3.Against the data in expressway domain,most business system insufficient use of Big Data technology analysis,a traffic flow prediction and analysis system of expressway is built.Using large-scale expressway information to conduct and analyse experiments which verifying the scalability,high efficiency and high precision of the prediction model,and presents the data result to the prediction analysis system.It provides reliable monitoring platform for the Henan provincial transportation department.The monitoring platform,has achieved the effect of real-time monitoring and prediction analysis of expressway traffic flow. |
