Algorithm Research Of Travel Time Forecast For Urban Expressway

Travel time forecast is one of the key issues in Intelligent Transportation System (ITS) Research. Real-time and accurate travel time forecasts could reduce traffic congestion, by providing effective information for route guidance to Advance Traffic Management System (ATMS) and Advanced Traveler Information System (ATIS). And the urban expressway is the main artery of the city transport system, so there is great significance in the travel time forecast of the urban expressway.The existing travel time forecast algorithms always predict the travel time directly, and the algorithm parameters often requires a large amount of real travel time data for identification. The performance of existing algorithms is not very satisfactory without great deal of travel time data. So based on theoretical analysis in the wavelet transform theory, BP (Back Propagation) neural network and data fusion, this paper presents a new algorithm to predict the urban expressway travel time indirectly. The algorithm, which applied to the urban expressway with a large number of fixed detectors, is based on the wavelet transform denoising, BP neural network and data fusion.Wavelet transform is introduced into traffic flow data preprocessing to enhance data accuracy. And the BP neural network predicts the future traffic flow velocity using the historical data. Because detectors in the section have spatial difference, Data fusion technology is adopted in merging every velocity into the average speed of the section. And the average speed is converted into the average travel time by two different methods. This paper chose the north of Beijing 3-ring as a test section. There are four microwave detector fixed on the test section. Video recognition technology is introduced into obtaining the travel time data, and this data is used to verify the algorithm. The results show that the error of the algorithm is around 10 percent. So the algorithm proposed in this paper is feasible and has higher precision. On the basis of the algorithm, we use the Visual C++, SQL Server and MapX technology to program a utility software.