Estimation And Prediction Of Link Travel Time For Urban Freeway

Link travel time is an important parameter to describe the traffic state, it can evaluate the patency of the road and reflect the efficiency of the road transportation very well, so it plays an important role in traffic planning, traffic management and traffic control, and it also occupies an important position in research and application of the current intelligent transportation systems.Considering the microwave detector technology has the characteristics of mature technology, easy data acquisition and low cost, this article has improved a travel-time domain algorithm to estimate link travel time of the urban freeway. Firstly, the algorithm has assumed that the real-time detection speed of microwave detector is the average speed of road unit in different time unit, then constructed the travel-time domain of vehicle’s trip, finally gotten the link travel time by imitating the vehicle passing through the travel-time domain. This algorithm is verified on the basis of the microwave detection data of Beijing second ring freeway. The result shows that, compared with the traditional static travel time estimation method, this new way significantly improves the accuracy of travel time estimationNot only getting the link travel time at the moment is very important, but also predicting the future travel time is very important. For the purpose of improving the accuracy of travel time prediction, this article has built a travel time prediction which is based on wavelet neural network. Taking the travel time of Beijing second ring freeway as experimental data which is obtained by travel-time domain algorithm, this model is also verified by creating some prediction examples according to different parameters and different sample data, and made comparison with BP neural network model for prediction error. The result shows that, the wavelet neural network model can reflect the mapping rules of input and output. By means of comparing each practical prediction result and combining with the research on previous prediction of link travel time, This article has also analyzed the causes of the error and found out the reason of higher prediction accuracy of this model. As for the research on the field of traffic parameter’s prediction, the link travel time prediction model and the related discussion about it in this paper has some innovative significance and reference value.