Forecasting of Time Headway Distribution in Merging Area of Urban Expressway Interchange Based on RBF

Analyzing the reasons why the time headway distribution in the merging area outmost lane of urban expressway interchanges is different from that in the merging area of urban expressways. The WEIBULL model is chosen to simulate the time headway distribution, then researching the relations between the traffic flow factors and the WEIBULL parameters, and designing a RBF artificial neural networks with three input nerve cells and three output nerve cells to forecast the WEIBULL parameters, then based on the field data to train the networks, and to verify the RBF neural network based on different field data, and forecasting the parameters based on the RBF calibrated, at last, inputting the parameters of the WEIBULL model forecasted to simulate the field data, the simulating results are very good, and the test results are all accepted, so, to forecast the time headway distribution in merging area of urban expressway interchanges based on RBF artificial neural networks is feasible, and the accuracy is acceptable. The method is useful to the control and management of the interchanges or the capacity analysis of interchanges.