Travel Time Prediction for Urban Networks: the Comparisons of Simulation-based and Time-Series Models

Travel time prediction for urban networks is an important issue in Advanced Traveler Information Systems (ATIS) since drivers can make individual decisions, choose the shortest route, avoid congestion and improve network efficiency based on the predicted travel time information. In this research, two algorithms are proposed to estimate and predict travel time for urban networks: the simulation-based and time-series models. The simulation-based model, DynaTAIWAN, designed and developed for mixed traffic flows, is adopted to simulate the traffic flow patterns. The Autoregressive Integrated Moving Average (ARIMA) model, calibrated with vehicle detector (VD) data, is integrated with signal delay to predict travel time for arterial streets. In the numerical analysis, an arterial street in Kaohsiung city in Taiwan is conducted to illustrate these two models. The empirical and historical data are used to predict and analyze travel time, including: travel time data from survey and historical speed data from vehicle detectors (VD).