Dynamic Freeway Path Travel Time Prediction Based on NPR Approach Using DSRC Data

摘要

The objective of this research is to develop a dynamic model to forecast path travel time betweenfreeway Interchanges (ICs) using path travel time data collected by Dedicated Short RangeCommunication (DSRC) system. This paper also includes the applicability of the proposedmodel using real-world data. This research was triggered by the drawbacks of the existing real52time based short-term prediction models, which have been widely applied for single intervalpredictions. The developed model is based on the k-Nearest Neighbor Non-ParametricRegression (KNN-NPR) using historical and current data collected by DSRC technology. In apractical test with real-time data, the presented prediction model performed effectively in termsof both prediction accuracy and computing time to the degree of the application of existingAdvanced Traveler Information Systems (ATIS).