Travel time estimation from loop detector data for advanced traveler information systems applications

The feasibility of the use of detector data to improve the performance of functions for advanced traveler information systems was examined. Relationships between travel times and flow/occupancy information were assessed using the NETSIM simulation model. Several simulation experiments were performed to incorporate variations in entry flows, turning movement percentages, and the randomness of traffic phenomena. Statistical models were developed, capable to convert flow and occupancy values to arterial through link travel times. Link-specific and generalized-type models were developed and validated extensively with simulated as well as empirical data.; The results indicate that travel time is independent from both flow and occupancy under low traffic demand conditions. Moreover, travel time predictions are not possible, when occupancy exceeds 90%. However, for occupancies between 30% and 90%, travel time and occupancy are significantly correlated. Regression equations are fitted for this occupancy range, including single variable, link-specific models as well as multiple variable, generalized models. The developed models are intented to be applied in a data fusion context, to provide improved estimates of arterial link travel time.