With the advent of Advanced Traveler Information Systems (ATIS), short term travel time prediction is becoming increasingly important. Over the last few decades researches in this field have mostly concentrated on the applications to freeway facilities, while very few studies have focused on urban networks. This preference of researches for freeways can be identified from the literature overview. There are no reliable and applicable methods for urban short term travel time prediction. The intention of this dissertation is to make an attempt on this subject. In conclusion, in this dissertation we present an accurate and robust model for short term urban travel time prediction. This research is the first attempt to combine a model based approach and data driven approach. The model has a generic structure. In that sense, it can be applied on any urban route equipped with traffic data collection systems (single loop detectors, license plate cameras and traffic signal timings). It can also be extended to easily include more influencing factors because of the nature of the flexible structure of the SSNN. (Author/publisher)
Abstract