Estimation Of Travel Time Distribution And Detection Of Incidents Based On Automatic Vehicle Classificatin

We study the problem of travel time estimation along a section of a freeway based on data derived from vehicle detectors at multiple locations. We pose the problem as one of pattern recognition. We derive algorithms that aim to recognize patterns which persist between the error-prone upstream detector samples and the error-prone downstream detector samples. We describe how these can allow us to estimate the distribution of the travel time between these detector locations. The most promising algorithm derived in this research is a dynamic programming based algorithm based on sequence matching techniques that would process the data as it arrives and could therefore provide continuously updated travel time estimates in real time. Keywords: Travel time estimation, pattern matching, sequence matching, dynamic programming.