This paper presents a dynamic OD demand estimation model which uses turning movement counts as observations. Based on an iterative bi-level estimation framework, the upper-level problem is to minimize a weighted objectivefunction of deviation between simulated link flows and real-time link counts and the deviation between estimated time-dependent demand and an a-priori historical OD table, where the weighting value is determined by an interactive approach to obtain the best compromise solution. A case study has been performed on US29 network in Maryland to compare the estimated tables of this approach with the one obtained from the traditional method which uses only approach link volume counts. The application illustrates considerable benefits of using turning movements instead of approach volumes in matching observed counts.
Abstract