Developing an iterative procedure to estimate origin-destination matrix based on two-point license plate tracking system

Origin-Destination (O-D) matrix, one of the most important elements in transportation planning, is usually estimated by various techniques such as mathematical modeling, statistical methods, and heuristic approaches. Since usage of electronic devices is rapidly increasing which helps decision-makers improve models' capabilities, an iterative procedure was proposed in this paper to estimate the O-D matrix through the detection of vehicles' license plates. The main concept is to track vehicles on the rst and last links equipped with plate camera over the shortest path from origins to destinations. A twostep procedure and mathematical models were developed to adjust assigning the passing trac to the network links by minimizing deviations between the observed and estimated truck trac volumes. The proposed procedure was elaborated using an illustrative example and it was validated using experimental road network that covered seven eastern provinces of Iran including 310 nodes, 400 two-way edges, and around 3600 origin and destination pairs. Results revealed that the proposed procedure could estimate O-D matrix when the network links were optimally located and equipped with road camera detection systems. In addition, similar to other heuristic approaches, the proposed procedure is sensitive to the number of iterations on the estimation accuracy.