Short-time OD matrix estimation for a complex junction using fuzzy-timed high-level petri nets

The OD matrix at a signalised junction estimated for traffic light cycle represents a crucial information for transportation systems. A new dynamic two-steps method is proposed to estimate such an OD matrix. First, a vehicle conservation law is build in a dynamical way for each traffic light cycle using fuzzy-timed high-level Petri nets (FTHN). It is represented by an under-determinate system of equations whose unknowns are the elements of the OD matrix. Second, a fuzzy linear programming approach is used to solve this system. Since the data used to estimate the matrix are imprecise, they are represented as fuzzy numbers. Experimental tests made with real data provided by video cameras installed on a junction show the benefits of using an FTHN tool for this problem.