A traffic simulation interacted approach for the estimation of dynamic origin-destination matrix

Conventional studies assumed that the transition matrix is known or at (east approximately known, which is unrealistic for a real world network. Due to the fact that traffic counts at a site, y, is easy to obtain and the O-D variable, x (path flow based in this research), is not easily observable, a Gaussian state space model is formulated to describe the relationships of x and y, observation equations, and the dynamics of x, state equations, with unknown transition matrix. Under the assumption of Gaussian noise terms in state space model, the distribution of random transition matrix F is derived. A robust solution algorithm combining Gibbs sampler and Kalman filter with arbitrary initial states is proposed to simultaneously estimate F and x, based on the latest available information. Moreover, a traffic simulator is incorporated with the solution process to provide travel time in determining the dynamic relationship between x and y and to make the whole solution procedure self-consistent. A randomly generated O-D data is used to verify the presented model and solution method. Preliminary results are generally satisfactory, showing that also in the unknown transition matrix case, significant estimates are achieved.