Task 4: local origin-destination matrices and travel times estimation from magnetic identification
The aim of Task 4 is to estimate the mean travel times and the dynamic origin-destination matrices on the highway in the south of Grenoble. The magnetometer signals depend on the speed and position of the vehicles. Thus, the first step consists in analyzing and correcting the signal distortion. The principle of vehicle recognition using magnetometer arrays is based upon the comparison of signature features collected on two successive magnetometer arrays. These recognition methods are called “re-identification methods”, associated with “magnetometer arrays” and allow us an anonymous identifying and re-identifying of the vehicles.
Consequently, the estimation of individual travel time or origin-destination matrix can be achieved. The re-identification algorithm proposed in this project uses a traffic model. The proposed method is tested on a database carried out in Task 3. The individual travel times are filtered, and allow calculating the mean travel times. The dynamic origin-destination matrices of “Rondeau” are estimated for several periods to define.
To validate the proposed method, the mean travel times and the dynamic origin-destination matrices estimated from magnetometers will be compared to other video measurements, carried out in this purpose