![[Translate to English:]](/fileadmin/_processed_/8/5/csm_mfund_titelbild_49453a0fa2.png "[Translate to English:]")
The knowledge of the current traffic relations in road networks is an essential basis for advanced traffic management procedures. The capacities of the traffic networks can only be used optimally if the traffic demand as well as the travel times of the routes are known. With these inputs, the traffic flow on the network can be distributed efficiently. This knowledge represents the core information of almost all traffic management procedures, such as change direction signs, construction site management or light signal control in urban road networks. However, this essential information cannot yet be measured at a reasonable cost. For this reason, traffic management procedures usually use simple assumptions about traffic relations in order to achieve the best possible use of the existing traffic infrastructure.
It is now possible to determine this basic information of traffic management up-to-date and with high accuracy by means of a suitable fusion of mass data. Such data is the data obtained from detector loops and probe vehicles data or so-called Floating Car Data (FCD).
The aim of the project is to reconstruct vehicle trajectories with the help of advanced analysis methods and to refine them into reliable information about arbitrary traffic relations by model-based fusion with data from detector loops. As a result, the spatial course of the traffic flows between any source and destination zones can be determined for selectable periods in the past and for the current point in time.
In the frame of this project a scientific method will be applied to increase the quality oft he estimated traffic demand. The Gravity model is used to derive traffic flow on the intersection turns by using aggregated FCD. FCD is also used to estimate the traffic assignment in the network instead of using the theortical approaches such as the Stochastic User Equilibrium assignment. Again aggregated FCD is used to generate what is called the initial matrix, which is an important first step to launch the information minimisation model (IM). The IM model is used to estimate Orign-Destination matrix using the initial matrix, the estimated traffic assignment and traffic demand from detector loops.