Executive summary

This research handled about microscopic traffic simulation. Traffic simulation is used for evaluating innovative infrastructures as well as driver support systems. In this study, the difference in working area between macro- meso- and microscopic traffic simulation is explained. Two microscopic models used at TNO are laid out. The first model is MIXIC, which is very reliable because of its detailed vehicle/driver behaviour. The second model is Paramics, which can use extensive road nets. MIXIC is primarily used for evaluating driver support systems. Disadvantage of MIXIC is that it only works on straight roads without intersections. Paramics has a very powerful potential to evaluate traffic on road networks. It however lacks well documented detailed driver/vehicle behaviour. To anticipate on this deficiency, the manufacturer provides a tool which enables redefinition of a part of the model.

At TNO the need exists for a detailed driver/vehicle model which can be used in complex road networks. Both models have to be studied to create a new model that can handle both requested features. MIXIC turned out to be complex, but is well designed, which enables easy access to its internal variables that are needed to control vehicles on road networks. Paramics is a rather closed program which only enables adaptations on specific points in the model. It is concluded, both models could be integrated. The following integration options are evaluated:


Synchronisation: Running MIXIC parallel with Paramics and reflect the vehicle positions in MIXIC to Paramics


Use MIXIC as a subsystem: Use the MIXIC model to determine the new position of each vehicle in Paramics


Total integration: Redefine the certain types of behaviour in Paramics using elements of the MIXIC model.

Option three has been discarded as impossible due to limits of the Paramics plug-in tool. The first two has been implemented. The first method did provide some valuable insights (for example when to release the vehicles) but was still bounded to a straight MIXIC road. The second integration option turned out to be working fine by using Paramics internals to store the MIXIC data and only the MIXIC driver/vehicle model for each vehicle in Paramics. However, the MIXIC model is not applicable to intersections, so Paramics takes over in these areas. A prototype version of this last solution is available.