抄録
Vehicular safety systems using ITS (Intelligent Transportation Systems) technologies have been proposed for supporting drivers' cognition and reducing traffic accidents. The vehicles equipped with such systems are named ASV (Advanced Safety Vehicle). To test the effectiveness of ASV technologies by simulations, the traffic simulator is needed to have the mechanism in which traffic accidents are caused by driver's cognitive error. In this research, we implemented the external world model to the traffic simulator "MATES" to deal with the cognitive errors in the simulation. MATES is the multi-agent-based traffic simulator which we have been developing. In our new model, a driver agent constructs its own storage for the information from the peripheral objects (signals and other cars), and makes its decision by following the stored information in every time step. Then cognitive errors are modeled as the storing errors since a driver cannot percept the non-stored information. We simulated the scenario in which traffic accidents occur at a single intersection and checked the reproducibility of this model.
