This paper presents the cognitive driver modeling effort during the project IMoST2 which deals with various driving maneuvers on the german "autobahn", e.g. free-flow, car-following, lane changes left / right. Starting with a wrap-up of the predecessor project IMoST1 and the flaws of the previous model, we derive the goals for the necessary extensions in IMoST2: improving the visual perception models, including peripheral and fovea vision; adding models for tactile and acoustic perception needed for the interaction with multimodal stimuli from modern driver assistance systems; improving the models mental representation of a traffic situation, based on a semantic network as memory component; models for distance and speed estimation plus a first implementation to create a spatial representation. We also present some selected empirical results to underpin the modeling assumptions. Finally, we outline the approach for model validation and show results for the lane change left maneuver.
10 / 2013
IMOST 2 Integrated Modeling for Safe Transportation 2