Visual Design Support in Dynamic Probabilistic Networks for Driver Modelling

Understanding inference in probabilistic networks is an important point in the design phase. Their causal structure and locally defined parameters are intuitive to human experts. The global system induced by the local parameters can lead to results not intended by the human expert. Comprehending the behaviour of dynamic probabilistic networks (DPN) for tuning the model is a time consuming task. Therefore this paper introduces tools supporting the design phase. The application of these tools is shown by means of a DPN for human driver modelling.