Images in mind – Design metaphor and method to classify driver distraction in critical situations

Abstract

The paper presents a driver model which classifies visual distraction based on the detection of atypical driving behavior. The model forwards the information to a driver adaptive collision mitigation system (CMS) and activates the acoustic warning earlier in case of distraction. Therefore the model requires the knowledge of the normal driving behavior. For that reason we introduce a design metaphor. We use the human memory and its ability to build up mental representations. Based on the idea to interpret multivariate time series as gray level images we adapted the concept of mental images to learn a situation based normal driving behavior. The model transfers the property of the long term memory to store, to interfere and to forget prototypes of mental images. We compare the stored prototypical image with the current image to obtain a distraction index. If the index reaches a certain threshold value the acoustic warning is presented.