Cognitive Driving Data Visualization and Driving Style Transfer

This study proposes a visualization method for imaging cognitive driving data, which records data related to driving operations and drivers’ cognitive states. The visualization yielded an easy understanding of the characteristics of the driving operation and cognitive states. This allows intuitive comparison using images. Some differences between novice and experienced drivers can be visually embossed. Furthermore, the image style transfer algorithm using deep learning can be adopted for driving style transfer by representing the driving data as an image. Therefore, the image of an experienced driver can be used as a style image, and that of a novice driver can be modified by the style of an experienced driver. In this study, some trials of driving style transfer were attempted. Consequently, the driving style of a novice driver can be modified to clear and smooth operations like those of an experienced driver. As for the cognitive state, although a novice driver has always felt stressed up, the style of relaxed driving based on road conditions can be applied.