Most traffic accidents are caused by human factors. The design of the driver environment has proven essential to facilitate safe driving. With the advent of new devices such as mobile telephones, GPS-navigation and similar systems, the workload on the driver has been even more complicated. There is an obvious need for tools supporting objective evaluation of such systems, in order to design more effective and simpler driver environments. Now video is the most used technique for capturing the drivers' visual behaviour. However, the analysis of these recordings is very time consuming and only give an estimate of where the visual attention is. An automated tool for analysing visual behaviour would minimise the post processing drastically and leave more time for understanding the data. In this thesis the development of a tool for visualising where the driver's attention is while driving the vehicle is described. This includes methods for playing back data stored on a hard drive, but also methods for joining data from multiple different sources. (Author/publisher) The report is available at the internet: http://www.ep.liu.se/exjobb/itn/2002/mt/019/exjobb.pdf.
Abstract