Calibration of an eye tracking system for variable message signs validation

The overall objective of this study is the calibration of an eye tracking system for assessing the effectiveness and efficiency of different Variable Message Signs. The study has been carried out in the driving simulation laboratory of the Inter-Universities Research Centre for Road Safety (Italy). The first step of the research was the evaluation of the reliability of the eye tracking system. Twenty subjects have been tested in the laboratory. They were required to look at a point on a flat screen sitting in the simulator cabinet. The position of the point changes nine times on the screen each five seconds. The system tracks automatically the point focused by the subject. For each position the software takes more than hundred frames (frequency greater than 20 Hz). The positions focused by the subject are collected in a file that is post processed. The statistics of the residuals between point positions and focused positions show that the system is very accurate and the standard deviation is negligible considering the objective of the study. The second step of the research was the validation of ten different Variable Message Signs. The Signs are always composed with a pictogram and some words, symbols or numbers. The twenty subjects were required to look the signs for five seconds sitting in the driving simulator cabinet. During the test, the subjects were interviewed about the meaning and comprehension of each sign. A cross correlation between the interview outcomes and the eye tracks demonstrates that as the sign comprehension increases the eye fixes for about the same duration alternatively the pictogram and the words message. Otherwise the eye fixes for a longer time what it is unclear. In conclusion the study puts in light as the eye tracking system integrated in a driving simulator is a promising tool for in depth human factors analysis.