The present research project aims to investigate if elderly drivers’ fitness to drive can be measured as good in a driving simulator as in real traffic. For this, a test course for driving simulators was developed. This course includes representative driving tasks of medium difficulty as well as scenarios being specifically sensitive to age-related impairments. For the assessment of subjects’ driving performance, trained observers register and classify driving errors online by means of a tablet application (S.A.F.E.). Based on the driving errors, a global assessment of driving performance is given. The course has been implemented in a high-fidelity-simulator as well as in a cost-effective compact simulator. The driving course developed in this project has been validated by comparison with a driving test in real traffic. This driving test is a standardized 60-minute course being structurally comparable with the driving course in the simulators. A 2x3 experimental plan with the dependent factor ‘method’ (three levels: real traffic vs. high-fidelity-simulator vs. compact simulator) and the group factor ‘age’ (25-50 years vs. >70 years) was realized. The data showed that the group of elderly drivers performed worse in various driving parameters compared to the control group. This is shown for both driving simulators as well as for the driving test in real traffic. Global ratings of fitness to drive and total number of driving errors during the driving tests in the simulation correlate with the global driving performance ratings during the driving test in real traffic up to r = .80. Findings from the high-fidelity-simulator and the compact simulator are highly correlating. In conclusion, the results of the validation study show that the driving simulator test developed in this project is suitable to assess the driving fitness of elderly drivers. This is true for both investigated configurations of driving simulation, i.e. high-fidelity as well as compact simulator. (Author/publisher)
Abstract