Abstract
In recent years an increasing number of driver assistance systems have been introduced supporting driver in avoiding accidents by providing warnings and by intervening automatically. Since such systems are currently not yet considered in consumer ratings, e.g. Euro NCAP, the European funded research project AsPeCSS (Assess-ment methodologies for forward looking Integrated Pedestrian and further extension to Cyclists Safety Systems, www.aspecss-project.eu) aims at developing a test procedure in order to evaluate the benefit of systems for ac-tive pedestrian protection. Due to the complex interaction between driver, vehicle and environment, a compre-hensive test methodology cannot be applied without taking the driver behaviour into account. In this paper, the parameterization of a model of driver’s braking reaction with focus on driver – pedestrian interaction by means of a driving simulator study is described.
Starting from a definition of requirements resulting from the test procedure, driver models that are available from literature have been reviewed. A suitable model approach has been chosen and possible influencing factors have been identified. In order to systematically compile information on dependencies between model parameters and influencing factors, a dependency matrix has been created and has been filled with information from literature. As this information is sometimes contradictory due to different sources and study designs and for some depend-encies no information is available, a driving simulator study has been performed to investigate selected influenc-ing factors and to check hypotheses that have been derived based on the collected information. The subject test has been conducted in a dynamic driving simulator with four test groups, varying the criticality of the situation (by varying the time-to-collision at which the pedestrian appears) and the availability of a pedestrian warning signal. As dependent variables the model parameters reaction time, gain and maximum braking force have been chosen.
Significant influences of the independent variables on the dependent variables, i.e. the model parameters, have been checked by means of an ANOVA. The analysis shows that a warning signal has significant influence on the driver’s reaction time. The influence of the criticality of the situation on the reaction time has a trend towards significance. Furthermore, an evaluation of questionnaires has been performed. These results are only valid for the specific test scenario. However, conclusions on the qualitative effect in comparable situations might be drawn.
The results of the driving simulator study show that a warning signal has significant influence on the driver’s braking behaviour. Also, the influence of criticality shows a trend towards significance. Additional work will investigate this influence in more detail. Furthermore, it has been shown that most independent variables have only influence on the model parameter reaction time.
KEYWORDS – driving simulator, warning, driver model, reaction time, braking behaviour