Effects of Driver Characteristics and Driver State on Predicting Turning Maneuvers in Urban Areas: Is There a Need for Individualized Parametrization?

摘要

In future, advanced driver assistance systems (ADAS) may be able to adapt to the needs of the driver, thus reducing the risk of information overload in complex traffic situations. One way of achieving this may include the use of predictive algorithms that anticipate the driver's intention to perform a certain traffic maneuver based on vehicle data, such as acceleration and deceleration parameters. In order to explore whether the predictive quality of such algorithms may be mitigated by individual driver-specific parameters such as driver characteristics (i.e. emotional driving [ED] and uncritical self-awareness [US]) as well as driver state (specifically stress), an empirical test-track study was conducted with N= 40 participants. The results indicate that maximum longitudinal and lateral acceleration vary significantly depending on driver characteristics. Moreover, analyses of the collected data suggest that incorporating psychological aspects into driver models can promote new insights into driving behavior.