2017 | OriginalPaper | Buchkapitel
Assistance-on-demand – development of a speech-based, personalized left-turning assistant
verfasst von : Martin Heckmann, Heiko Wersing, Dennis Orth, Dorothea Kolossa
Erschienen in: Fahrerassistenzsysteme 2017
Verlag: Springer Fachmedien Wiesbaden
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
We recently developed the concept of “Assistance on Demand”. This describes an advanced driver assistance system (ADAS) which supports a driver in an inner city scenario only if she asks for assistance. A key element is the control of the ADAS via speech which allows the driver to flexibly formulate her requests for assistance while the situation develops. Our application scenario is turning left at unsignalized urban intersections. After the driver has activated the system via a speech command it monitors the right side traffic and informs about suitable gaps to enter the intersection, just like a co-driver would do.In a first user study together with the Würzburg Institute for Traffic Sciences we investigated this concept in a driving simulator. The results showed that drivers clearly preferred our speech-based system to a visual system implemented via a HUD and to driving manually without system support.We assume that drivers differ in what they perceive as a suitable gap to make the left turn. To test this hypothesis we have performed a second simulator study using CarMaker where 9 participants were turning left in crossing traffic from both sides. We deploy a probabilistic method to estimate the smallest accepted gap of each driver, so called critical gap. The results reveal that there is, as postulated, a significant inter-individual difference in the critical gap between the drivers. Next we investigate how well we can predict if a driver will accept a gap presented to him. We show that a prediction based on a driver’s personalized critical gap can achieve an accuracy of more than 90%.