Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

AuthentiCap - a touchless vehicle authentication and personalization system

: Frank, Sebastian; Kuijper, Arjan


Braun, Andreas (Ed.); Wichert, Reiner (Ed.); Gomez, Antonio Maña (Ed.):
Ambient Intelligence. 13th European Conference, AmI 2017 : Malaga, Spain, April 26-28, 2017, Proceedings
Cham: Springer International Publishing, 2017 (Lecture Notes in Computer Science 10217)
ISBN: 978-3-319-56996-3 (Print)
ISBN: 978-3-319-56997-0 (Online)
ISBN: 3-319-56996-1
DOI: 10.1007/978-3-319-56997-0
European Conference on Ambient Intelligence (AmI) <13, 2017, Malaga>
Conference Paper
Fraunhofer IGD ()
Authentication; personalization; user interaction; automotive industries; Guiding Theme: Smart City; Research Area: Human computer interaction (HCI)

Current authentication systems in vehicles use portable keys or biometric and/or touch based inputs. They can be outwitted by stealing the keys or by copying the biometric information and analyzing the touch marks. This has to be inhibited, since vehicles are not only an expensive property, that would be lost in non-authenticated hands, but wrong permitted access also can unleash heavy machine power to inexperienced drivers or even people without a driver's license.
We present a system that authenticates drivers and unlocks personalization features without any portable keys or touching. Moreover, it is invisibly integrated into a vehicle structure, the steering wheel. In contrast to biometric authentication, the password pattern is adjustable and changeable. With the presented system, vehicle manufactures are able to install driver authentication systems without any visible design changes. The manufacturer thus provides more freedom and responsibility to the driver by giving him the option to choose his own unlock pattern. Still, the security is increased by avoiding common vulnerabilities like smudge attacks, the stealing of keys, or copying of biometric data. Our experiments show excellent recognition rates for multiple string patterns. A small user study shows that our system achieves 86% accuracy for inexperienced users, up to 96% for experienced ones. The users appreciated the easy of use.