Driver-vehicle interaction: Plenty of inhomogenous research works
The search term "driver-vehicle interaction study" results in 2,690 Google Scholar hits of research papers published in the past 5 years (2010-2015). This huge number clearly points out the problem researchers (particularly, new to this field) are exposed to: A lot of decisions regarding the setting, (e.g., lab/field, low-/high-fidelity simulator, within/between subjects, sample size, biased subject, learning effect, sensor technology, mobile hardware, synchronization issues, briefing, etc.) have to be made very early in the design phase, but there are no best practices, code of conduct, etc. available supporting them in the process of identifying the optimal solution to answer a specific research question. This workshop invites a) people active in the field to share their experiences in executing studies to measure driver behavior or vehicle conditions (driver-vehicle interactions), and b) young researchers to draft research questions, present their problems, and discuss possible solutions with the other participants.
Modeling of drivers and driver-vehicle interactions is still in its infancy...
In the latest years, advances in sensor technology have allowed to push the boundaries of transportation with a clear trend towards achieving fully intelligent transportation systems. Multiple efforts have been made to instrument vehicles with complimentary sensor arrays that utilize cameras, LIDAR, infrared and ultrasonic sensors to make sense of short and long-range environment context along the road. Some of these vehicles have already made it to the roads [4] and further trials keep industry excited [5]. A more limited approach has looked at instrumenting the interior of the cockpit with similar sensor arrays [6, 7]. Whether it is due to the big brother concern of the commercial application or to the lack of real-time value for in-vehicle information (IVI) systems, user modeling is still in its early stage. Our aim in this workshop is to present and review the available sensor technology, understand the limitations and present state-of-the-art methods, measures and modeling to achieve a complete understanding of the user behaviors in the vehicle. This exercise aims at having an expert review discussion *) from sensory data that can be utilized in non-autonomous vehicles for Advanced Driver Assistance Systems (ADAS), for enhancing adaptive in-vehicle user interfaces and for learning user preferences in a system where personalization is so important for both brand value and user comfort.