Baselining behaviour: driving towards more realistic simulations
chapterposted on 16.02.2006 by Alex W. Stedmon, Chris Carter, Steven H. Bayer
Division of a book, which in a scholarly context usually treats a part of a larger subject in a stand-alone manner.
Automatic Speech Recognition (ASR) allows systems to be operated by speech input and may potentially improve the usability and safety of in-car systems. SPEECH-IDEAS is a LINK IST project investigating the use of ASR interfaces for in-vehicle systems. The success/application of in-car ASR relies on designing interfaces to match the expectations, preferences and abilities of various user groups. Driver workload (underload or overload) is a primary factor affecting the integration of in-car systems. Using multiple measures of workload, it is possible to assess relationships between actual task difficulty (objective measures), perceptions of task difficulty (subjective measures) and how individuals react to their perception of task difficulty (psychophysiological measures). This experiment sought to define a range of baseline driver workload factors in a simulator. Traffic behaviour (density, flow, speed changes, etc) and road layout/conditions (geometry, speed restrictions, fog, etc) were manipulated to assess the validity of different workload levels. Overall, the findings illustrated that increased workload affected driver performance and provide suitable baseline information for proceeding to assess the impact of in-car ASR on driving behaviour.