Driver Behavior Interface Test Equipment – D-BITE and the Road-Sense Project

A.V. Thomas

Jaguar Cars Ltd, Coventry, United Kingdom

The EC co-funded RoadSense project comprising TNO, five automotive manufacturers and three university partners was established to provide an evaluation framework for the Human-Machine Interface of new Advanced Driver Assistance Systems (ADAS) during development, to enable the offering of innovative in-car ADAS for safety, comfort and support.

The accidentological background was established for human-vehicle interaction where an element of driver failure at the HMI was a contributory factor, so identifying critical scenarios for HMI design quality. Human Factors Indicators for driver behavior were then selected against a host of criteria, and objective and/or subjective metrics identified or created for these indicators. The engineering quantities associated with these metrics were identified and related to sensors for onboard use. Validation tests for the indicators and metrics were used to refine the selections using simulators, simple road layouts and tasks on test track facilities, through to full-scale driving studies. Advanced instruments were used such as TNO’s lane position sensor, Peripheral Detection Task systems, and various optical image analyzers for traffic hazards.

D-BITE open-architecture modular systems were then developed for four vehicle demonstrators fitted with a range of ADAS features, using Linux, Windows 98, 2000 or XP platforms and integrated up to seven video inputs, full vehicle onboard data, radar, infra-red, gaze trackers, the aforementioned specialist sensors, audio and keyboard monitoring and control, with capacity for more. These were all synchronized to allow real-time and posttest analysis, regardless of data channel time delays. Synchronizing and recording a D-BITE network in more than one vehicle for traffic interactions was also demonstrated. Further HF testing was conducted using these demonstrators. In all, over 170 drivers performed over 210 runs.

The output from D-BITE can be further processed in the laboratory to isolate events of interest, compare exact timing for sequences of incidents, stimuli and responses of the driver and vehicle, and make sophisticated comparisons of an individual across tasks, between drivers or between technologies. Naturally, the data can be further processed by third-party systems for detailed ergonomic usability studies, etc. Other applications in highly dynamic HMI environments may also be possible.

Paper presented at Measuring Behavior 2005 , 5^th International Conference on Methods and Techniques in Behavioral Research, 30 August - 2 September 2005, Wageningen, The Netherlands.

© 2005 Noldus Information Technology bv