An extended field study on driver distraction and drowsiness was conducted in an instrumented vehicle, which was equipped with an eye tracker and other sensors. This report focuses on the findings concerning driver distraction. Seven participants used the car just as their own car during a period of one month each. During the baseline phase, which comprised of the first ten days of the trial the distraction warnings were deactivated. During the treatment phase, consisting of the remaining 20 days, the warnings were activated, meaning that the driver received a vibration in the seat whenever an algorithm called AttenD determined that the driver was distracted from the driving task. The participants’ subjective opinion about the warning systems was assessed with the help of three questionnaires.
The method is promising for driver distraction research, as it investigates naturalistic behaviour in a naturalistic setting. The employed eye tracker held up to the expectations, even though it is recommendable for future research to use more than two cameras. With the current setup, there was a tendency that tracking was lost just when driver distraction occurred. A robust data acquisition system is a requirement.
The main finding was that the drivers’ gaze behaviour was not influenced much by the distraction warnings. The drivers received distraction warnings at about the same frequency during the treatment phase as they would have during the baseline phase. This indicates that they did not avoid the warnings. Performance indicators like “percent road centre” and the newly developed percentage of glances within the “field relevant for driving” did not change from baseline to treatment phase. The standard deviation of gazes did not change, either. The average percentage of very long glances decreased slightly in the treatment phase, suggesting that the warning had an effect on the more extreme glance behaviour. There are also indications that the system helped prevent further extended glances away from the road immediately after a warning was issued.
The results from the questionnaire indicate that the drivers were satisfied with AttenD. Their expectations had been positive, and they indicated no disappointment. The drivers stated that they trusted the system, that the warnings were not experienced as disturbing, and that the system made them more aware of what they did while driving. Some drivers reported using their cell phones less while driving as a consequence of the warnings.
The analyses presented here are of a rather general nature, and more detailed analyses could provide new insights and a more differentiated picture of the usefulness of the driver distraction warning system. It is also important to investigate whether AttenD influenced driving behaviour like speed choice or steering variables.
A general problem with driver distraction research is the absence of a ground truth, which could be used as a benchmark, against which distraction detection algorithms could be compared and evaluated.
The similarity algorithm calculates how much two contents in the system are similar to one another. So far, similarity is calculated based on similarity of the project type, area of interest and user type. Generally, if two contents have more parameters in common they are more similar to each another. More information.
Copyright to all material on FEHRL Knowledge Centre are reserved. FEHRL Knowledge Centre's content (documents, reports, presentations, etc) can be cited, or excerpted in a sensible and proportionate manner, or e.g. included in non-commercial, on-line news digests, with proper reference (including a link) to FEHRL Knowledge Centre as the source, and to the author, by name, of any referenced post