Using current sensing technology, a wealth of data on driving sessions ispotentially available through a combination of vehicle sensors and drivers'physiology sensors (heart rate, breathing rate, skin temperature, etc.). Ourhypothesis is that it should be possible to exploit the combination of timeseries produced by such multiple sensors during a driving session, in order to(i) learn models of normal driving behaviour, and (ii) use such models todetect important and potentially dangerous deviations from the norm inreal-time, and thus enable the generation of appropriate alerts. Crucially, webelieve that such models and interventions should and can be personalised andtailor-made for each individual driver. As an initial step towards this goal,in this paper we present techniques for assessing the impact of cognitivedistraction on drivers, based on simple time series analysis. We have testedour method on a rich dataset of driving sessions, carried out in a professionalsimulator, involving a panel of volunteer drivers. Each session included adifferent type of cognitive distraction, and resulted in multiple time seriesfrom a variety of on-board sensors as well as sensors worn by the driver.Crucially, each driver also recorded an initial session with no distractions.In our model, such initial session provides the baseline times series that makeit possible to quantitatively assess driver performance under distractionconditions.
展开▼
