lundi 10 avril 2017

Automated vehicles: How to assess inattention using physiological data?

Advances in technological and industrial research in the automotive sector suggest the introduction of Level 3 automated personal vehicles (an automated car requiring sporadic supervision by the driver, the latter having to be able to regain control of the vehicle if necessary) or even Level 5 (fully driverless automated vehicle) beyond 2020. Traffic policies are evolving to allow the deployment of autonomous vehicles, but these technological advances also ask questions in terms of road safety. Notably these questions concern human-machine interactions and the allocation of the attention during the use of such automated systems. Future vehicle should give the opportunity for the driver to do something else than driving. For these reasons, it will become necessary to supervise drivers in their activities to keep them in the loop when needed.
Although Level 3-4 automated vehicles prototypes are being developed rapidly, vehicles currently on the market are equipped with Level 2 automation systems. Taking back the control of such vehicles is still needed to ensure safety of driver and other road users. Research on automation has shown that the use of systems such as Advanced Driver Assistance Systems (ADAS) can lead to over-reliance, over-dependence or misuse, even if the system is not completely effective (De Waard, et al., 1999). According to Parasuraman and Riley (1997, p238-239), “excessive trust can lead to rely uncritically on automation without recognizing its limitations or fail to monitor the automation’s behavior”.
Rely on the system can lead to a passive fatigue in the sense that the driver should have few things to do while driving (May & Baldwin, 2009; Saxby, et al., 2013) and even an under activation that could hamper the mobilization of attentional resources (Young & Stanton, 2002). These two effects could be detrimental on the management of attention and de facto on the different steps needed while taking the driver back in the loop. A minimum state of vigilance is required to monitor the system optimally (Parasuraman, 1987). Inattention episodes are already observed when driving actual vehicles although drivers’ attentional resources should be entirely dedicated on the driving task. Indeed, epidemiological studies have already shown an increase in the risk of being involved in an accident while being responsible when the driver is inattentive either because of the presence of ruminations associated with emotional stress (Lagarde, et al., 2004) or because of mind wandering (Galera, et al., 2012). The proportion of accidents linked to inattention to driving being not negligible, it is necessary to address these new road safety issues by notably taking into account driver’s state during the use of automated vehicle.
The aim of this post doctoral contract is to work with several researchers of Lescot: 1) to specify physiological indicators (e.g., heart rate variability, respiratory rate, respiratory volume, electro-dermal activity) describing the driver’s attentional states as a function of the level of inattention; 2) to determine the thresholds of the physiological data indicating the changes of the driver’s attentional states; 3) to identify the technical possibilities to counteract the negative effects of inattention while driving so as to increase road safety.
Several experiments will be conducted in order to reach these three objectives. The project will be managed with several major industrial partners.
- Candidates to this position have a Ph.D. in cognitive science, cognitive neuroscience or related field, and research experience should follow the same lines. Evidence of top quality research on the above specified areas in the form of published papers in top conferences/journals and/or patents is therefore mandatory.
- Candidate will have experience in one or more of the following areas: neuroscience
background, electrophysiological measures, acquisition and analysis of EEG, ECG
EDA data, signal processing, programming in Matlab, interest in mental workload
assessment. Knowledge concerning data fusion would be valuable.
Contract duration: benefited position with funding for an initial period of one year, pending
satisfactory performance the position can be extended.
Salary: about 2400 € monthly gross salary; full time offer
Work environment: our team is based at Ifsttar-TS2-Lescot, Bron, France.
Application: Applicants are invited to respond no later than April 21st 2017 by submitting a
curriculum vitae including publication list, a one‐page summary of research interests and
expertise a brief description of research that might be undertaken, and names and contact
details of two referees to alexandra.fort[at]
Further details and informal enquires can be made by email to