Studying risk situations safely

Science topics February 2016 InnovationTransportModelling and computer simulationHuman behaviourRoad safety

Catherine Berthelon, Director of the LMATS2 Department

In the sphere of driving, precisely because they are “risky”, “risky situations” can only be investigated in a limited way under real driving conditions. Consequently, driving simulators are a possibility to be borne in mind.

Studying risk situations safely - Ifsttar - [Translate to English:] Image Pix5  Id: 15534543  Auteur: totolekoala

Identifying risky situations

We need to make a distinction between risky situations and the concept of “risk taking” which covers a set of complex phenomena which every individual tends to relate to their subjective perception of risk.

Risky situations correspond to situations that are likely to cause accidents and which are often difficult for drivers to deal with. Epidemiological studies
1 and detailed accident studies provide us with reliable indications about them and show they are the outcome of a combination of factors present in the human/vehicle/infrastructure system.

Human factors, which are frequently mentioned, such as speeds, driving under the influence of alcohol, drugs or medication, paying insufficient attention due to distraction (by a telephone, for example) and hypovigilance. The youngest (males) and oldest age groups are the most vulnerable, and riding a motorcycle or walking increase the severity of the outcome. Last, accident risk on trunk roads other than motorways is high.



Some studies made possible by driving simulators

Driving simulators have become indispensable tools for understanding and analysing interactions between the driver, the vehicle and road geometry. Their use has increased over time because of their many advantages: the reproducibility of the situations, control of parameters and absence of real risk as the driver operates in a virtual environment. They are thus particularly well-suited to studying situations that occur randomly or very rarely as well as ones that involve risk. For example, it is possible to simulate risky situations and present them to different groups of drivers. The aim of this research is to better identify the mechanisms responsible for a malfunction of the human-vehicle-environment system that have the potential to cause an accident.

A wide range of issues are investigated in connection with risky situations. These relate, for example, to the state of the driver in relation to the environment:

  • Vigilance i.e. “sleepiness” due to driving in excessively monotonous situations2;
  • Workload, or difficulties encountered in a complex environment3;
  • Distraction and inattention4 ;
  • Ingestion of psychoactive substances (alcohol, drugs of medication)5,6.



Studying risk situations safely - Ifsttar - A driving simulator (Photo credit IFSTTAR)

In spite of their benefits, simulators have some limitations (Photo credit IFSTTAR)




Nevertheless, in spite of their benefits they have some important limitations

The absence of a real risk, which is considered an advantage in driving simulation, nevertheless raises a number of questions when studying situations that are objectively risky. The issue of “behavioural validity” relates to the ability of a simulator to generate driving behaviours that are identical or similar to those which would be observed in a natural situation: the absence of risk can thus make it more difficult to interpret behaviours. Moreover, the “physical validation” of the simulator is linked to the match between its dynamics and that of vehicles. It therefore assumes that the dynamics of the simulator correspond to those of the real vehicle, but also that the way in which the road environment is simplified does not generate actions which differ from those in the real driving situation. Cognitive fidelity, which is related to knowledge acquisition, thus becomes a crucial factor in simulator validation.

The most effective way of carrying out this validation is to compare the driving behaviours obtained in the simulated situation with those obtained in the real situation. This is however difficult, and even impossible in the case of risky situations on ethical grounds, for example. So, from the safety standpoint, making a driver drink or testing a driver who has been deprived of sleep is not acceptable on the road while it is possible on a simulator. Simulators are therefore essential for studying risky situations, in particular in order to evaluate changes in behaviour due to the driver’s state. 

1 These are studies of populations that investigate the frequency and distribution of health-related problems over time and across different locations as well as the role of the factors that cause them. 



Find out more...


2Contardi, S., Pizza, F., Sancisi, E down., Mondini, S., Cirignotta, F. (2004). Reliability of a driving simulation task for evaluation of sleepiness. Brain Research Bulletin 63, 427–431.
3Paxion, J., Galy, E., Berthelon, C. (2015). Subjective overload depending on driving experience and situation complexity: which strategies faced with a pedestrian crossing? Applied Ergonomics, 51, 343-349.

4Fort, A., Gabaude, C.,  Lagarde, E., Lemercier, C., Cour, M., Maury, B. (2003). Impacts des inattentions sur la conduite automobile : approche multidisciplinaire. Compte-rendu de fin de projet ANR-09-VTT-04, 33 pp.

5Berthelon, C., Gineyt, G. (2014). Effects of alcohol on automated and controlled driving performance. Psychopharmacology, 231, 2087-2095.

6Bocca, M-L., Marie, S., Lelong-Boulouard, V., Bertran, F., Couque, C., Desfemmes, T., Berthelon, C., Amato, J-N., Moessinger, M., Paillet-Loilier, M., Coquerel, A., Denise, P. (2011). Zolpidem and zopiclone impair similarly monotonous driving performance after a single nighttime intake in aged subjects. Psychopharmacology, 214, 699-706.