Developing movement simulators for the study of behaviour

Science topics February 2016 InnovationTransportModelling and computer simulationHuman behaviourRoad safety

Stéphane Caro, Research Engineer - COSYS Department, LEPSIS Laboratory



In contrast to what the name suggests, driving simulators, and more generally motion simulators, do not strictly speaking simulate driving. Above all, they generate a virtual environment in which participants drive and move by using an interface. The real world is replaced by a synthetic environment obtained by simulation.

From the components of simulators…

In order a simulator to be valid, it must simulate the different elements of the real world, which include:

  • The operation of the vehicle that is driven by participants, including its engine1, its dynamics, as well as on-board devices such as driving aids and sensors2;
  • The behaviour of the other characters and elements in the scene such as the vehicles that make up the traffic3, pedestrians and cyclists whose movements must be simulated too4 ;
  • Sources of light and sound, as well as the propagation of light5 and sound6.

In order to replace the real world and give participants the impression they are moving, the results of the simulations must be expressed. To do this simulator hardware is made up of an interface equipped with sensors and sensory reproduction devices. A large amount of previous and ongoing research has focussed on the sensory

cueing and simulators’ immersion capabilities.


… to a process of development

The natural approach when developing a simulator is to reproduce the sensory stimulation produced by the real world as faithfully as possible. This has been done, for example in the bicycle simulator by reproducing the dynamics of the rear wheel of a real bicycle7. In this case, the torque participants must apply to the pedals and the inertia of the bicycle (i.e. the difficulty of starting and stopping) are the same as in real situations.

However, it is often technically or physically impossible to faithfully reproduce the characteristics of the real world. This would require, for example, the ability to reproduce the luminance of a sunny day on a screen8 and the acceleration of emergency braking. In order to overcome these difficulties, researchers have developed techniques that create illusions for participants, by reproducing only the relevant stimulations9,10. The development of these techniques relies on an understanding of the mechanisms of perception and tests conducted with participants.



Creating an “illusion” for participants by reproducing only certain accelerations (rights IFSTTAR)




In spite of the use of these algorithms, the lack or inevitable distortion of some sensory cues can interfere with the perception and control of movement in the virtual world. In some cases it is better to develop models which no longer represent physical reality but which are based on the expectations of the participants in terms of operation and sensory feedback. This is what was done for example for the control of balance on the motorcycle simulator11 or for steering wheel torque feedback as currently implemented on a number of IFSTTAR’s simulators.


And the choice of a suitable configuration

In order to select the most appropriate models and reproduction techniques, we need to know the benefits and drawbacks of each and, obviously, take account of how the simulators will be used.



Find out more ...


1See the article on eco-driving in this section and the IFSTTAR MODYVES internal research project (Dynamic and energetic models of conventional electric and hybrid vehicles for driving simulators) (B. Jeanneret & D. Ndiaye).

2Among other things, the "pro-SiVIC Recherche" simulator aims to provide physically realistic simulation of on-board sensors, communication devices and moving objects for the prototyping, testing and evaluating  of Advanced Driving Assistance Systems (ADAS) (Dominique Gruyer). More information available at

3Espié, S. & Auberlet, J.M. (2007). ARCHISIM: A behavioral multi-actors traffic simulation model for the study of a traffic system including ITS aspects. International Journal of ITS Research, 5, 7-16. Access the publication in HAL


4Research is under way to produce realistic animation of pedestrians and cyclists (Isabelle Aillerie).

5One example is research that attempts to simulate fog.

Dumont, E. (2002). Caractérisation, modélisation et simulation des effets visuels du brouillard pour l'usager de la route. Thèse de doctorat de l'université René Descartes, Paris.

Dumont, E., & Cavallo, V. (2004). Extended photometric model of fog effects on road vision. Transportation Research Record, 1862, 77-81.

6Research in progress on sound synthesis and propagation in collaboration with Genesis in the framework of the FUI Yellow project (Fabrice Vienne). More information available on the Movéo Competitive Cluster website.

7Presentation of the bicycle simulator project during the 2013 workshops of the Scientific and Technical Network of the Ministry of the Environment and Sustainable Development (MEDDE). Access the presentation of the workshop.

8The use of new technologies can make it easier  for the system to approach 1: 1 reproduction. See the 2014-2015 IFSTTAR internal research project that deals with high luminance dynamic screens (Céline Villa & Maud Ranchet).

9Algorithms that reproduce tones make it possible to represent wide ranges of luminance on screens with limited capacities. See for example:

Petit, J. (2010). Génération, visualisation et évaluation d’images HDR. Application à la simulation de conduite nocturne. Doctoral Thesis : Université Claude Bernard, Lyon.

Petit, J., Bremond, R., & Tom. A. (2013). Evaluation of tone mapping operators in night-time virtual worlds. Virtual Reality, 17, 253-262.

10Motion cueing algorithms allow motion bases with only limited capabilities to produce the essential characteristics of accelerations. See for example:

Mohellebi, H. (2005). Conception et réalisation de systèmes de restitution de mouvement et de retour haptique pour un simulateur de conduite à faible coût dédié à l’étude comportementale du conducteur. Thèse de doctorat de  l’université d’Evry-val d’Essonne.

11Patent No. FR 14 53836 pending

Caro S., Espié S., Lobjois R., Benedetto S., & Vienne F.  Méthode de calcul des composantes latérales (roulis et lacet) et du retour d’effort guidon pour simulateur de conduite moto.