The scientific study of the behaviors of road and street users can be difficult or even impossible in real-world situations, where experimental conditions cannot be controled. Virtual reality offers an interesting alternative in several ways: it allows manipulating independent variables, comparing configurations ex ante, and safely studying dangerous situations. One must first establish that data collected in the simulated environment are consistent with what would be found in the real situation.

Immersive simulators are complex devices designed to interactively engage participants in the situation of interest by rendering the visual, cinetic and aural environment. Part of our research aims at evaluating and improving the physical, experiential, ethologic and/or psychologic validity of such devices, depending on the questions they are implemented to address. This involves studying the characteristics and configuration of the simulator as independent variables.

In order to study the attentional and sensorimotor processes underlying the task of locomotion, we focus on the restitution of visual and inertial information. We aim at identifying the dynamic model of the motion plateform of the simulator, using estimators and observers when sensors are unavailable. We also work at populating the road or urban environment with virtual travelers with scripted or autonomous behaviors, in order to study interactions among users.