Research interests Pedestrian traffic I have been part of the PEDIGREE project on pedestrian traffic. We have realized several experiments in various settings, including bidirectional traffic in ring corridor, one-dimensional motion on a circle, binary or multiple interactions of pedestrians, oscillations at a bottleneck, etc. Data analysis With Asja Jelic, and in collaboration with the other teams of the PEDIGREE project, we have analysed the experimental data obtained for the one-dimensional flow of pedestrians. Microscopic features as well as macroscopic ones could be extracted. The relation between velocity and headway has been shown to exhibit two sharp unexpected transitions in the following behavior: while for large distances pedestrians move at their preferred velocity, they need to adapt to their predecessor when the distance decreases below 3m. This adaptation occurs on a time scale which suddenly shrinks when the distance becomes less than 1.1m. The relation between step size and velocity was shown to be very different from the one found in locomotion studies. Indeed, in the latter, only isolated pedestrians are considered, while in our case the lowest velocities result from steric constraints due to the presence of other people. Modeling From experimental data obtained for a one-dimensional pedestrian flow, in a collaboration between 3 teams of the PEDIGREE project, we found that the best correlation between the acceleration of a pedestrian on the one hand, and a function of the velocity difference / distance with its predecessor on the other hand, was obtained only if the following observation was taken into account: while a pedestrian can determine almost instantaneously the distance to its predecessor, it needs more time to evaluate the velocity difference. Based on this observation, we have proposed a ped-following model able to reproduce both the heterogeneities and the wave propagation that occurs in quasi-one-dimensional pedestrian flows (corridors). In collaboration with IMT, we have also proposed a macroscopic model for bi-directional flows in corridors. This model reproduces several features of the experiment realized in a ring corridor by the PEDIGREE project, qualitatively and quantitatively. In collaboration with IMT, we have more recently proposed some 2D pedestrian models derived from microscopic models. Chevron effect When two perpendicular pedestrian flows are crossing each other, some diagonal patterns were known to occur. In his PHd, J. Cividini (in collaboration with H. Hilhorst and C. Appert-Rolland) has shown, in the frame of a model based on exclusion processes, that this diagonal pattern was produced by a linear instability. Besides, J. Cividini has evidenced an unexpected non-linear effect, namely that the diagonals are not completely straight lines, but rather have the shape of chevrons. He has shown that indeed, the dynamics of the system could sustain a non linear mode with this chevron structure. This chevron instability could also be understood in terms of an effective interaction between pedestrians of the same type, mediated by the perpendicular flow. This result can also be seen as a contribution to the large field of effective interactions in soft matter. Here it is the first example of such an interaction in out of equilibrium systems, that can be solved exactly at the microsopic scale.