Towards Parameter Identification of a Behavioral Model from a Virtual Reality Experiment

In this paper, we investigate the calibration of a mathematical model describing different behaviors occurring during natural, societal, or technological catastrophe. This was developed in collaboration with geographers and psychologists. To collect information on level stress, psychologists LPPL laboratory Nantes (France) led virtual reality experiments. These experiments consisted immersing individuals situation catastrophe measuring their electrocardiogram. From physical biological data collected, present methodology to calibrate behavioral model. First, theoretical analysis is carried out determine (i) if parameters can be uniquely estimated, (ii) minimal number discrete measurements required for estimation. Then, from these analyses, an estimation procedure performed at least have order magnitude parameters. Through work, will show simulations that proposed system makes it possible apprehend non observable human processes.