Ground reaction forces GRF are influenced by different factors. Because of that, the experimental dynamic response of a lightweight footbridge under pedestrian crossing may differ from the one estimated by simulation. For the loading model to be realistic, the best option is use real recordings. Even though the pedestrian is following a metronome, each footprint is different. This provides a variability that is not taken into account in the loading models normally used. If a big enough extensive database were available, the differences could be statistically characterized and some loading sequences could be computationally generated. A procedure to generate virtual GRF is presented in this paper. Time domain and frequency domain estimators are used in order to compare experimental and virtual GRF histories. As results, some differences will be analyzed and the main conclusions will be drawn.
