The investigation of pedestrian dynamics is a young and lively field of research. Along with interesting self organization phenomena there are a multitude of applications, like the evaluation of escape routes in the context of crowd management or the optimization of pedestrian facilities for urban development. Our aim is the quantitative description of pedestrian dynamics by using microscopic models of self-driven particle systems. For model validation we cooperate with several universities on a systematical enhancement of the empirical data basis. Moreover we develop software to automatically extract accurate pedestrian trajectories from video recordings.