
Environmental perception of vehicles is essential to ensure safe automated movement in road traffic. One specific class of error patterns is particularly striking in this process: the "environmental hacks" in terms of manipulation of environmental perception. Environmental perception must reliably detect stop signs, for instance, but a relatively minor manipulation, not necessarily recognizable as such by the human eye, can result in misclassification. This misclassification can, in turn, result in a risk for people, so the system must be resistant to such manipulations. Since the minor soiling of traffic signs can already cause similar effects, safety considerations also cover the intended functionality. The question is how to deal with "environmental hacks" since they concern both security (manipulation) and safety (effect). In his presentation, Stefan Kriso shows how to address this problem with examples taken from practice.