Evidence-Based ITS [President´s Message]

Looking at the many methods and functions we have developed in Intelligent Transportation Systems (ITS) evidence-based ITS is an exception rather than the rule in our field. New car models are validated during the homologation process by various experiments. This process yields a high evidence level for many components of the vehicle. E.g., after a mechanical part has lasted various driving profiles the trust in its strength is deemed sufficient. However, for Intelligent Transportation Systems these experiments rarely provide significant evidence. The reason for this is that even for relatively simple systems like an Adaptive Cruise Control the space of possible real world data y includes driving situations and thus its dimensionality is too high to allow a representative coverage in experiments. How many successfully driven kilometers in which countries and by which drivers would be required to guarantee sufficient performance of such an advanced driver assistance system? Clearly, that example is solved as the driver is still kept in the driving loop and has to take on responsibility. But how can we ever gain sufficient evidence to trust a safety-relevant driver assistance function? It is still a long way before we may ever reach evidence for safe ITS but I am convinced that it worth to proceed on it.