Human-in-the-Loop simulator studies as a tool to design and validate interactive communication between automated vehicles and vulnerable road users

Abstract

Nowadays, human-in-the-loop experiments are widely used in order to answer research questions, e.g. in the areas of human driver behavior or automated and connected driving. Often, human reactions to unusual situations or future technologies are in the focus of these studies. One of the main advantages is the ability to reproduce critical situations as often as required without endangering the subjects. The development of highly automated vehicles, especially in urban areas, requires a high degree of interaction with pedestrians, cyclists and other vulnerable road users (VRUs). Hence, more and more pedestrian and bicycle simulators are built and used in order to answer research questions. Since the possibility of evaluating driver assistance systems or their interaction with other road users also depends heavily on the quality of the simulation, a particular focus when setting up a simulator is on the quality of the underlying simulation modules and the degree of reality perceived by the test persons. Results obtained within this type of simulator studies can be used to design and evaluate communication concepts of automated vehicles via external HMI or the effect of their behaviour in difficult situations (e.g. way of right). Furthermore, the obtained data on pedestrian and cyclist behaviour can be used to model these road users, which is an important step on the way to a valid simulation-based testing of automated vehicles. The importance of a thorough design of VRU-simulators will be explained at the example of the new DLR VRU-Lab. The VRU simulator set-up is expected to give the simulator user a realistic impression in terms of the haptic, acoustic and visual presentation of the virtual environment. Thereby, it shall allow for a realistic walking and cycling behaviour. Hence, the simulator is expected to provide high immersion and low simulator sickness and thus should be well suitable for the exploration of a wide range of research questions concerning the interaction of pedestrians, cyclists and manual car drivers or automated vehicles.