SAFIR Cluster 1: Simulation-based test systems for the pre-crash phase
Impulse Project 1: Mixed reality experimental environment for safety-critical functions in highly automated driving
Automated driving is a key technology of the mobility revolution. The introduction of highly or fully automated vehicle functions will significantly change the interaction between driver, vehicle and driving environment. However, the technical complexity and in particular the functional safety of these vehicles cannot be adequately captured and verified with classic endurance tests, as they have been carried out in the development phase to date. The impulse project starts at this point and aims to create a modern test environment for automated vehicles in urban traffic by means of a mix of simulation-based methods and real tests. A further problem of classical endurance tests arises from the fact that, especially in mixed traffic, different weather conditions, mental states of all traffic participants as well as sensor aging and failures have to be examined in a reproducible setting.The central research question of IP 1 is how traffic safety in urban areas will be changed by highly or fully automated driving. The aim of the project is the automatic generation of test cases and, based on this, the determination and evaluation of future traffic scenarios and their effects on traffic safety. The implementation of a continuous so-called "mixed reality test environment" will create a basis for valid and reproducible validation and validation of the high complexity of situations in urban areas with variable reality reference. For this purpose, the proportion between virtual and real test can be varied in several discrete steps in the four dimensions, driver models, sensor technology, environmental properties and vehicle characteristics of the test setup. With the resulting configurations, all test cases occurring in reality should be generated, evaluated with regard to their criticality and HAF/VAF functions should finally be secured in a reproducible manner.
Explanation Video 1: A key component in the introduction of automated driving is to achieve a high level of acceptance and trust in the technology among users. In driving simulation studies, corresponding parameters are collected and the effect of assisting systems (e.g., system transparency through augmentation) is investigated.
Publication: Philipp Wintersberger , Frederica Janotta , Jakob Peintner , Andreas Löcken und Andreas Riener. 2021. Evaluating feedback requirements for trust calibration in automated vehicles, it-Information Technology. De Gruyter Oldenbourg, Online 16 January 2021, DOI: https://doi.org/10.1515/itit-2020-0024
Explanation video 2: Simulation can never reflect reality, e.g. due to the lack of actual danger. Therefore, in comparative studies, laboratory tests are replicated with instrumented (automated) vehicles, including on the CARISSMA outdoor test track, the differences are analyzed and, in turn, conclusions are drawn for the simulation. The aim is to obtain results in simulations that are as close to reality as possible.
Publication: Anna-Katharina Frison, Philipp Wintersberger, Clemens Schartmüller, and Andreas Riener. 2019. The real T(h)OR: evaluation of emergency take-over on a test track. In Proceedings of the 11th International Conference on Automotive User Interfaces and Interactive Vehicular Applications: Adjunct Proceedings (AutomotiveUI '19). Association for Computing Machinery, New York, NY, USA, 478–482. DOI: https://doi.org/10.1145/3349263.3349602
>>> Read more