REACT

REACT is considered to be a strategically important project for the ASR department; the content is also a core topic in the Autonomous Driving Competence Center (CCAD). The overall goal of REACT is a systematic, safe and validatable approach to the development, training and use of digital reality to achieve safe and reliable action by autonomous systems, especially in critical situations. For this purpose, methods and concepts of machine learning - in particular deep learning and (deep) reinforcement learning (RL) - are used to learn low-dimensional submodels of the real world. In this way, a suitably wide range of existing critical situations should be recorded and identified so that they can be simulated in virtual space. Using this digital reality, the otherwise missing sensor data for critical situations can then be virtually synthesized and used to train the autonomous systems and vehicles for the safe and reliable handling of critical situations. The ultimate goal of the project is to systematically validate and constantly improve the capabilities of autonomous systems through continuous alignment with reality and the necessary adaptation of models.