Virtual Testbeds for the Development and Validation of Sensor-Based Systems Using Digital Twins

The presented application of simulation-based approaches (simulation-based systems engineering, digital twins, virtual prototyping) for the engineering of sensor-based systems has the potential to dramatically shorten development cycles and to simplify the generation and annotation of data for verification and validation purposes. This is especially true for application areas where the environment is hazardous or hard to reach (like spaceflight), and/or the situations that need to be evaluated occur rarely or are especially dangerous (like crashes or specific lighting conditions in automotive applications). This paper presents a novel approach to support the development of optical sensors and sensor data processing systems with a virtual testbed. We describe the concepts of our sensor and environment simulation, validation strategy, and annotation mechanisms. We evaluate our sensor simulation by comparison with real-world experiments and demonstrate the application of our approach in two real-world engineering processes (LiDAR-based spaceflight and camera-based driver assistance systems) derived with input from our industrial partners.