Scalable and Physical Radar Sensor Simulation for Interacting Digital Twins
Jörn Thieling, Susanne Frese, Jurgen RoBmann
Abstract
Automation, artificial intelligence and sensor-enabled applications are part of everyday life comprising technical systems that utilize environment-perceiving sensors. Here, rough environments often require robust sensor technologies such as radar, coupled with application-specific software evaluating the measured data. The rising complexity of underlying sensor systems (hard- and software) and complex physical interactions with dynamic environments require development-related validations that use virtual methods primarily focusing on close-to-reality sensor simulation. To allow for a wide range of applications, we propose the use of reusable building blocks that generically model and mimic the physical structure from the real world. This includes the whole path between sensor and 3D environment with respect to sensor-intrinsic characteristics as well as environmental influences such as rain. At the same time, computational scalability of each building block allows physical simulations generating reliable sensor data even in real-time applications. As a result, Digital Twins of relevant components, systems and environments are no longer acting on their own but interacting within Virtual Testbeds allowing for virtual validation of radar-based applications with respect to the operational environment.