AVL and Rohde & Schwarz Announce Strategic Vehicle-in-the-Loop Collaboration

Rohde & Schwarz and AVL have intensified their collaboration with the integration of a Rohde & Schwarz radar test system into the AVL DRIVINGCUBE toolchain, creating new possibilities for testing radar enabled ADAS features and validating autonomous driving functions with a vehicle-in-the-loop test bed. Now complex driving scenarios can be generated and tested in a safe and reproducible environment.

The validation of complex advanced driver assistance systems (ADAS) and automated driving (AD) functions such as emergency braking, adaptive cruise control and highway chauffeur in all possible driving scenarios and environmental conditions requires huge amounts of functional and non-functional validation and can be risky when conducted on roads. Additionally, the new challenges of automated driving means that conducting conventional real-world tests is often not feasible on roads. Instead, test methods are virtualized using an X-in-the-loop (XIL) approach such as the AVL DRIVINGCUBE.

The AVL DRIVINGCUBE combines both simulation and real vehicles on a chassis dynamometer and powertrain testbed. It provides a new way to speed up the validation and approval process of ADAS and AD systems. The key concept of this solution is operating the real vehicle in a virtual environment, taking into consideration all parts of the "sense, plan, act" chain. Connecting the virtual environment to the real built-in sensors requires cutting-edge simulation technology. The conditions are completely different to those of an isolated sensor lab – particularly for test beds.

Following their work together on GNSS simulation, AVL and Rohde & Schwarz are now collaborating on radar target simulation on the AVL DRIVINGCUBE. The Rohde & Schwarz radar test system opens up a completely new range of possibilities for testing radar-enabled ADAS features and ensuring the safety of autonomous driving functions with vehicle-in-the-loop test beds. Innovative antenna array technology allows complex artificial objects to be generated for the radar sensors at variable distances and with variable radial velocity, size and azimuth – without physically moving antennas or devices.

A responsive real-time interface and seamless integration into AVL's virtual testing toolchain allow even challenging, complex and risky driving scenarios to be generated and tested. The system is fully future-proof thanks to a modular and scalable concept. This means that the same test bed can be used for tests of all types of vehicles with a different number of radar sensors and various sensor variants. The solution is used for the validation and application of adaptive cruise control (ACC), lane keeping assist (LKA) and other ADAS or AD functions. Furthermore, Euro NCAP scenarios can be validated in a reproducible and safe environment.