With the new release 7 of its simulation environment for virtual driving tests, DYNA4, the tool manufacturer Vector simplifies the integration of mass simulations of virtual test drives into Continuous Testing (CT) pipelines.
DevOps workflows and Continuous Software Testing are now considered indispensable in the development of complex modern vehicle control functions. The use of these methods contributes significantly to ensuring compliance with the high quality and safety requirements of automotive software. The DYNA4 simulation tool supports customers throughout the entire development cycle with virtual test drives – from model-based controller design to hardware-in-the-loop ECU testing. The new DYNA4 Release 7 offers features to facilitate the integration of virtual test drives into CT pipelines.
The DYNA4 vehicle and scenario simulation models are based on freely adaptable models in Simulink. Lean simulation packages for Windows or Linux can be generated from these models. These can be exported as FMUs, for example, or can be distributed and operated in Docker containers. Both the generation and the execution of these DYNA4 Run Packages can be automated in CT pipelines. These mechanisms can then be triggered, for example, by a pull request from an ECU software developer or by a test engineer who has expanded the catalogue of ADAS test scenarios. Scaled execution of mass simulations with millions of virtual test kilometres is therefore easy to realise, for example with the help of a Kubernetes cluster.
Smart cameras and comprehensive ground truth data
The key to implementing ADAS/AD functions is sensing the environment. Consequently, sensor models are an essential model component of modern simulation tools for virtual test drives. Depending on the system boundaries of the ADAS/AD function to be tested, a suitable abstraction level for the sensor model must be selected. These range from idealised object lists to raw sensor data. DYNA4 R7 closes a gap in the middle processing level of camera sensors by making the generation of segmented images as output of an intelligent camera sensor now fully configurable. Various distortion filters can be parameterised and the segmentation can optionally be based on individual object instances or on object categories. The latter can be adjusted by the user to correspond to those categories and even the respective colours that are output by the real smart camera sensor. Further improvements concern the ground truth data as an essential element of the simulation.
Three DYNA4 product editions
The new Release 7 of DYNA4 introduces three product editions that are tailored to different use cases and operating environments and thus ideally support modern DevOps workflows. The “Server Edition” described above offers functions that support the fully automated and scaled execution of virtual test runs in CT pipelines. The “Desktop Edition” is designed for simulation experts who work on the creation of simulation content, conduct individual simulation studies or prepare the automated operation of test benches and mass simulations. The “Test Bench Edition” is geared towards the manual and automated operation of dedicated SIL or HIL test benches.