Application Story


Enabling automated vehicle verification and validation with real world data and ASAM standards

Member:  Siemens Digital Industries Software

Featured Standard:  ASAM OpenDRIVE®, ASAM OpenSCENARIO®


  • Millions of hours of recorded data are being generated during the development and testing of the next generation of autonomous vehicles. This data is stored and used for improving the development process, but without proper tooling it is not feasible to manually process it.
  • In order to achieve the safety standards required by both the authorities and the society to make the automated vehicle a reality, it is important to exploit every single bit of data available to improve neural networks, sensors, HMIs, comfort, …
  • The usage of that data makes sure that after real world testing, we can confront our decision and driving algorithms as well as our sensor fusion algorithms with the same situations found during real-world testing, as well as enabling variations over those encountered scenarios.
  • ASAM OpenDRIVE and ASAM OpenSCENARIO offer the foundation to generate simulation scenarios that can be easily stored, shared and test automated. In this way, not only are we able to virtually replay the scenarios encountered while testing but also virtually enhance them and add variations to challenge our application with edge cases derived from the initial recorded data.


At Siemens we support ASAM because we believe that standardized solutions benefit the whole industry and accelerate the development and adoption of safer, more sustainable and more accessible means of transport by both authorities as well as general public, leading to a conscientious society.

Joan Roca – Product Manager of Simcenter Prescan

Application Story

At Siemens we are developing the virtual Validation and Verification solutions that will enable the development and Certification of the automated vehicles of the future.


Customers face the ever-increasing challenge of validating their ADAS algorithms and meeting the authorities’ requirements when certifying a newly developed vehicle. Simulation is already playing an important role in the industry to solve this challenge and it will play an even more essential role in ensuring the safety of the newly developed self-driving vehicles.


But virtual testing doesn’t come without challenges. One of the focus areas of the auto makers nowadays is to make virtual testing as efficient as possible, by using edge test case detection. These edge cases need to be detected before the vehicle fleet is in service. As part of our Validation and Verification framework, we have set up a methodology to find such edge cases. We start with leveraging the data collected from vehicles which are on the road, that can be the test fleet or regular cars. This data is converted into Simulation scenarios, which are then used for determining critical scenarios by varying the scenario parameters. The Simulation results provide a feedback loop to improve and adapt the algorithms improving their robustness.


In order to satisfy this need, at Siemens we are enabling the pipeline that allows the import of real-world data into Simulation. First of all, the data is collected in the vehicle. In order to turn all this data into relevant simulations, we need to make sure that the recordings are enriched with as much Metadata as possible. The SCAPTOR solution offers in-vehicle tagging capabilities to ensure that the posterior automated tagging already has a basis to work with. Thanks to these two technologies, manual tagging is kept at a minimum, thus reducing costs on the whole pipeline. Subsequently, it is important to identify the relevant information, or the edge cases that we will challenge our application with. In order to guarantee efficiency and visibility, a fast and reliable searching tool is required, which will quickly identify and help visualize the stretches of data from our Database that are relevant to us and to our test case. Finally, this data will be further processed into ASAM’s OpenDRIVE and OpenSCENARIO open Simulation formats.


The key benefits of using ASAM OpenDRIVE and ASAM OpenSCENARIO are enhanced shareability, storage and test automation capabilities, making them the obvious choice for our framework. Compared with test data, the logical road and scenario description in ASAM OpenDRIVE and ASAM OpenSCENARIO is the key to easily apply variations of the measured traffic situation, and thus be able to cover a large parameter space of the adjacent situations. Once the real-world data has been transformed into ASAM OpenSCENARIO models, it can be replayed in Simcenter Prescan. In Prescan, users can choose to virtually modify the sensors, algorithms and even the vehicle characteristics to ensure that the challenging situation can be overcome by the next generation of vehicles. Also, with our framework users will be able to explore and vary these challenging situations to ensure that all concrete scenarios can be derived from a logical scenario, guaranteeing full coverage of a specific scene.


Similarly, databases with critical scenarios can be created and later used to check the compliance of newly developed vehicle models and algorithms. These databases can then be used in a test campaign to ensure that all requirements are satisfied first in a high-fidelity Simulation environment such as Simcenter Prescan, which provides validated sensor models, vehicle dynamics and asset quality and material definition. After the Simulation-based Validation and Certification, users can proceed to proving grounds for their real-life test campaign with their validated Simulation tests results. By using Simulation first, we can displace the risk and the edge test detection to a virtual environment, ensuring safety is prioritized early in the process, reducing development time and reducing also time between prototype iterations (rapid prototyping). Furthermore, ASAM OpenSCENARIO databases with edge cases can be used in a cloud computing environment to ensure continuous integration of newly developed algorithms, enabling nightly Verification in a CI/CD environment.


All in all, it is important to support all possible data streams when overcoming the multiple challenges that developing ADAS applications and autonomous vehicles entails. Even when using high fidelity Simulation, real data plays an important role, enhanced by our edge case detection and test coverage methodologies, and bringing real-life traffic situations into play. Combining real-life data with synthetically generated data (either fictional scenarios or extrapolated from real data via test automation) will be key for functional test coverage and will ensure that the autonomous vehicles are faced with realistic traffic situations before getting on the road, making them safer and more comfortable to operate.