The development and use of advanced driver assistance systems (ADAS) and autonomous driving (AV) is rapidly growing. Especially autonomous driving- technologies are becoming increasingly complex. The validation of such systems poses significant challenges for OEMs, Tier1 as well as new entries in this market.
This is mainly due to the high number of potentially unknown and unsafe situations in which a vehicle’s behavior and residual risk in terms of safety must be assessed. In addition, it is not possible to fully test all these situations in real-world testing environments. Hypothetically, billions of test kilometers on public roads must be completed before a highly automated vehicle is ready for certification.
Achieving the highest possible safety standard for all kinds of vehicles is crucial. In 2022, for example, the General Safety Regulation (GSR) introduced new mandatory conditions for the implementation of certain ADAS features required in Europe. Market acceptance and further development of these innovative technologies are decisively determined by validated and reliable test concepts. In short, the higher the level of vehicle automation, the more intensive the required validation effort becomes.
In particular, virtual testing methods will be given a more important role in the validation process of automated vehicles. They enable to carry out the required high number of test scenarios in a wide variety of driving situations - thus reducing safety risks and the expense of real-road tests. Thereby massive simulation in the cloud, Model- and Software-in-the-Loop, Hardware- and Vehicle-in-the-Loop, as well as Driver-in-the-Loop will be important virtual testing environments for bringing these systems on the road.
Another big challenge, however, is the development of suitable tools for creating, describing, and managing proper test scenarios and cases. Moreover, according to safety assessment argumentation frameworks, appropriate methods are required for test coverage optimization and critical scenario evaluation. Thus, scenario-based verification and validation methods are applied in combination with virtual and real road environments
As a leading test system provider, we offer a scenario-based Verification and Validation (V&V) toolchain. The flexible, modular toolchain enables test preparation, execution, and evaluation for multiple test environments – from real-road, to proving ground and fully virtual environments. Our toolchain strategy allows the integration of different tools and content, such as traffic or environment simulation tools with proper scenario content. That method forms an important aspect for the virtual- and real-world validation of ADAS/AD-vehicles. It enables users to analyze KPIs, create test reports and to gain valuable information out of the vast amount of data generated through simulations, virtual testing, and real road testing.
Intelligent methods for generating test cases, such as via ontologies, ensure efficient testing with optimal test coverage. Our toolchain also uses statistics from real-road traffic to optimize test coverage in terms of real-world driving behavior. Pre-packaged scenarios and KPI evaluations support the testing of NCAP and GSR and ensure compliance with those regulations. We are partnering with key players in the area of safety argumentation and test methodology, for example, TNO.
Scenario Based Test Preparation
The development, testing and release of vehicles with ADAS/AD functions require safe operation and a great perceived safety. Thus, a huge number of predefined test plans (such as for NCAP or GSR), but also scenarios that go beyond them, must be tested. Furthermore, results have to be documented in a traceable way- in simulation as well as at the proving ground and on the road.
The AVL SCENIUS™ supports the complete process from scenario management and test case generation to test allocation and KPI- reporting. It helps to move from scenarios to specific test plans, which can be executed in simulation and other virutal as well as in real world testing environments. It integrates a systematic approach to define safety requirements and validation targets, and to generate the relevant test cases. Different test planning methods integrated into the toolchain allow intelligent optimization with respect to critical test coverage and the test environment
Test Execution Environments
Through the combination of physical and virtual tests you can validate your vehicles over large distances under limitless conditions to ensure system robustness and safety. The most efficient validation and approval will be done by those who will use the best combination of the different test environments. We provide testing solutions considering all relevant test execution environments:
- Real Road Testing: An accurate benchmark of the real world allows the measurement of the accuracy of the vehicle’s perception systems and therefore, engineer their safety. The AVL Dynamic Ground Truth System™ enables a faster, reliable, and precise ground truth collection over a whole vehicle fleet.
- Proving Ground Testing: The proving ground is a necessary part of the process for the verification of functional targets, active safety and legislative requirements. Our holistic and open toolchain consisting of our modules AVL Smart ADAS Analyzer™ and AVL SCENIUS™ support this complex workflow efficiently
- Vehicle-in-the-Loop (ViL)Testing: ViL-Testing combines both, simulation and a ready-to-drive vehicle, on a chassis dynamometer or powertrain testbed. Through the reproducible execution the AVL DRIVINGCUBE™ speeds up the validation and approval process of ADAS and AD functions in a safe test environment.
- Hardware-in-the-Loop (HiL) Testing: It is used in the early phase of development, where there is not yet a road-ready vehicle. The AVL-system-HIL-testbed enables tests for all ECUs using the SPA (Sense-Plan-Act) approach.
- Software-in-the-Loop (SiL) Testing: Testing the production software code in a virtual environment offers great performance and reproducibility. Fully automated SiL-Testing as part of a continuous integration cycle can be achieved by using the powerful programming interface of our virtual integration platform Model.CONNECT™.
- Model-in-the-Loop (MiL) Testing: In the MiL development process, the first virtual prototypes are basic models of the desired system functionality. Our integrated and open development platform Model.CONNECT™ allows an easy integration and simulation of controller and plant models from different tools.
KPI Evaluation and Data Management
During the development, verification, and validation process of automated driving functions, a huge volume of data is generated. Statistical evaluations of large volumes of representative real-world driving data are applied, to assess and improve the performance of ADAS/AD sensors and their algorithms. Millions of pieces of simulation-generated data ensure that the diversity of real-life driving is considered during development. The Data-Driven Development (DDD) approaches become a core enabler of the higher the degree of automation (SAE level 0 to 5) of a vehicle.
The AVL Big Data Management and Analytics Platform™ supports the entire workflow of the data driven development and helps to store and manage data efficiently and to analyze it automatically. This increases efficiency and accelerates ADAS/AD development.
AVL offers a comprehensive, scenario-based verification and validation toolchain to generate test scenarios, create the relevant test cases, and execute testing and homologation of assisted and automated driving functions in virtual and real road environments. Only through a combination of physical tests and simulations manufacturers will be able to validate their vehicles over such large distances and under limitless conditions.
- Thomas Guntschnig, Program Manager ADAS/AD Portfolio
- Seamless scenario-based testing toolchain, including simulation, vehicle-in-the-loop and real-world testing
- AI-Based methods for scenario extraction and tools for scenario creation and management
- Comprehensive methods for test case generation and optimizing test coverage, such as Ontology and real road statistics
- Scaleable big data solutions with build-in AI based algorithms to enable data driven development efficiently
- Up-to-date regulatory standard scenario-packages for NCAP, EU-GSR, UNECE ALKS
- Strong know-how in verification and validation processes, toolchains, and integration into customer environments
- Cooperations with best-in-breed partnerships with unique technologies such as radar and steering injections
- Experiences in model and entire toolchain validations to enable simulation and virtual homologation
- Strong engagement in regulatory and standardization bodies to accelerate a safe deployment of ADAS/AD systems
AVL, as a leading system provider, is connected with experts all over the world. By combining our know-how with other partners in the automotive and technology industry we are continuously strengthening our ADAS/AD Verification and Validation portfolio. Our evolving partner ecosystem involves key players in partnership simulation, technology, scenario content and methods and standards.
Years of Experience
Automotive Customers and Institutes
ADAS/AD Experts Worldwide
Industry and Research Partners
In this webinar, AVL and Alp.Lab will introduce an ADAS/AD testing and validation from real world data recording, scenario extraction to scenario-based test planning and execution in virtual environments.
Ready to combine the best of two worlds?The demand for ADAS/AD testing and validation in simulation and real world is increasing rapidly due to the growing number of ADAS/AD functions and vehicle variants.