AUTOEV@L, multi-vehicle automation through the development, verification and technological validation of automated functions.

The Challenge

The European automotive sector is governed by strict testing and validation rules, which require a thorough assessment of all possible situations that an automated function will face in the real world: scenarios, hazards, unusual situations and difficult conditions. As a consequence, SAE-L3 test procedures and methodologies remain a key unresolved challenge for the validation of highly automated driving functions. The starting point is the approach where all possible scenarios to be tested are defined to validate functionality, an approach that is unfeasible and unaffordable in practice.

The main challenge of the AUTOEV@L project was to design and develop a methodology that can effectively and efficiently ensure the verification and validation of highly automated functions of all types of vehicles (automotive, rail and commercial vehicles), as well as the development of procedures for the automated testing of such advanced functions, in order to facilitate their rapid implementation and deployment and improve their effectiveness.

The Proposed Solution

The methodology proposed in AUTOEV@L revolves around the simulation and interoperability of the models that make up an automated system.

The AUTOEV@L development process is based on the model-based design (MBD) strategy with the aim of reducing development time and improving the quality of the system to be developed. To this end, models are created to represent the system, which are then subjected to a series of simulations and modifications to verify and test their correct operation.

Specifically, the MBD development cycle chooses to evaluate each of the phases, allowing errors to be detected as soon as they are introduced. In this case, verification is no longer considered as a final step but becomes a continuous process starting with design simulation, followed by real-time testing. Furthermore, the AUTOEV@L validation methodology follows the established automotive principles of scenario-based testing.

Another pillar of the methodology proposed in AUTOEV@L is simulation. It is not feasible to perform all tests in all domains, so it is necessary to combine physical and virtual tests in order to have a complete view of the scenarios that the advanced function will face.

to combine physical and virtual tests in order to have a complete view of the scenarios that the advanced function will face. In this way, a holistic view of the system's performance in different situations can be obtained, allowing for greater reliability and accuracy in system evaluation.

Evaluation with virtual environments allows the emulation of critical and meaningful situations by excluding any interaction that could lead to an incident. Thus, it is possible to perform testing of various systems in a controlled and safe manner, and with a lower investment in material costs. During the project, different simulation environments are adapted and developed to meet the virtual validation needs of the advanced functions developed. In addition, middlewares such as ROS2, RTMaps or ADTF, as well as standards (e.g. ASAM OpenLABEL) are integrated to achieve an interoperable validation and verification framework.

The Result

The methodology proposed and developed in AUTOEV@L was put into practice, applying it to different use cases in 3 different environments: automotive, railway and industrial. Some of these use cases include: platooning, parking assistance, comfort assessment, advanced localisation or robotic navigation.

The results obtained throughout the project were also published in different international conferences and journals, such as the IEEE International Conference on Intelligent Transportation Systems ITSC 2023.

AUTOEV@L had a multidisciplinary character, formed by a consortium of 8 partners of different typologies, including multifocalised R&D technological centres (Vicomtech, Tecnalia, CEIT, Ikerlan, BCAM), intermediate technological agents (AIC) and universities (EPS-MU and UPV/EHU).

AUTOEV@L is a project funded by the Basque Government (Elkartek 2021 programme) and led by Vicomtech.