The sixth biennial ASAM International Conference was held in Munich, Germany, on December 4-5, 2024, and delved into the future of automotive engineering and how standards facilitate interoperability, data exchange and toolchain integration, thereby improving processes and providing value.
For this year’s program, ASAM (Association for Standardization of Automation and Measuring Systems) expanded the scope beyond ADAS/AD to address a wider range of ASAM standards, exploring topics ranging from early-stage development to real-time interaction during operation, encompassing critical aspects such as over-the-air (OTA) updates and diagnostics.
The House of Communication in Munich hosted the conference and provided a suitable backdrop for discussions, embodying the association’s commitment to innovation and collaboration. The venue reflected what ASAM CTO Benjamin Engel (left) wanted the company to represent, that standards are about communication, serving as the “glue” to seamlessly integrate components along the whole development pipeline and enabling the exchange of data and information. According to CEO Marius Dupuis, “We try to get everyone to talk to each other and exchange ideas. ASAM is pre-competitive. You can compete outside ASAM’s doors, but here it’s all about cooperation.”
Speakers included BMW, Porsche Engineering, 51Sim, Akkodis, Deontic, AVL, Tier IV and the Bavarian State Chancellery.
An exhibition area complemented the two-day conference, enabling companies to demonstrate and showcase their innovations. Exhibitors included Automotive Artificial Intelligence (AAI), Modelica Association, Deontic, dSPACE, Foretellix, Karakun, MathWorks, TrianGraphics and UKi Media & Events (organizers of ADAS & Autonomous Vehicle Technology Expo Europe , Automotive Testing Expo Europe and associated publications ADAS and Autonomous Vehicle International and Automotive Testing Technology International.
According to AAI’s general manager, Basit Khan, “At AAI, we believe that collaboration drives innovation. Sponsoring the ASAM International Conference reflects our commitment to fostering industry-wide partnerships and advancing the standards that shape the future of ADAS and automotive AI.”
Day 1 Conference highlights
Streams included ‘From simulation to real’, ‘The future of data recording’, ‘Beyond the road’ and ‘Verification & validation (V&V)’, with contributions from vendors 51Sim, Foretellix and Japanese deep-tech startup Tier IV. Discussions covered topics ranging from ‘Synthetic data generation with scenarios in ASAM OpenScenario XML’ and ‘AI-driven end-to-end synthetic data simulation and ASAM OpenX for autonomous driving’ to ‘ASAM OpenX standards for off-road applications’.
Dr René Grosspietsch from BMW Group delivered the first keynote, providing an OEM perspective and business case consideration for standards. Among other things, he assessed whether standards are creating additional value and if they make a difference to the customer. “L3 is a product liability for the OEM, you cannot blame the driver,” he explained. “You have to have a good argument if you deviate from state of the art. Standards help with argumentation. It’s not easy to change standards, for example to SOVD (service-oriented vehicle diagnostics); you must decide between alternatives. A whole ecosystem around the standards poses a value question: does the new standard outperform the legacy? The more advanced the organization, the more it takes care for added value.”
Tille Karoline Rupp (above), head of simulation at Porsche Engineering, spoke about using large language models (LLMs) to generate critical driving situations for virtual and hybrid ADAS/AD testing, particularly how to use GenAI to keep up with complexity and demand for automation.
“We want to increase speed of development, increase maturity and develop new features faster,” she said. “Critical driving situations should be generated in standardized formats and integrated into a common process for validation. We always start with natural language to generate the scenario. To include it in the process, we used ChatGPT to create the artefacts in standardized formats, such as ASAM OpenDrive. It was broadly successful; we got variation, lane change behavior showed variety in the speed at which a lane was changed.”
Exhibitors had the opportunity to present their current projects: Deontic showed fine-tuning models, created in collaboration with the University of Warwick, that use ChatGPT to capture more through natural language; Mitsubishi Heavy Industries demonstrated how to measure sensor degradation, supporting ASAM for ADAS (lane and fog test systems and cooperation ideas); AAI showed how to get very high fidelity maps; ASCS demonstrated trust, transparency and innovation by using an open-source project so all can collaborate and secure quality in a data space; and and AVL showed how to ensure and validate based on ASAM standards.
Hans-Dirk Walter at Karakun considered ‘Old standards for new applications – How ASAM ODS can boost efficiency in engineering, now and in the future’. “To accelerate engineering, its crucial to have data, the foundation of which are ASAM standards,” he explained. “Standards boost engineering; any AI based framework can work on data. A closed process chain allows engineers to focus on work.”
ASAM member companies including Mississippi State University discussed how they have been considering the application of ASAM OpenX standards to the modeling and simulation of ADAS and autonomous vehicles operations in off-road environments.
ASAM project leader Matthäus Lang said, “We have a concept project focusing on using OpenX standards in the off-road environment, which is going to be finalized in the summer of next year. On that basis, there will be a proposal for a next development project which will result in a possible standard or standardized approach in using OpenX standards in autonomous/off-road applications. We are focusing on possible diverse applications in agriculture, construction, mining and even drones which have lidars and cameras. We are looking at existing standards that can be used and adapted here. We are creating the standards, the building block used by the tool vendors and used according to the needs of their customer.”
Keynote from the Bavarian State Chancellery
Dr Florian Herrmann (above), head of the Bavarian State Chancellery and State Minister for Federal Affairs, presented a keynote regarding the current state and outlook of the automotive industry and how standards help to bridge the gap between regulation and industry. He also contributed to a panel discussion.
Dr Herrmann emphasized that Bavaria is a key hub for Germany’s automotive industry, with the automotive, commercial vehicle and supplier sectors contributing significantly to the state’s economic output. He stated that the Bavarian state government is committed to supporting this vital industry and fostering innovation across all technological pathways, including hydrogen, e-fuels and electric mobility. To ensure competitiveness and sustainable growth, it is crucial to create framework conditions that encourage technological advancements and provide flexibility for exploring the best solutions for future mobility.
A dinner reception at the end of the first day offered attendees an evening of food, drinks, conversation and live music.
Day 2 Conference highlights
Streams included ‘Software-defined vehicles (SDV)’, ‘Generative AI’ and ‘Automation and Virtualization’, with discussions including ‘Generative AI for creating simulation scenarios in ASAM OpenDrive and ASAM OpenScenario’, ‘Driving innovation in vehicle diagnostics: Adoption of ASAM’s SOVD standard for modern POSIX architectures’, and ‘Collaborating to standardize ASAM SOVD over-the-air’, among others.
ASAM CTO Benjamin Engel set the scene on day two of the conference, presenting a recap of the previous day’s topics and highlighting the interoperability of standards as well as the need to foster an environment where they work well together.
“Discussions have highlighted the growing relevance of such topics as AI and SDV. While there is very strong industry interest in these topics, for us it remains about the data,” he said. “For example, are the outputs and inputs of AI of sufficient quality? Data quality is a key topic that is just as relevant here and one that we are committed to supporting. Our quality checker has just been released and will be continued over the next few years. Speakers have also looked at simulation quality and credibility, its role in homologation and V&V. We will look at this over the next few years.
“Where should ASAM put its focus? There are new topics such as driver monitoring systems, with pressure from the regulatory side, but we should also look at it from the standards side. Data is the core competence of what ASAM and its standards are there for, to ensure the fast and efficient flow and exchange of data and at good quality, independently of these use cases.”
With demand for mobility growing, Andreas Richter (above) from Volkswagen Group encouraged engineers to apply learnings across domains, such as from automated driving to railways, thereby making systems more robust and saving money on maintaining infrastructure. Volkswagen is currently working on robotaxis as part of its Mobility as a Service alongside the German government.
Harsha Jakkanahalli Vishnukumar from Akkodis posited a novel approach for scenario extraction and generation in ASAM OpenScenario DSL using AI-core meta and ELM. An AI-based methodology was used to extract scenarios using a reinforcement learning approach and low-level and high-level fusion to look at extracted data. AI clustered the scenarios, analyzing coverage. Harsha explained that if a model is well trained, it can create a new scenario to increase coverage; the accuracy continually increases.
Yves Peirsman from Deontic discussed how the company is using GenAI to help test engineers simulate vehicle behavior and Steve Schwinke from Sibros looked at how flexible, scalable SOVD is important for enabling SDVs. The standardization of the data pipeline and its importance for SDVs was analyzed by Mark Singer from the eSync Alliance.
The ASAM International Conference will next take place in 2026. In the meantime, you can revisit this year’s event by looking through the online gallery and official video.