May 14 – 16, 2024 | Germany

Auto[nom]Mobil
May 15 – 16, 2024
Here you see a recap of the program of the Auto[nom]Mobil 2023. As soon as the program of the 2024 conference is finalized, you will find it here.
Wednesday, May 24, 2023
Intro Prof. Klaus Kompaß - KKo4Safety, formerly BMW Group; Direktor & Professor Andre Seeck - BASt - German Federal Highway Research Institute |
Keynote: 'Please press Ctrl, Alt and Park to reboot your car' – Introduction to the Software defined Car (SdC) Dr.-Ing. Michael Fausten - Robert Bosch GmbH |
Automated Urban MobilityChair: Udo Steininger - TÜV SÜD Rail GmbH |
Wishes and Requirements of the Retail Industry of the Future for Autonomous Vehicles Frank Schmitz - REWE Digital GmbH
|
Autonomous On-Demand Mobility - Fata Morgana or the Revolution in Public Transport? Thorsten Möginger - Rhein-Main-Verkehrsverbund Servicegesellschaft mbH
|
Vay’s Teledrive-first Approach to Autonomous Driving Ole Hans - Vay Vay develops teledriving technology designed to remotely drive vehicles in urban areas. With that, Vay aims to launch a sustainable, affordable, door-to-door mobility service with remotely driven ("teledriven") cars. While starting with teledriving, autonomous functions will be gradually introduced over time as it is safe and permitted to do so. Vay was the first company to drive on European public roads without anyone in the car, no Safety Driver needed anymore. At the example of Vay’s teleoperations in Hamburg, you will learn about how driving without anybody in the vehicle in a complex urban environment is possible today. |
Artificial Intelligence and Automated DrivingChair: Dr.-Ing. Michael Fausten - Robert Bosch GmbH |
Applications and Challenges of AI in Autonomous Driving
Prof. Dr. Torsten Schön - Technical University of Applied Sciences Ingolstadt Groundbreaking AI systems such as ChatGPT, BARD or DALL-E(2) currently dominate the media and impressively demonstrate the capabilities of current AI technologies. At the same time, OEMs and high-tech companies have outdone themselves in recent years with announcements on autonomous vehicles. So far, however, one finds only a few, very low-key products on the market, predominantly with only Level-3 autonomy in sometimes very limited Operational Design Domains (OODs). In this talk, we will address the reasons for this reluctance and the challenges of current AI methods in autonomous driving. |
Safe AI for Autonomous Driving: Processes and Methods Jonas Schneider - e:fs TechHub GmbH The rapid development in the field of Artificial Intelligence have lead to big leaps forward in the development of autonomous and automated driving functions. To be able to release these AIs into series vehicles, the developing companies are facing the challenge to show the process-conformity of the development process and the safety of the AI function. Driven by laws, standards and industry-wide consortiums, a consensus on possible solutions to this challenge is building. This talk shows examples for such solutions from the consortium project KI-Absicherung and presents possible approaches on how the development of AIs in companies can be done in compliance with norms and processes. |
Hybrid AI-Methods: Approaches for AI in Safety Critical Applications Prof. Dr.-Ing. Michael Botsch - Technical University of Applied Sciences Ingolstadt Trained AI-algorithms are implementing a deterministic mapping. So, from this perspective “HOW” the system works is fully transparent and can be inspected. “WHY” a component is where it is and what purpose it has in the whole system is in general not transparent. Thus, the metaphor “black box” is often used and the validation of systems that contain AI-components cannot be realized using well established methods. Especially in the area of automated driving this is a central aspect. Hybrid methods, that combine AI with domain knowledge, are providing possibilities to use AI-approaches also in safety critical applications. The talk will address this topic and present some approaches for such hybrid methods. |
Developing Deep Neural Network Models for Behavior Prediction and Vehicle Interaction Dr. Gonca Gürsun - Robert Bosch GmbH Predicting the behavior of traffic participants is a challenging task in the context of autonomous driving. Producing reliable predictions that system functions can count on requires accurately modeling the interactions across traffic agents as well as their road environments. In our activity, we develop deep neural network based interaction models that capture the complex dynamic nature in such multi-agent settings. Our goal is to enable safe and comfortable natural driving via reliable, data-efficient and robust AI-based models. In this talk, we share our experiences in developing such AI models and present our first results. |
Regulation and its exemplary ImplementationChair: Prof. Klaus Kompaß - KKo4Safety, formerly BMW Group |
Framework for Regular Operation of Level 4 Vehicles in Germany Arne Zielonka - Bundesministerium für Digitales und Verkehr |
Validation Demands for the Automation of Commercial Vehicles in Different Environments Dr. Arno Hinsberger - ZF Friedrichshafen AG The presentation outlines technical challenges for the automation of major commercial vehicle applications. Specifically, the argumentation of applicable regulation and safety turns out to be one of the key enablers for the real-world operation of automated vehicles in different environments. Corresponding requirements are not only influencing the system-/software architecture (Design for Test), they also determine the efforts for Validation & Verification. Since argumentation of safety is fundamental for a product, related concepts are of strategical importance to design an economic solution. Before this background the main operation environments of commercial vehicles are analyzed. Naturally, the introduction of automated commercial vehicles for the operation in controlled environments is an important first step to take but even here potentially unexpected challenges need to be considered. In the light of still evolving regulation and homologation procedures, the presentation finally drafts an idea of how the application of different AD building blocks can be iteratively extended for increasingly complex operation environments. |
Wrap up Conference Day 1 Prof. Klaus Kompaß - KKo4Safety, formerly BMW Group |
Thursday, May 25, 2023
SimulationChair: Reinhard Böswirth - Veoneer Germany GmbH |
Stochastic Cognitive Model – a Driver Behavior Model for Safety Assessment of Automated Vehicles Alexandra Fries - BMW Group
|
Validation of a Neuro-Cognitive Driver Model in Safety-Critical Urban Scenarios Dr. Lukas Brostek - cogniBIT GmbH In order to assess the safety and compliance of automated and autonomous driving functions it is necessary to use simulation-based testing methods. However, this approach only provides valid test results if the validity of all simulation components can be proven for the specific use case and ODD to be tested. In my presentation, I will show that our ‘driveBOT’ driver model in urban scenarios does not differ significantly in a number of relevant statistics from real driving studies data, and that this can provide initial evidence of the validity of the simulation results. |
Realistic Simulation of Critical Situations in Road Traffic - Data-driven Simulation Optimization Florian Lüttner - Fraunhofer-Institut für Kurzzeitdynamik, Ernst-Mach-Institut The effort for assessing new automated driving functions up to autonomous vehicles is constantly increasing. Realistic traffic flow simulations are making it possible to design large parts of this protection virtually and to minimise this effort. For several years, Fraunhofer EMI has successfully taken on the challenge of mapping observable traffic validly and as realistically in traffic flow environments through data-driven optimisation. In doing so, interactions of all real-world traffic participants and resulting critical scenarios between them are taken into account. In this presentation, Fraunhofer EMI will give a brief overview of its approaches for data-driven optimisation of traffic flow simulations, the identification of critical scenarios in traffic flow data, and approaches for modelling non-motorised VRUs in a traffic flow simulation environment. |
Safe V2X as Enabler for Cooperative and Distributed Automated Driving Dr. Alexander Geraldy - Robert Bosch GmbH V2X communication prepares to be an important enabler for complex AD use cases on the road as well as in special environments, e.g., on logistic yards or in industrial plants. While safety in the vehicle as well as in the infrastructure is an established topic, V2X needs to evolve to directly support safety-critical distributed driving functions across diverse communication partners. The receiving vehicle must be able to decide if the data received is reliable enough and suitable to trigger a safety-critical maneuver (data qualification). Furthermore, the V2X messages must include information about the ability of the sender to support the receiver’s driving task (sender qualification). In this talk, we motivate the use of V2X communication for safety-critical AD driving tasks and present the challenges for V2X to support resilient and distributed AD systems in a safe way. Additionally, we will give an outlook to the German publicly funded project “ConnRAD” and a new ETSI work item addressing these safety challenges. |
Consistent Usage of OpenSCENARIO Across Different Simulation Environments Starting From openPASS Jan Dobberstein - Mercedes-Benz AG; Arun Das - BMW Group IT
|
The Human and AutomationChair: Dr.-Ing. Alexander Huesmann - BMW Group |
Driver Monitoring as an Essential Building Block for System Safety of Level 2 and 3 Automation Prof. Dr. Dietrich Manstetten - Robert Bosch GmbH It is widely recognized, that a well working driver monitoring is a key component for a safe-to-use level 2 and level 3 system. We will first lay the foundations by looking at standards and regulations in the field. We demonstrate the effectiveness of driver monitoring with results of some empirical studies, mostly with level 2 systems covering hands-on and hands-free driving situations. We give an overview of the state-of-the-art in driver monitoring technology, looking particularly at camera-based detection of visual attention. To close, we give an outlook to current research in cognitive distraction and situation awareness. |
Research on Kinetosis During (Highly) Automated Driving Andreas Hartmann - TU Berlin Studies show that as many as two-thirds of the population have experienced kinetosis (also known as motion sickness) while in a car at some point in their lives. With the increasing automation of vehicles, there is a higher likelihood of experiencing kinetosis due to intensified sensory conflicts. Therefor, research in this area is crucial to address this issue. In particular, subject studies are conducted, but due to the complexity of this phenomenon, high demands are placed on these studies. In this presentation, important influencing factors to consider (e.g. the influence of individual sensitivity), a kinetosis test vehicle, and first findings from subject studies on the effect of various driving dynamics on kinetosis in vehicles will be presented. |
Validation & Verification of Automated DrivingChair: Prof. Dr.-Ing. Thomas Helmer - Technical University of Applied Sciences Ingolstadt |
ISO TS 5083 Road vehicles — Safety for Automated Driving Systems — Design, Verification and Validation: Status Report and Schedule
Dr. Florian Raisch - BMW Group
|
Risk Management Core – Towards an Explicit Representation of Risk in Automated Driving Nayel Fabian Salem - Technische Universität Braunschweig While current automotive safety standards provide implicit guidance on how unreasonable risk can be avoided, manufacturers are required to specify risk acceptance criteria for automated driving systems (SAE~Level~3+). However, the ‘unreasonable’ level of risk of automated driving systems (SAE~Level~3+) is not yet concisely defined. Solely applying current safety standards to such novel systems could potentially not be sufficient for their acceptance. As risk is managed with implicit knowledge about safety measures in existing automotive standards, an explicit alignment with risk acceptance criteria is challenging. Hence, we propose an approach for an explicit representation and management of risk, which we call the Risk Management Core. |
V&V Activities of the next MB ADAS Generation. From Requirements and KPIs to a Release Recommendation Florian Dumas, Felix Erlewein - Mercedes-Benz AG
|
End of the Auto[nom]Mobil 2023 and Farewell Prof. Klaus Kompaß - KKo4Safety, formerly BMW Group |