奥迪

In contemporary virtual vehicle development based on systems engineering and the V-Model methodology, two primary approaches are used for simulating longitudinal dynamics: quasi-static backward simulation models and dynamic forward simulation models. Each approach offers distinct advantages and disadvantages. The main benefit of quasi-static backward simulation models is their computational efficiency. However, this type of model does not capture vehicle dynamics, which is a strength of forward simulation models. Additionally, forward simulation models inherently require longer computation times, which significantly increases the overall simulation effort, particularly when combined with optimization across multiple iterations. Nevertheless, in the context of requirement-based simulation, backward simulation models are generally more suitable than forward simulation models. Therefore, an obvious solution is to include dynamics in a backward simulation model.This publication presents a new time-efficient backward simulation approach that incorporates vehicle dynamics, with a particular focus on the acceleration of battery-electric vehicles (BEVs). The model is constructed using basic component representations, which can be replaced by more detailed models in future work, in line with systems engineering principles. For validation, the outputs of the proposed model are first compared against a quasi-static backward simulation model, a forward simulation reference model and vehicle measurement data. Secondly, a parameter analysis is conducted using a forward simulation model to analyze the chosen parameters of the backward simulation model.

For autonomous vehicles, safe navigation in complex environments depends on handling a broad range of diverse and rare driving scenarios. Simulation- and scenario-based testing have emerged as key approaches to development and validation of autonomous driving systems. Traditional scenario generation relies on rule-based systems, knowledge-driven models, and data-driven synthesis, often producing limited diversity and unrealistic safety-critical cases. With the emergence of foundation models, which represent a new generation of pre-trained, general-purpose AI models, developers can process heterogeneous inputs (e.g., natural language, sensor data, HD maps, and control actions), enabling the synthesis and interpretation of complex driving scenarios. In this paper, we conduct a survey about the application of foundation models for scenario generation and scenario analysis in autonomous driving (as of May 2025). Our survey presents a unified taxonomy that includes large language models, vision-language models, multimodal large language models, diffusion models, and world models for the generation and analysis of autonomous driving scenarios. In addition, we review the methodologies, open-source datasets, simulation platforms, and benchmark challenges, and we examine the evaluation metrics tailored explicitly to scenario generation and analysis. Finally, the survey concludes by highlighting the open challenges and research questions, and outlining promising future research directions. All reviewed papers are listed in a continuously maintained repository, which contains supplementary materials and is available at https://github.com/TUM-AVS/FM-for-Scenario-Generation-Analysis.

Automated vehicles (AVs) promise to enhance transportation safety and efficiency. However, ensuring their reliability in real-world conditions remains challenging, particularly due to rare and unexpected situations known as edge cases. While numerous approaches exist for detecting edge cases, a comprehensive survey reviewing these techniques is lacking. This paper bridges this gap by presenting a hierarchical review and systematic classification of edge case detection and assessment methodologies. Our classification is structured on two levels: first, by AV modules, including perception and trajectory-related (encompassing prediction, planning, and control) subsystems; and second, by underlying methodologies and theories guiding these techniques. Furthermore, we introduce "knowledge-driven" approaches, which complement data-driven methods by leveraging expert insights and domain knowledge to identify cases absent in training datasets. We then examine techniques and metrics for evaluating edge case detection methods, including detection performance, practical deployment (e.g., computational overhead), and domain-specific measures (e.g., crash rates and severity analysis). We conclude by highlighting key challenges for edge case detection, including data availability and quality issues, validation and interpretability limitations, the simulation-to-real gap, and computational constraints. The hierarchical classification and review of methods and assessment techniques in this survey enable modular and targeted testing frameworks by guiding the selection of detection methods for specific AV subsystems while considering methodological principles. It also supports practical testing by facilitating scenario generation in simulation and focused subsystem validation in the real world.

Where do the economic, environmental and social responsibilities of a car company begin and end? This is a question that often prompts discussion. For Audi, the focus of corporate responsibility is primarily on its products—and always considers the entire life cycle of a vehicle, from its production and use to the end of its life span: recycling. To lay the groundwork today for the future, the premium carmaker has to consider a wide range of issues: supplier management, protection of the environment, the research and development of innovative technologies like electrical mobility and synthetic fuels, and the challenge of recycling as much as possible of the materials used in cars. In doing so, Audi always considers how people will be living a few years from now, since creating an intelligent design for future individual mobility also depends on that. Audi bases all corporate decisions and activities on the sustainability of its products and processes. "We live responsibility" is the basic philosophy that runs like a common thread throughout the company.

Transportation systems are built up with the integration of numerous mechatronic systems that collectively enable the desired functionalities. The term software-defined vehicle (SDV) represents a paradigm shift in engineering. While the physical hardware components-such as sensors, actuators, and electronic control units-are modified in relatively slower development cycles due to manufacturing constraints and validation requirements, the software changes evolve at a higher iteration frequency. This accelerated software development enables faster deployment of new features to continuously optimize the performance and usability, or to remediate security vulnerabilities and perform maintenance updates. Various approaches have been published, adressing the rising system complexity with hierarchical decomposition, to gain insight, analyse, and maintain the system's performance characteristics from a system's approach to modelling. Various load situations, hereinafter referred to as use cases, can be analysed in detail. An open question is often how structured all use cases have been covered. For this purpose, expert knowledge or Monte Carlo simulation is often used. However, both approaches have the disadvantage that they have not built and structured the use cases through a clear methodology. In this paper, we present a modeling method to systematically describe the use cases for a transportation system using a structured modeling process. In this context, modeling already focuses on the fact that all relevant and possible use cases are considered during modeling, so that statements can also be made about completeness and significance in the analysis of the use cases of the system. We present the application and results of the method using an example of a use case model of an exemplary vehicle's themal energy system and a safety feature, as it could be part of autonomous mobility.