AVL LIST GMBH

As vehicle automation advances, integrating automated vehicles into the existing transportation system is crucial, considering technical but also social factors. This chapter investigates two Austrian pilot sites, Graz and Pörtschach, by assessing user preferences through a novel “supertester” approach that included experiential elements as well as interviews, questionnaires and workshops. The supertester approach is a within-subjects empirical method in which the same group of individuals experiences various use cases. Employing this approach allowed a comparative analysis across diverse settings, use cases, vehicle types and user perspectives. The study underscores the critical role of fundamental safety functions and the relation between different vehicle types and corresponding expectations of passengers.

Automated and autonomous driving systems are increasingly integrated into modern vehicles to support the vision of safe, efficient, and comfortable transportation. Given the complexity of these systems, thorough testing and close monitoring of their behavior is inevitable to prevent hazardous situations and unexpected behaviors. This paper presents different approaches for testing ADAS/ADS, focusing on generating critical scenarios for virtual validation and monitoring approaches applied during operation. Specifically, we provide a comprehensive overview of our previous work on combinatorial and search-based testing methodologies, highlighting their application in generating robust test suites. Additionally, we summarize our work on intelligent monitoring approaches to detect operational issues. Our findings emphasize the necessity of advanced testing solutions and continuous monitoring to identify and mitigate potential failures, demonstrating their applicability in enhancing the safety and trustworthiness of ADAS/AD.

In this paper we propose an extension to the MagicGrid framework to support virtual prototyping for early system performance Validation & Verification (V&V). Model-Based Systems Engineering (MBSE) is at a maturity where V&V of system performance is expected to be automated for mature analysis. However, current practices do not adequately cover nor describe how this can be enabled in standard MBSE processes using SysML models as the baseline for the system descriptions and knowledge capture. Therefore we propose an extension of the industrially accepted MagicGrid framework to cover virtual V&V in a tool and process agnostic method, supporting practitioners to develop and use models in MBSE for this purpose without a specific vendor or method lock-in. The framework extension is discussed for each new cell in the grid, and we provide guidelines and best practice discussions for how V&V should be enabled for each cell. Specifically, we discuss simulating/analysing SysML directly or through (co-)simulation. Automotive development is used as a running use case.

In a polymer electrolyte membrane (PEM) fuel cell, the following degradation mechanisms are associated with the catalyst particles and their support: carbon support corrosion triggered by carbon and platinum oxidation, platinum dissolution with redeposition, and particle detachment with agglomeration. In this work, an electrochemical model for those degradation effects is presented as well as its coupling with a three‐dimensional computational fluid dynamics PEM fuel cell performance model. The overall model is used to calculate polarization curves and current density distributions of a PEM fuel cell in a fresh and aged state as well as the degradation process during an accelerated stress test with 30 000 voltage cycles. The obtained simulation results are compared to measurements on a three‐serpentine channel PEM fuel cell with an active area of 25 cm ² under various temperatures and humidities. The experimental data are obtained with a segmented test cell using respective degradation protocols and test conditions proposed by the United States Department of Energy. In addition to the temperature and humidity changes, the influence of geometry and material parameters on the degree of degradation and the resulting fuel cell performance is explored in detail.

Due to complexity increases in modern systems and the digitalization paradigm shift, industrial development requires the integration of new technologies and methods to keep product quality high while reducing time to market. One emerging paradigm in the Systems Engineering (SE) discipline is ModelBased methods and technologies, and correspondingly Model‐Based Systems Engineering (MBSE) is seeing increased adoption. With mature MBSE application, several benefits can be expected from the availability of models, even from the very early stages of development, enabling increased communication clarity, cross‐domain collaboration, traceability, and analysis. Notably, MBSE enables (Co‐)simulation even at the early stage of architecture/design by leveraging model‐based capabilities. Co‐simulation specifically enables a smooth and seamless integration of different models defined across layers of abstraction, for example, system logical architecture and system physical architecture. However, while MBSE is assisting with many aspects of development it is still a predominantly isolated set of activities throughout the development, especially on the left‐hand side of the traditional V‐model. In this work we discuss the status of Co‐simulation in industrial MBSE and list several existing challenges, then we propose a novel framework for implementing Co‐simulation and exemplify using a real scenario how we might address the observed challenges. The framework hinges on the newly proposed SSP standard and extends the currently industrially adopted FMI (version 3) standard through embedding the FMI file format using various scripts, demonstrated in the Python language for this paper. Finally, we propose a set of recommendations for future investigations to strengthen Co‐simulation in industrial MBSE.