Autor*innen mit den meisten Dokumenten
Auflistung nach:
Neueste Veröffentlichungen
- TextdokumentIntegration of GPU RTL IP into System-level Simulation through TLM: A Case Study with Vortex RISC-V GPU in the METASAT Platform Digital Twin(Software Engineering 2025 – Companion Proceedings, 2025) Lazzara, Lorenzo; Stazi, Giulia; Valerio, Valerio Di; Sinisi, Stefano; Ulisse, Alessandro; Bonet, Marc Solé; Wolf, Janis; Kosmidis, LeonidasIn the aerospace sector, the increasing system complexity of on-board processing for next-gen functionalities (e.g., autonomy) requires high-computational power. This typically leads to advanced cyber (SW and HW) platforms adopting multicores, GPUs, and Artificial Intelligence (AI) accelerators. In this setting, anticipating the integration verification activity earlier in the V development is crucial to save time and costs; and this can be enabled by migrating from real HW prototypes to virtual platforms where the on-board SW applications can run on top of them and fed by real operational data (namely, an as-designed digital twin of the system under test). This capability is among the objectives of the METASAT Horizon Europe project where a Model-Based Engineering approach is developed to build an as-designed digital twin of a high-performance platform coupling an open-source RISC-V GPU with a RISC-V based CPU Instruction Set Simulator (ISS) integrated with an AI accelerator. To build the virtual HW platform of such a digital twin, this paper presents an approach based on open standards (SystemC and TLM 2.0) for the integration of a GPU’s register transfer level (RTL) intellectual property (IP) with an HW platform model. Finally, such an approach has been applied to the METSAT case study.
- TextdokumentViability of Rust for Avionics Software Development -- Current status and way forward(Software Engineering 2025 – Companion Proceedings, 2025) Sommer, Jan; Rojo, Tamara Gutierrez; Lund, Andreas; Abdelmaksoud, Hany Ibrahim Erfan; Lüdtke, DanielThe trend towards more software functionality in less time is also true for avionics software. New development tools are needed to meet this trend, including the programming language. Rust provides language features such as extended memory safety and concurrency that are desired in avionics software. However, it must compete with an established software development ecosystem around C/C++. We propose requirements that a new programming language should meet to justify the effort of adoption in the aerospace industry and evaluate the current state of Rust accordingly. As part of the evaluation, we provide an example of how to add support for a new target to the Rust compiler for the RTEMS operating system used in the aerospace industry. Overall, Rust looks like a promising candidate for avionics software development and offers a gradual path of adoption, including qualification to aerospace standards.
- TextdokumentArguing machine learning assurance for certification(Software Engineering 2025 – Companion Proceedings, 2025) Varchev, Tihomir; Staudacher, Stephan; Daw, Zamira; Holloway, MichaelAviation certification traditionally relies on standards shaped by expert consensus on best practices. However, for emerging technologies like machine learning (ML), establishing these practices is difficult, particularly when in-flight testing is not feasible before certification, which can slow innovation. Argument-based certification offers a promising solution to bridge this gap by providing a structured and transparent framework for discussions between applicants and certification authorities. It allows authorities to compare proposed compliance methods across similar technologies from different applicants, helping to identify best practices and assess risks associated with new innovations. In this paper, we apply the Overarching Properties organizing principle to define and assess the means of compliance (MoC) for three publicly available ML aviation applications. This paper demonstrates how OPs can be used to support the demonstration of compliance. Although we observed some commonalities in the three arguments, the specificities of the technology and its application highlight the differences in the strategies used to demonstrate assurance.
- TextdokumentReconfigurable Computing Hypervisors: State-of-the-Art and Ways Ahead(Software Engineering 2025 – Companion Proceedings, 2025) Janson, Vincent; Nöldeke, Phillip; Kleine, Samuel; Durak, UmutIncreasing complexity in automation and autonomy features in aircraft, particularly with the introduction of Machine Learning (ML) based approaches is leading to a growing interest in highly parallel processing architectures, Graphical Processing Units (GPUs). However, GPUs come with challenges, such as certification, weight and thermal design. Another solution is the use of Commercial of the Shelf (COTS) System on Chips (SoCs), combining traditional Processing System (PS) with a Central Processing Unit (CPU) with a tightly coupled Programming Logic (PL) consisting of a Field Programmable Gate Array (FPGA). Through the use of a hypervisor within the PS, multiple partitioned software applications can be concurrently executed on a single computing platform, even if they have distinct criticality levels, while the PL lends itself as a dedicated and configurable, highly deterministic ML accelerator. However, depending on available logic gates within the PL, the complexity of the ML algorithm itself and the number of overall ML algorithms, the PL might not have enough resources to host all required accelerators at once. A potential solution is discussed in this paper: Reconfigurable Computing (RC) Hypervisors. In this work, classical hypervisors and RC hypervisors will be examined regarding their functionalities and key differences. Further, relevant publications in this field are compared with respect to their reconfiguration mechanism and functionality. Lastly, the limitations regarding potential aviation applications, both concerning performance and safety, are discussed. Based on the discussed topic, a new RC hypervisor concept is presented.
- TextdokumentAutomotive Software Engineering in an increasingly Data-Driven Automotive Sector(Software Engineering 2025 – Companion Proceedings, 2025) Denninger, Oliver; Axmann, Joachim K.; Kacianka, Severin; Westphal, BerndAutomotive trends such as power-train electrification, personalization, connectivity, and automated driving are not well supported by the classical approach to hardware/software architectures that centre around numerous, dedicated electronic control units (ECUs) where software is delivered as part of the ECU and it and its environment does not change much after vehicle assembly. Similarly, current electronic architectures and vehicles do not exploit data-driven software development practices and do not have the capability to make use of unprecedented amounts of data on the vehicle, but also its environment and the Internet. These trends ask for a data-driven approach where the development, production, and operation data of automotive software feed back into continuous correction, improvement, and personalization. In this paper, we report findings from the Transformation Hub Automotive Software Engineering (TASTE) with two years of intensive discussions and workshops with a wide range of companies regarding the challenges facing the German based automotive industry in general, as well as individual companies from Original Equipment Manufacturers (OEM) to different suppliers (TIER-n). We discuss how previously different approaches need to be integrated into new software-centr
- TextdokumentAutomotive Security Engineering: A Demonstration of an Integrated Approach to EAST-ADL and Security Modeling(Software Engineering 2025 – Companion Proceedings, 2025) Fischer, Alexander; Kolagari, Ramin TavakoliThe automotive industry’s increasing reliance on software innovations has introduced significant security challenges. With connected vehicles and semi-autonomous systems becoming commonplace, automotive cyberattacks are frequently covered in the media. Public awareness is growing that cybersecurity plays a crucial role, akin to the life-saving innovations of the 1970s, such as mandatory seatbelts and speed limits. It is evident that security must be integrated throughout the entire development lifecycle and operation of automotive software systems, with every stakeholder considering security aspects. However, this is complicated by the vast and often complex nature of the security landscape and the limited accessibility of essential security frameworks like MITRE ATT&CK, which are primarily oriented towards implementation rather than serving as an entry point for comprehensive security analysis. In response to these challenges, this paper presents the Security Abstraction Model (SAM), a metamodel-based approach designed to address security concerns across all stages of automotive software development, focusing on the early phases and on understandability among all stakeholders. SAM offers a structured framework for defining, analyzing, and implementing security requirements within automotive systems, leveraging established metamodeling techniques. This paper demonstrates how SAM can be integrated with EAST-ADL to create a systematic approach for securing complex automotive software architectures. In order to illustrate its applicability, an integrated demonstration is developed, showcasing the combined use of EAST-ADL and security modeling.
- TextdokumentWorkshop Summary(Software Engineering 2025 – Companion Proceedings, 2025) Schweiger, Andreas; Durak, Umut; Reich, Marina; Annighoefer, BjoernSystems and software engineering in aerospace is subject to special challenges. For covering these the AvioSE’25 workshop connects academia, industry, and certification authorities through selected scientific presentations, keynote talks, and a panel discussion.
- TextdokumentResearch Knowledge Graphs for Sustainable Literature Reviews in Software Engineering and Beyond(Software Engineering 2025 – Companion Proceedings, 2025) Karras, OliverIn this extended abstract, we present the ongoing work of several researchers on using research knowledge graphs (RKGs) for sustainable literature reviews (LRs) in software engineering (SE) and beyond. SE and academia face a significant increase in secondary studies and, in particular, LRs due to the ever-increasing number of publications. Most LRs are not sustainable as they do not build on previous ones due to the unavailability of extracted and analyzed data. Our ongoing work aims to address the challenge of conducting sustainable LRs by utilizing RKGs to ensure their data is FAIR, open, and long-term available. We developed and still refine an approach for using the Open Research Knowledge Graph (ORKG), one representative RKG, as a technical infrastructure for sustainable LRs, based on the KG-EmpiRE use case, which organizes empirical research in requirements engineering. The KG-EmpiRE use case demonstrates the applicability of our approach for conducting sustainable LRs by building, publishing, maintaining, updating, and analyzing a knowledge graph of empirical research in requirements engineering in the ORKG. The use of ORKG and RKGs in general can significantly enhance the sustainability of LRs, promote collaboration and updating of LRs, and ensure the quality, reliability, and timeliness of their research results through improved archiving, retrieval, replication, and (re-)use.
- TextdokumentRequirements Classification for Requirements Reuse(Software Engineering 2025 – Companion Proceedings, 2025) Märdian, JuliaIn various domains, standards are used to ensure a high level of product quality. During standard tailoring, requirements from the applicable standards are specialized and integrated into the project. The requirement type influences the way the standard requirement interacts with project requirements. Yet, manual classification of large existing standards is time-consuming. This thesis presents a machine learning pipeline to compare four algorithms for this task: k-Nearest Neighbor (kNN), Support Vector Machine (SVM), Logistic Regression (LR), Multinomial Naive Bayes (MNB), as well as an ensemble model combining all four. The models are trained and tested with 466 requirements from the European Cooperation for Space Standardization (ECSS). SVM and LR achieve the best results with F1 scores around 0.85. The integration of term contexts could potentially further increase the prediction accuracy. Yet, the improvement for our dataset is insignificant.
- TextdokumentError categorization in novice code(Software Engineering 2025 – Companion Proceedings, 2025) Just, NadjaIn a longitudinal study accompanying an advanced programming course, we manually analyzed 710 programming errors and categorized them based on an established framework of programming errors. Students initially make few syntax errors, but numerous semantic and logical errors. Over time, semantic errors almost completely vanish. The number of logical errors also decreases, but they never vanish completely. This suggests that students quickly grasp the programming language, but take more time to understand concepts and express them in a programming language.