Conference & training programme

“Anyone got a free board right now?” That question appears in almost every developer chat. The software build runs fully automatically, but the path to real hardware still leads through colleagues, calendars and a walk to the laboratory. This presentation shares Vector's experience turning globally distributed hardware into a shared platform that serves both automated CI/CT pipelines and everyday developer work. It examines what it takes to make hardware bookable and remotely controllable, integrate it into existing build pipelines, and close the organisational gap that often separates software teams from the people who own the test hardware. Attendees will learn how the Commit to Hardware concept connects modern CI/CT practices with physical product validation. Using real examples from Vector's journey, the presentation shows how treating hardware as a shared platform enables faster feedback, higher automation and more efficient use of globally distributed test resources.

Speed-to-market is critical for today's automotive ecosystem, yet modernising E/E architectures remains bottlenecked by legacy formats, fragmented scripts and proprietary toolchains. FLYNC tackles this challenge at its foundation by bringing a true Configuration-as-Code approach to system-level architectures. As an open-source description language and toolchain, FLYNC centralises vehicle topology, interface definitions, deployment and communication into a unified, single source of truth managed natively in Git. This presentation explains how FLYNC's schemas, validation tooling, and human- and AI-friendly syntax transform system integration. By replacing fragmented workflows with collaborative, code-centric practices, FLYNC enables modular ownership across OEMs, Tier-1s and Tier-2s while establishing the foundation for automated, agentic development loops. Ultimately, this open-source paradigm compresses integration and validation cycles, delivering the speed, scalability and agility required for the next generation of software-defined vehicles.

Modern vehicle architectures are moving from isolated ECUs and functional domains towards highly interconnected, zone-based networks. As a result, communication that appears stable when tested within a single device or domain may behave very differently once it shares network resources, timing dependencies, gateways and processing paths with the rest of the vehicle. Traditional device- and domain-centric validation therefore no longer provides sufficient assurance for pre-SDV and SDV architectures. Validation must evolve towards a system-level approach that observes Ethernet, CAN, LIN and other communication simultaneously and correlates all relevant data on a common time base. Only by reconstructing the interaction between distributed traffic, events and processing delays can engineers identify where congestion, timing shifts, dependencies or unexpected interference originate. The objective is no longer simply to prove that each component works—it is to prove that the complete vehicle works as one interconnected system under realistic load conditions.

Automotive Ethernet is the enabler for zonal networks. However, the high gigabit bandwidth used between high-performance computers and zonal controllers is not required in every area. Zones containing peripheral actuators, sensors and other components require much less bandwidth. For these applications, the 10BASE-T1S physical layer can be used. The presentation explains how a 10BASE-T1S network differs from other Automotive Ethernet technologies and how it supports bus-type architectures similar to CAN while retaining the benefits of Automotive Ethernet protocols. It also shows how 10BASE-T1S can replace CAN FD networks and support an all-Ethernet vehicle architecture. The audience will learn why Automotive Ethernet is important in zonal architectures, how software-defined vehicles are built on zonal networks, how network topologies are changing, how 10BASE-T1S operates, and the benefits of replacing legacy networks with Automotive Ethernet.

As automated-driving programmes expand across regions, vehicle platforms and software releases, validation teams face a growing challenge. The primary bottleneck is no longer simulation capacity alone, but the ability to systematically translate Operational Design Domain constraints, engineering requirements, safety objectives and regulatory expectations into complete, testable and traceable scenario coverage. As the ODD approaches the open world, the challenge is not simply producing more scenarios, but understanding the quality of the resulting coverage and proving it with defensible evidence. This work presents Deontic's prompt-driven, ODD-first approach to scenario engineering and validation-evidence generation. The framework ingests ODD definitions, technical requirements and compliance inputs, then uses agentic workflows to author, parameterise and manage simulation-ready scenario libraries. The generated scenarios cover road geometry, traffic infrastructure, actor behaviour, weather, visibility, work zones and long-tail interactions. It also produces requirement-aligned parameter sets, coverage plans, test matrices and provenance metadata, maintaining traceability from each ODD condition and requirement to scenario intent, simulation execution, coverage results and validation evidence. Deontic operates as an intelligence and API layer across the engineering toolchain, feeding simulators, test-management platforms, continuous-integration pipelines and evidence-management systems with structured scenarios and defensible test logic. The approach improves scalability, strengthens governance and improves release efficiency. IPG Automotive contributes as a validation-technology and ecosystem partner, enabling generated scenarios and test plans to be managed and executed through VIRTO and IPG CarMaker. The result is a shift from simulation enablement towards validation assurance, with evidence that remains traceable, reusable and ready for audit.

AI agents are increasingly able to interact with engineering tools through APIs and structured tool interfaces. Instead of manually navigating complex applications, engineers may describe a task in natural language and allow an agent to select and execute individual operations. This presentation examines what changes when an AI agent becomes an additional layer between the engineer and the tool. Using examples from automotive development environments, it explores practical opportunities such as easier tool access and workflow automation, as well as open questions around context, permissions, reproducibility and technical traceability. The aim is to provide a practical engineering perspective on the questions that should be considered when moving from experimental demonstrations towards repeatable, tool-supported workflows.

Teaching AI to Drive ADTF via MCP A lighthearted look at how we turned two decades of automotive tooling into a prompt-driven experience — and why your next ADAS session might be configured by a chatbot.

Validating perception algorithms in simulation requires more than visual fidelity; it demands a seamless, synchronised data pipeline that mirrors the physical vehicle architecture. This presentation details VERITAS's technical architecture. The session presents VERITAS, a production-ready virtual-validation ecosystem designed to bridge the gap between simulation and reality. It focuses on the rigorous Sim2Real correlation methodology used to validate NCAP scenarios, demonstrating that virtual environments can achieve the fidelity required for certification. Attendees will learn how to architect a validation pipeline that satisfies both engineering rigour and emerging requirements for credible digital assets.

The increasing automation and connectivity of modern vehicles mean that visual and acoustic vehicle signals—including warnings, display changes and driver-related feedback—play an increasingly central role in functional safety and system behaviour. Robust detection methods are needed to analyse these signals reliably and validate their correct timing. These methods can precisely capture visual features such as changes in shape and colour, as well as specific acoustic patterns. Industrial cameras, high-speed image-processing algorithms and microphones are used to analyse characteristic sound profiles. The resulting trigger and timing information enables an objective evaluation of safety-critical functions in complex scenarios, including driver-assistance systems and warning logic. Across the entire development process—from early simulations and hardware-in-the-loop setups to real-world vehicle testing—such detection systems provide a common and consistent database. This presentation provides an overview of the technological fundamentals of visual and acoustic detection and highlights typical challenges in evaluating time-critical signals.

This presentation introduces Child Presence Detection (CPD) and demonstrates how CPD systems can be tested and validated reliably.

As Automotive Ethernet becomes the communication backbone of modern Software-Defined Vehicles (SDVs), requirements for network monitoring, recording and analysis are increasing significantly. Beyond traditional diagnostic and development activities, performance analysis, TSN validation, cybersecurity testing, OTA-update investigations and distributed software debugging require reliable access to network communication throughout the vehicle lifecycle. Traffic can be captured using passive network TAPs, switch-port mirroring, media converters, transparent gateways, embedded monitoring or software-based endpoint capture. Each method offers different advantages and limitations in network transparency, timestamp accuracy, latency impact, scalability, implementation effort and the ability to manipulate traffic. This presentation provides a practical comparison of these architectures and their suitability for common automotive use cases. Capturing packets may be straightforward, but ensuring data is captured completely, accurately and with correct timing information is significantly more challenging. A dedicated section addresses preventing data loss during recording. High-bandwidth Gigabit Ethernet networks can exceed the capabilities of storage devices, memory buffers, mirror ports and processing units. Buffering mechanisms, trigger-based recording, circular buffers, continuous logging and capture-interface performance are examined, together with trade-offs between recording duration, capture rate and system cost. The presentation also examines measurement-system behaviour during vehicle sleep and wake-up transitions. ECU-startup analysis, wake-up validation and cybersecurity monitoring often depend on information exchanged immediately after network activation. Always-on capture, autonomous edge recording, synchronised startup and trigger-based approaches are compared. Accurate time correlation across multiple measurement points is essential for deterministic vehicle networks. Local hardware timestamping, centralised time servers, GPS-synchronised systems, IEEE 1588 Precision Time Protocol (PTP), Automotive Ethernet timing profiles and TSN synchronisation are evaluated for accuracy, complexity, scalability and suitability. Attendees will gain a technology-neutral framework for balancing measurement accuracy, network transparency, flexibility, synchronisation quality, implementation effort and cost when designing Automotive Ethernet monitoring and recording solutions. 30-minute agenda 1. Introduction and motivation: Automotive Ethernet in SDVs, capture challenges and common measurement pitfalls. 2. Capture methods: passive TAPs, SPAN, media converters, transparent gateways, and embedded or endpoint capture. 3. Time synchronisation: hardware and software timestamps, IEEE 1588 PTP, Automotive and TSN profiles, and multi-device correlation. 4. Preventing data loss: bottlenecks, storage and RAM, triggered recording, circular buffering and high-bandwidth strategies. 5. Sleep and wake-up: startup communication, always-on monitoring, autonomous recording and trigger strategies. 6. Decision matrix: use-case fit and practical trade-offs between accuracy, visibility, latency and flexibility.