Resources |
◎ COVESA Events |
Join/Sign Up |
◎ Join COVESA |
COVESA All Member Meeting ~ April 22-23, 2026
Day 1, Apr 22, 2026
Start Time | End Time | Title | Type | Description | Presentation | Speakers |
|---|---|---|---|---|---|---|
9:00 | 9:25 | COVESA Member Keynote | Business | Overview of Member Survey results and JASPAR MoU announcement |
| Steve Crumb, Executive Director, COVESA, Kazuo Tsobouchi, Chairperson of the API Standardization WG at JASPAR. Honda |
9:30 | 9:55 | API Standardization in JASPAR for SDV | Business | Presentation of JASPAR’s work on standardizing APIs using COVESA VSS |
| Yasushi Ando, Vice-Chairperson of the API Standardization WG at JASPAR, Toyota |
9:30 | 9:55 | Data Expert Group - Community Update | Technical | Data Expert Group Status Update
|
| Erik Jaegervall (Bosch), Adnan Bekan (BMW), Stephen Lawrence (Renesas), Daniel Alverez (BMW), Chaitanya Podalakuru (Ford), Ulf Bjorkengren (Ford) |
9:30 | 9:55 | AOSP App Framework Standardization Expert Group - Community Update | Technical | Expert Group Status Update
|
| Richard Fernandes (GM), Sabine Hofschen (BMW), Jose Freitas |
10:00 | 10:25 | Progress in SDV: A Global Comparison | Business | The 2026 Omdia SDV survey, in collaboration with Sonatus, builds on the 2025 SDV survey presented at the COVESA All Member Meeting in Berlin, offering a deeper exploration of the evolving landscape of software-defined vehicles (SDVs). |
| Maite Bezerra, Sr. Principal Analyst, Omdia; John Heinlein, CMO, Sonatus |
10:00 | 10:25 | Introduction to VSS: A Deeper Dive | Technical | Taking the next steps with your understanding of VSS. Learn from the pros as they walk through some how-tos. Get your questions answered and learn how to get involved. Description: Taking the next steps with your understanding of VSS. Learn from the pros as they walk through some how-tos. Get your questions answered and learn how to get involved. |
| Erik Jaegervall (Bosch) |
10:00 | 11:00 | AOSP App Framework Standardization Expert Group - Working Session OEM VHAL | Technical | As vehicles become software‑defined, OEMs might want to expose while at the the same time fully control reliable, consistent vehicle data to insurance partners, fleet operators, mobility services, ADAS analytics, anfd in‑car digital applications. Today, an OEM can offer either all vehicle data via OEM signature or close none at all. Today, every OEM solves this differently — resulting in fragmentation, high integration costs, and slow time‑to‑market. An OEM AOSP‑based VHAL offers a harmonized and scalable way to publish core vehicle data across models and regions using Android’s native Vehicle HAL framework.
In this workshop, we will walk through a hands‑on AOSP VHAL showcase using an emulator, demonstrate how standardized data flow and websockets enable partner integration, and collaboratively identify high‑value use cases. |
| Jan Kubovy (BMW) |
11:00 | 11:25 | Data Expert Group - VSS Next and VDM Next | Technical | What is next for the Data Expert Group: VSS and VDM |
| Daniel Alvarez (BMW), Adnan Bekan (BMW) |
11:00 | 11:25 | AOSP App Framework Standardization Expert Group - Working Session - Entertainment | Technical | Entertainment Ecosystem - Why isn’t it taking off and what we can do about it. |
| Chris Goebel (Appsfactory) |
11:00 | 11:25 | Navigating the Future: Omdia’s SDV Forecast and Regional Comparisons | Business | This session explores the expectations for how Software-Defined Vehicles (SDVs) will evolve in the coming years through a blend of forecasts, data-driven insights, and thought-provoking industry expertise. |
| Maite Bazerra, Sr. Principal Analyst, Omdia; Santhosh Jogi, ED, Systems & Core Architecture, Ford |
11:30 | 11:55 | Title: VDM in Practice: End-to-End | Technical | This presentation explores the implications of using the Vehicle Data Model (VDM) as a conceptual and governance layer and how it can influence application and database design in practice. Using a simple EV charging station domain as a proof of concept, it demonstrates an end-to-end workflow from conceptual modeling and The work illustrates how a unified conceptual model can act as a single source of truth across teams and technologies, enabling a common vocabulary while allowing |
| Daniel Alvarez (BMW), Rami Pinto (MongoDB) |
11:30 | 11:55 | Reimagining Customer Connection in the Era of Connected Vehicles | Business | Today, the rise of connected vehicles and the rapidly evolving in-vehicle digital experience are reshaping the innovation model. These advancements challenge us to rethink how we experiment, iterate, and gather customer insights, enabling a deeper, more dynamic emotional connection with the vehicles of the future. |
| Cathi Chinn, VP Innovation & Connected Vehicle Services |
11:50 | 12:30 | AOSP App Framework Standardization Expert Group - Working Session - Future Model For AOSP group and future topics | Technical |
|
| Jan Kubovy (BMW), Richard Fernandes (GM), Sabine Hofschen (BMW), Jose Freitas (FORVIA), Chris Goebel (Appsfactory) |
12:00 | 12:55 | Building Knowledge Graphs to Harmonize Data Models and Standards for Connected Vehicles | Technical | Connected-vehicle ecosystems increasingly span multiple data models and standards ,e.g. vehicle signals and services, fleet/telematics schemas, simulation and validation formats, and emerging regulatory interfaces. Each layer often introduces its own “local” vocabulary (optimized for a product, UI, or pipeline), and the translation cost between these vocabularies compounds over time. This session presents knowledge graphs as a pragmatic harmonization mechanism: a semantic spine that preserves local flexibility while enabling shared discovery, reuse, and cross-domain analytics. Starting from COVESA artifacts (e.g., VSS/VISS and CDSP data pipelines), we describe a pattern language for building an automotive knowledge graph that connects to complementary standards such as ASAM OpenX (OpenDRIVE/OpenSCENARIO/OpenLABEL/OpenODD), ISO/ITS concepts, and generic vocabularies (SSN/SOSA for sensing, QUDT for quantities/units, and SAREF extensions). The focus is incremental adoption: lightweight mappings, versioning and governance to prevent drift, and “just enough” semantics to enable validation and reusable queries across vehicles and environments. Attendees will leave with an actional playbook for reducing integration friction without forcing a one-size-fits-all ontology. |
| Danh Le Phuoc (TU Munich) |
12:00 | 12:25 | Turning AI Potential into Engineering Productivity in the Software-Defined Vehicle Era | Business | As the automotive industry moves toward software-defined vehicles, companies are exploring how generative AI can deliver practical business value across the engineering lifecycle. | No slides | Maurice Dantzler, Executive Director, Advance Electronic Systems, Cummins; Dr. Yansong Chen, CEO of Ropix Tech |
|
|
|
|
|
|
|
2:00 | 2:40 | Automotive AI – Unleashing New Possibilities | Business | Creating opportunities for collaboration within the Automotive AI community can be challenging but generative AI potentially can help the mobility and connected vehicle achieve a safer and more intuitive experience for both driver and passengers. |
| Dr. Yansong Chen, CEO of Ropix Tech; Harry Powell, Global SVP - Data & AI, Endava; Ian Etheredge, ML Expert, Cinemo; Georg Doll, CTO, Automotive & Mobility Industry, Microsoft; Mihai Potoceanu, Sr. Product Manager, RTI |
2:05 | 2:25 | Commercial and Fleet Vehicle Expert Group - Community Update | Technical | Group Status Update
|
| Ted Guild (Geotab), Mouham Tanimou (Bosch) |
2:05 | 2:30 | Accelerate software-defined vehicle (SDV) workflows | Technical | COVESA’s Vehicle Signal Specification (VSS) becomes standard for SDV platforms and app development for OEMs and 3rd app companies in Automotive industry. Cloud-native browser based solutions can help OEMs by leveraging automated VHAL generation and signal mapping. By transforming VSS-compliant vehicle signal models into a standardized Vehicle Hardware Abstraction Layer (VHAL). Applying this clear approach enables early, consistent, and URI-flattened access to vehicle signals—effectively enhancing how industry standards can be applied in real development workflows. This approach allows developers to use standardized vehicle signals as a foundation for building reusable, standards-based Android Automotive applications, entirely within virtual cockpit environments, long before physical vehicle hardware is available. By combining VSS-driven signal standardization with virtualization, new SDV toolchains accelerate prototyping, enable test automation through synthetic and SOME/IP signal feeds, and creates consistent integration points across SDV toolchains. The session will also highlight how cloud-based browser based virtual development environments leads some added values and benefits—such
|
| Ayhan AK (Elektrobit) |
2:25 | 3:05 | Extending VSS to commercial vehicles - Panel-Discussion with ACEA's truck & bus-OEMs and COVESA | Business | Upcoming in-vehicle API will offer FMSv5 migrated to VSS organized in the Commercial and Fleet Vehicle Expert Group-Community. | Panel - No Slides | Armin Keller, Managing Partner, LogiCom; Abdelkader Sellami, Specialist System Design Engineer, Volvo Trucks; Ted Guild, Connectivity Standards Lead, Geotab; Achim Henkel, Director, Systems Engineering & Technical Strategy Mobility, Bosch; Steve Crumb, Executive Director, COVESA |
3:05 | 3:30 | Joint Vehicle Signal Specification/Commercial and Fleet Vehicle Expert Group Working Session | Technical | Topics:
|
| Ted Guild (Geotab), Erik Jaegervall (Bosch), Adnan Bekan (BMW) |
2:45 | 3:20 | Operationalizing Vehicle Data in Insurance, Tires and Other Sectors | Business | The automotive industry has long promised that vehicle data would unlock new value in adjacent sectors, from insurance and logistics to smart infrastructure. Yet, scaling from pilot projects to reliable revenue has proven difficult. | Panel - No Slides | Nicola Concer, Product Manager, Automotive Processors, NXP; John Heinlein, CMO Sonatus; Michael O’Shea, MOTER Technologies |
2:35 | 3:00 | Multi-Domain E/E Testing: AOSP, Open AD Kit and Open/Closed-Loop Simulation | Technical | A demonstration of how virtual E/E architecture simulation - combining open-loop (drive recording playback) and closed-loop (synthetic scenario) approaches - can accelerate and standardize multi-domain vehicle software development and testing. We show how an Android Automotive OS emulator and an Autoware OpenADKit AD/ADAS stack can run within a single virtual E/E architecture definition, with vehicle signals mapped toward a common signal abstraction layer. Participants will see a live demo transitioning from open-loop signal replay - feeding recorded drive data into both IVI and AD stacks - to closed-loop simulation where the AD stack perceives and actuates against a synthetic scenario while the IVI reacts to the resulting vehicle state in real time. Key takeaways include a practical framework for multi-domain virtual testing, signal abstraction as an enabler for cross-domain interoperability, and how a virtual-first approach can accelerate vehicle software development across domains. |
| Per Sigurdson (RemotiveLabs) |
3:05 | 3:30 | VSS and AI: | Technical | AI both consumes and generates vehicle data, challenging static data catalogs and VSS’s ability to describe derivation. The talk highlights selective, purpose-driven data collection as essential for scalable AI lifecycle management and software-defined vehicles. Description: Artificial Intelligence (AI) changes the way data can be produced, processed, and collected. Not only does AI change the way data is processed in a vehicle: AI does not only consume data, it also continuously generates new, derived or synthetic data whose availability and semantics depend on the resultant model itself. Dynamic data preprocessing already challenges the notion of a static data catalogue for an automotive telematics system. AI further challenges this notion: changing an AI model can directly influence which data exists and can be collected. While the Vehicle Signal Specification (VSS) has a notion of derived data, its ability to describe how data is derived is limited. Due to their opaque and non-deterministic nature, AI models make the task of describing how derivation takes place nearly impossible. Add in the need for more complex data such as pictures, video, and lidar data, and it becomes clear that not all data can be transmitted all the time. This talk will discuss these challenges and present use cases that motivate them. Using concrete use cases, it demonstrates why VSS only scratches the surface of the problem and why selective, purpose-driven data collection is essential. Building and maintaining AI models require vast amount of data, both for accomplishing its tasks, as well as for retraining. However, training, inference, validation, and retraining have fundamentally different data requirements. Determining what data is needed where and how often is key to success. It is also key to a real Software Defined Vehicle. Takeaways
|
| James Hunt (aicas) |
3:25 | 3:50 | BMW Keynote | Business |
| Slides not available | Graham Smethurst, VP of E/E Architecture R&D, BMW Group |
3:35 | 3:55 | COVESA Central Data Service Playground (CDSP) Introduction and Status Update | Technical | The COVESA Central Data Service Playground (CDSP) is intended to support collaborative investigation of data centric architectures and for that work to be publishable in the open. After a short project introduction for newcomers, we will cover some of the major newer features and the future roadmap | 
| Stephen Lawrence (Renesas) |
3:35 | 3:55 | VSS Beyond V2X: Standardizing Research Datasets Across Simulation and Real Vehicles | Technical | VSS is known for vehicle-to-cloud and inter-ECU communication, but automotive researchers face a different standardization problem: data heterogeneity across platforms. A colleague's ML project analyzing autonomous driving algorithms required aligning 200+ features from multiple vehicles in different formats—manual feature mapping became a major bottleneck. When combining simulation data (BeamNG, CARLA) with real vehicle telemetry, researchers waste time on format conversion instead of analysis. VSS could solve this, but it hasn't been available in simulation tools. We've developed VSS-compliant telemetry exporters for BeamNG and SimHub (tested across multiple sim racing games). These tools demonstrate a new VSS application: dataset standardization for research workflows. By generating simulation data in VSS format, we eliminate preprocessing overhead when integrating with real vehicle datasets. This isn't just simulation-to-vehicle integration—it's showing VSS as a practical solution for the data wrangling problems researchers actually face. We will be extending this to CARLA and exploring other widely-used simulation platforms. The goal: establish VSS as the common data schema for automotive research, enabling seamless dataset portability between virtual and real testing environments. We welcome collaboration with COVESA members to identify priority platforms and develop guidelines for VSS in research contexts. | Christos Papadopoulos (University of Memphis) | |
4:10 | 4:50 | Integrating Open-source Semantic | Technical | COVESA have established the Central Data Service Playground (CDSP) as a neutral, open environment for experimenting with data-centric architectures across VSS/VISS, cloud backends, and adjacent-industry use cases. With the addition of rule-based semantic reasoning in the Knowledge Layer, CDSP is starting to support deterministic, explainable decision logic that moves beyond scattered IF-ELSE code. However, the current Knowledge Layer Server depends on a single non-open source reasoner, which creates practical barriers for open adoption, long-term evaluation, and reproducible benchmarking across the community. This presentation is a call to action to make CDSP “multi-reasoner ready” by integrating permissively licensed open-source semantic reasoners alongside existing implementations. We will report our initiative (BMW and TU Berlin) in implementing and benchmarking open source semantic reasoners. This initiative aims to provide a minimal Reasoner Adapter interface (rules/ontology loading, incremental updates, query/inference API, and explanation hooks), a shared benchmark harness built around representative VSS/VDM scenarios, and a contribution workflow that lets members plug in new engines and compare them fairly. Attendees will leave with three concrete outcomes: (1) an agreed integration blueprint and backlog for an OSS reasoner option in CDSP, (2) a candidate short-list and evaluation criteria (performance, determinism, deployability, licensing), and (3) a roadmap to turn these contributions into a community-maintained capability that accelerates SDV and fleet-data innovation. |
| Christian Muehlbauer (BMW), Haonan Qiu (BMW), Danh Le Phuoc (TU Berlin) |
4:10 | 4:50 | Generative AI Group - Working Session - Generative AI for Automotive Engineering: From Concept to Code | Technical | The COVESA Generative AI Group is building practical AI tooling for automotive digital engineering. We present our first concrete output — the vss-translator MCP server enabling AI assistants to author, validate, and convert VSS signal specs — and dive into the DBC/CAN-to-VSS translation use case. The second half covers our ASPICE MAN.3 agent framework and opens the floor: bring your use cases and help shape the group's roadmap. 4:10–4:15 - Welcome + Who are we? - Yansong / Georg |
| Georg Doll (MS), Yansong Chen (Ropix), Bogdan Racotea (SmartEngineering), Ragip Selcuk, Susan Butler |
4:10 | 4:50 | Deriving Actionable Insights from Vehicle Data to Advance Business, Safety, and Operations | Business | Modern road administrations, fleet operators, insurance companies, and a host of other organizations are increasingly leveraging vehicle data generated by connected car platforms to improve how they operate, maintain, and manage their roads, fleets, and products. | Panel - No Slides | Roger Lanctot, CEO and Founder, StrategiaNow Consulting; Jose Carlos, CEO, Xouba; Harry Powell, Global SVP - Data & AI, Endava; Michael O’Shea, COO & CTO, MOTER Technologies; Ted Guild, Connectivity Standards Lead, Geotab |
4:40 | 5:30 | Ready to Go Together: Unveiling JASPAR’s New OSS Project | Technical | JASPAR, after two years of activities in its API Working Group, is now ready to release the results as open source. These outcomes extend and further reinforce the long-standing efforts of COVESA. In this session, we will provide the world’s first in-depth overview of what will be released as OSS and share the direction for accelerating future collaboration with COVESA. |
| Shinichi Suzuki (JASPAR/Toyota), Takafumi Yoshida (JASPAR/Astemo) |
Day 2, April 23, 2026
Start Time | End Time | Title | Type | Description | Presentation | Speakers |
|---|---|---|---|---|---|---|
9:00 | 9:30 | Vehicle Information Service Specification - Community Update | Technical | Project Status Update What has been done since last AMM What is the focus/being done at this AMM What is the focus until next AMM |
| Ulf Bjorkengren (Ford) |
9:00 | 9:30 | Agentic AI in the Smart Cockpit: Emerging Security Gaps at the Edge | Technical | As in-vehicle AI systems evolve from passive voice interfaces to multimodal, agentic applications, the security model of the smart cockpit must be re-evaluated. Unlike earlier infotainment systems, these AI architectures reason, plan, invoke tools, and increasingly operate across vehicle domains. This expanded capability fundamentally changes the system’s risk profile, introducing threats that traditional automotive cybersecurity approaches were not designed to address. Prompt injection and jailbreak techniques can manipulate model tool-using decisions rather than exploit code, enabling the hijacking of AI models and agents. Meanwhile, the erosion of domain isolation assumptions increases the potential impact of compromised AI behavior across vehicle services and user data. This presentation will explore emerging attack paths, including tool misuse, agent context manipulation, and unintended cross-domain actions, and understand why model-level safeguards alone are inadequate. It will examine why cloud-centric AI guardrail models are insufficient for in-vehicle deployments, where security cannot come at the cost of real-time performance or interaction quality. Finally, it will outline a forward-looking, risk-based security architecture for edge AI systems — an approach that preserves performance while enforcing trust boundaries, validating tool outputs, protecting memory integrity, and inspecting AI inputs and outputs where decisions are made. |
| Gregor Knappick (VicOne)- |
9:30 | 9:55 |