Results & Downloads

Vehicle-to-Everything Trust: Enabling Autonomous Trust Assessment of V2X Data by Vehicles

Authors: Nataša Trkulja, Artur Hermann, Paul Lukas Duhr, Echo Meißner, Michael Buchholz, Frank Kargl, Benjamin Erb

Final newsletter

Challenges and Priorities towards Trustworthy AI

Authors: Ioannis Krontiris, Frank Kargl, Thanassis Giannetsos

D6.2 details the evaluation results of CONNECT’s Trust Assessment Architecture when integrated/deployed in the context of the envisioned use cases unlocking the assessment of such advanced trust mechanisms as part of the end-to-end operational profile of CCAM functionalities. Building on top of the benchmarking results (per standalone component) that were documented in the previous version of this deliverable, D6.2 validates the framework’s tangible impact through three core CCAM use cases: Intersection Movement Assistance (IMA), Cooperative Adaptive Cruise Control (C-ACC), and Slow Moving Traffic Detection (SMTD). Each was tested using the most appropriate methodology – from Large-Scale Simulation scenarios to Hardware-in-the-Loop and Real-World Living Labs environment following the development pipeline adopted in V2X security and safety engineering: from conception activities and prototyping to real-world evaluation. Our findings confirm the framework’s benefits in allowing trust-aware decision making, achieving all identified KPIs and positioning it a stepping stone for future research towards converging security and safety, which can further drive industry-wide adoption.

A Framework for Dynamic Trustworthiness Assessment in Cooperative and Automated Vehicles

Contributors from CONNECT: Denso Automotive, Huawei Technologies

Quantification Methods for Trust in Cooperative Driving

Authors: Artur Hermann, Nataša Trkulja, Patrick Wachter, Benjamin Erb, and Frank Kargl

Demo: Quantifying Trust in a Trust Assessment Framework

Authors: Artur Hermann, Nataša Trkulja, Echo Meißner, Benjamin Erb, Frank Kargl

A Real-Time Implementation of an Edge-Assisted Collision Warning System for Urban Intersections

Authors: Christian Vitale, Panagiotis Kardaras, Panayiotis Kolios, Carla Fabiana Chiasserini and Georgios Ellinas

This deliverable completes the description of the work of WP5 on the MEC-enabled networking technology to support the CONNECT task offloading needs, the CONNECT Distributed Ledger Technology, the use of Verifiable Credentials and the crypto protocols used to ensure the security and privacy of data in the CONNECT eco-system. Trusted task offloading, the CONNECT Distributed Ledger Technology and the use of attributes for access control and for signing and encrypting failed attestation data have all been described, implemented and measured. All of these extensions work in tandem with the MEC-enabled Orchestrator (described in earlier deliverables) managing the secure lifecycle management of CCAM services deployed over the continuum.

D4.3 presents the final update of the CONNECT Guard architecture including security extensions for virtualization and edge computing in CCAM systems. It introduces the latest trusted computing mechanisms, secure container deployment, and the TALOS framework for verifiable enclave migration. The document also evaluates cryptographic protocols for privacy-preserving trust assessment, enabling secure, scalable, and resilient service orchestration across vehicles, MEC, and cloud. It successfully concludes the mission of WP4 of CONNECT.

Dynamic Group Signatures with Verifier-Local Revocation

Authors: Callum London, Daniel Gardham, Constantin Catalin Dragan

This deliverable describes the final version of the standalone trust assessment framework (TAF), the federated TAF and the digital twin version of the TAF (TAF-DT). For the standalone TAF, the federated TAF and the TAF-DT, the final architecture, implementation details, and evaluation results are provided. In addition to that, the approaches for trust quantification, the approach for calculating the required trustworthiness level (RTL), and details about the risk assessment framework are provided.

Fourth newsletter

Leaflet for the CONNECT Workshop on Trustworthy AI, held on March 25^th-26^th 2025 in Frankfurt.

GenSUMO: GenAI-Creation of Critical Scenarios for Autonomous Vehicle Testing

Authors: S. Angelini, D. Gasco, C.Casetti

The CONNECT project addresses security, privacy, and trust in Cooperative, Connected, and Automated Mobility (CCAM) systems. D2.2 extends the overall architecture outlined in D2.1 and focuses on ethical analysis, the final architecture, detailed threat modelling and requirements, and the definition of the CONNECT Minimal Viable Platform. By prototyping this architecture, CONNECT will ensure secure collaboration across vehicles, MEC, and cloud, providing a foundation for privacy-preserving, trustworthy mobility solutions.

Third newsletter

AI/ML-based Services and Applications for 6G-connected and Autonomous Vehicles

Authors: Claudio Casetti, Carla Fabiana Chiasserini, Falko Dressler, Agon Memedi, Diego Gasco, Elad Michael Schiller

On Enhancing Intersection Applications With Misbehavior Detection and Mitigation

Authors: Jiahao Zhang, Ziyi Liu, Ines Ben Jemaa, Francesca Bassi, Fawzi Nashashibi

Pluralism and the Design of Autonomous Vehicles

Authors: Adam Henschke, Chirag Arora

This deliverable presents the second and final version of the CONNECT Data Management Plan. In addition, the deliverable includes a roadmap definition as it pertains to the ethical and legal aspects of the Trust Assessment part, which plays a central role in CONNECT.

Personalized and Context-aware Route Planning for Edge-assisted Vehicles

Authors: Dinesh Cyril Selvaraj, Falko Dressler, Carla Fabiana Chiasserini

Adaptive Autopilot: Constrained DRL for Diverse Driving Behaviors

Authors: Dinesh Cyril Selvaraj, Christian Vitale, Tania Panayiotou, Panayiotis Kolios, Carla Fabiana Chiasserini and Georgios Ellinas

OScar: An ETSI-Compliant C-ITS Stack for Field-Testing with Embedded Hardware Devices

Authors: Marco Rapelli, Francesco Raviglione, Claudio Casetti

Second newsletter

Maneuver Coordination Using V2I to Improve Traffic Efficiency at Intersections

Authors: Lorenzo Farina, Marco Rapelli, Barbara Mavı́ Masini, Claudio Casetti, Alessandro Bazzi

ETSI-Compliant Protection of Connected Vulnerable Road Users: Simulation and Field Trial

Authors: Marco Rapelli, Francesco Raviglione, Claudio Casetti

All project videos in one place!

D5.2 follows up the D5.1 contributions to deliver the first version of the CONNECT orchestrator and the mechanisms for secure data exchange. The here-in reported work covers the task-offloading logic and the implementation details of the CONNECT orchestartor software. Lastly, the work is completed with the Blockchain architecture and data handling principles to facilitate the CONNECT accountable attestation and monitoring (when combined with the D5.1 claims).

D4.2 documents the high-level design of the complete CONNECT security architecture as developed by Workpackage WP4. Compared to D4.1 we refine and add users stories and designs for the Multi- Access Edge Cloud (MEC). We also add detailed cryptographic protocols that implement the trust assessment of the services of CONNECT. Finally, we detail the concepts and interactions for secure migration and up- grade of TEE-protected workloads. This is the technical foundation for Secure Task Offloading. The key questions that this deliverable answers are “How can the Multi-access Edge Cloud (MEC) be architected to provide robust security and privacy guarantees?”, “How can cryptography help to remotely assess trust of services while maintaining sound privacy guar- antees?”, and “How can vehicle and edge collaborate while maintaining security?”.

This deliverable builds upon D3.1 and describes major build- ing blocks for the standalone trust assessment framework (TAF), the federated TAF and the digital twin version TAF-DT. For the standalone TAF, a fine-grained architecture underlying our current prototype is provided, federated TAF and TAF-DT are described as high-level architectures.

This deliverable presents the CONNECT dissemination, communication, clustering, standardization, and exploitation activities up to M18. Furthermore, it contains information highlighting open-source contributions and workshop organization activities.

WIP: A Trust Assessment Method for In-Vehicular Networks using Vehicle Risk Assessment

Authors: Artur Hermann, Natasa Trkulja, Anderson Ramon Ferraz de Lucena, Alexander Kiening, Ana Petrovska, Frank Kargl

In-vehicle Trust Assessment Framework

Authors: N. Trkulja; A. Hermann; A. Petrovska; A. Kiening; A. Ramon Ferraz de Lucena, F. Kargl

Sphinx-in-the-Head: Group Signatures from Symmetric Primitives

Authors: Liqun Chen, Changyu Dong, Christopher J. P. Newton and Yalan Wang

MS-VAN3T-CARLA: An Open-Source Co-Simulation Framework for Cooperative Perception Evaluation

Authors: C. M. Risma Carletti, C. Casetti, J. Härri and F. Risso

ms-van3t: An integrated multi-stack framework for virtual validation of V2X communication and services

Authors: F. Raviglione, C. M. Risma Carletti, M. Malinverno, C. Casetti, C. F. Chiasserini

D4.1 documents the initial results of Workpackage 4. It speci- fies the core security architecture of CONNECT revolving around the use of HW-based trust anchors for supporting both the continuous trust as- sessment of the CCAM-wide continuum as well as the secure lifecylce management of all comprised devices. D4.1 refines the architecture (com- ponents and flows) presented in D2.1 and puts forth a deteiled description of all interactions and sequence of actions with the underlying Trusted Computing Base for providing node- and data-centric trustworthiness ev- idence. Furthermore, it documents user stories that specify desired fea- tures, and finally outlines the key hierarchy and formalized requirements. This overarching TEE architecture will guide the design of all security pro- tocols and cryptographic mechanisms to be documented in D4.2.

D5.1 paves the way to the CONNECT orchestration im- plementation by first defining the task migration/offloading problem and then presenting a comprehensive analysis of the state-of-the-art. Derived insights suggest candidate approaches for the CONNECT implementation. Furthermore, the document explores identity management and proposes verifiable and presentation claims to verify a CONNECT entity’s attributes or identity. The involved data model to hold the needed information is finally introduced.

Machine-Checked Proofs of Accountability: How to sElect who is to Blame

Authors: Constantin Catalin Dragan, François Dupressoir, Kristian Gjøsteen, Thomas Haines, Peter Rønne, Morten Rotvold Solberg

D2.1 defines the technical requirements of CONNECT, alongside the specification of the conceptual reference architecture, the functional components, and interfaces between them. It also provides an analysis and point of reference for CONNECT in relation to the use cases and reference scenarios including an analysis of the trust models, that need to be considered for capturing the security and privacy requirements of each of the target CCAM services, and the trusted computing anchors to be further investigated and deployed as part of all actors – both at the MEC and Edge. Its purpose is to define the parameters for the rest of the CONNECT project and provide the necessary input for the design and implementation of all security enablers and models towards the dynamic trust assessment of complex CCAM ecosystems.

This Project Quality Plan constitutes a set of project templates and explanations on the project management process, review process, quality checks and meeting organisation, which are communicated to all partners. It further includes the risk assessment plan and shows how potential risks are assessed and mitigated.

Inaugural newsletter

Hash-based Direct Anonymous Attestation

Authors: Chen L.; Dong C.; El Kassem N.; Newton C.J.P. and Wang Y.

Safer Roads using a Vehicle Trust Assessment Framework

Edge-assisted ML-aided Uncertainty-aware Vehicle Collision Avoidance at Urban Intersections

Authors: Dinesh Cyril Selvaraj, Christian Vitale, Tania Panayiotou, Panayiotis Kolios, Carla Fabiana Chiasserini, Georgios Ellinas

Deliverable 3.1 presents initial work towards the Trust Asssessment Framework (TAF). It includes definition of core terminology for trust assessment, a state of the art analysis and requirements engineering which motivates our decision for a TAF based on Subjective Logic. Subsequently, an ap-proach to trust modeling is described including a methodol- ogy for deriving trust relationships and examples for such trust relationships and trust sources are given based on the four CONNECT use case. Finally, a high-level architecture of the TAF is presented and first considerations for the evaluation of the TAF are described.

CONNECT Fact Sheet 2: TRUSTWORTHINESS
Authors: Chirag Arora, Adam Henschke

ZEKRA: Zero-Knowledge Control-Flow Attestation (pre-camera version Open Access)

Heini Bergsson Debes; Thanassis Giannetsos; Edlira Dushku; Ali Marandi

The Internet of Vehicles (IoV) — Security, Privacy, Trust, and Reputation Management for Connected Vehicles
Authors: A. Drobot, T. Zhang, M. L. Buonarosa, F. Kargl, S. Schwinke and B. Sikdar

Securing Cooperative Intersection Management through Subjective Trust Networks

Authors: Frank Kargl, Nataša Trkulja, Artur Hermann, Florian Sommer, Anderson Ramon Ferraz de Lucena, Alexander Kiening, Sergej Japs

S-LDM: Server Local Dynamic Map for 5G-based Centralized Enhanced Collective Perception

Authors: Carlos Mateo Risma Carletti, Francesco Raviglione, Claudio Casetti, Filippo Stoffella, Girma Mamuye Yilma, Filippo Visintainer

Comparative Evaluation of PKI and DAA-Based Architectures for V2X Communication Security
Authors: A. Angelogianni I. Krontiris, and T. Giannetsos

How To Bind A TPM’s Attestation Keys With Its Endorsement Key

Authors: Liqun Chen, Nada El Kassem, Christopher J P Newton

CONNECT Fact Sheet 1: TRUST ASSESSMENT
Authors: Ana Petrovska, Natasa Trukulja, Artur Hermann, Frank Kargl

Click on images in the fact sheet for videos to start.

This deliverable presents the CONNECT communication kit including the project’s visual identity as well as communication and dissemination material to be used within the project. Furthermore, a detailed standardization plan is documented constituting a guideline for the liaison and further dissemination of the CONNECT project to external target groups and related standardization working groups. A timeline of short- and long-term dissemination activities is presented.

An ML-Aided Reinforcement Learning Approach for Challenging Vehicle Maneuvers

Authors: Dinesh Cyril Selvaraj, Shailesh Hegde, Nicola Amati, Francesco Deflorio and Carla Fabiana Chiasserini

The official leaflet of the Horizon Europe project CONNECT is now available, containing information about the consortium, the project’s mission, vision and objectives.

Announcing the launch of the Horizon Europe project CONNECT!