CoTwin: Collaborative Digital Twins for Smart Mobility
CoTwin (Collaborative Digital Twins for Smart Mobility) was a project led by MobilityXlab & Lindholmen Science Park and financed by Vinnova/FFI to demonstrate a secure and sustainable data-sharing solution to solve challenges for local public authorities and society
The transport sector generates enormous amounts of data: camera footage, GPS signals, radar readings. But almost none of it is shared. Legal uncertainty, technical barriers, and GDPR concerns have kept this data locked inside organisational silos, limiting its potential to improve road safety, logistics, and urban planning.
CoTwin (Collaborative Digital Twins for Smart Mobility) was a Vinnova FFI-funded research project that ran from November 2024 to March 2026. The consortium brought together Lindholmen Science Park, MobilityXlab, Volvo Group, Ericsson, CanaryBit, AstaZero, and the City of Gothenburg to test whether these barriers could be overcome.
CoTwin aimed to develop and validate a federated infrastructure for secure, privacy-compliant, real-time data sharing between vehicles and road infrastructure. Specifically, the project set out to:
- Build and demonstrate a federated infrastructure combining mobile edge computing with Trusted Execution Environment (TEE) technology.
- Validate a GDPR-compliant data governance framework for vehicle sensor data.
- Run at least one end-to-end data sharing use case using real vehicle data at a test facility.
- Assess the commercial potential of the platform through customer research and market analysis.
- Engage external innovators through an open hackathon to generate and test use case ideas.
How the project was structured
The project ran across six work packages, each led by the partner best placed to deliver it. Work packages covered everything from infrastructure deployment and AI model development to regulatory engagement and open innovation.
Coordination
Project management and reporting. Led by Lindholmen Science Park / MobilityXlab.
Trusted environment
Ericsson and CanaryBit deployed and tested the confidential edge computing infrastructure at NorthStar/AstaZero.
Data and use cases
Volvo Group developed the on-vehicle AI application and edge integration, using real camera footage from Lindholmen area test drives.
Regulatory compliance
CanaryBit, Volvo Group, and Ericsson participated in eight structured workshops with IMY to obtain formal legal guidance on GDPR and TEEs.
Business models
MobilityXlab led 16 customer interviews, workshops, and concept development sessions to assess commercial viability.
Hackathon
A two-day open innovation hackathon at MobilityXlab in November 2025, engaging 60+ participants across startups, industry, and academia.
Key results and conclusions
Technical demonstration
The project deployed a confidential computing cluster at AstaZero and validated a two-stage AI pipeline for automated road sign detection. An on-vehicle model performed binary detection at 82% accuracy, passing only high-probability frames to an edge-based model that classified road sign types at 84% accuracy. The system reduced bandwidth compared to transmitting raw video, preserved privacy throughout, and was validated using real footage from a Volvo truck in Lindholmen. This addressed a concrete operational need identified by the City of Gothenburg.
Regulatory guidance
CoTwin participated in IMY's regulatory sandbox programme, which reached three conclusions. TEEs qualify as a meaningful technical safeguard under Article 32 of GDPR, providing technically verifiable control over data processing in a way that conventional cloud solutions cannot. Processing vehicle camera footage inside a TEE constitutes personal data processing under GDPR and must be handled accordingly. A mobile operator that provides compute capacity without access to the data, and without control over attestation or encryption keys, is unlikely to be classified as a data controller or data processor.
Market and business model
The business model assessment concluded that CoTwin's strongest near-term market is not general-purpose data processing, but high-stakes applications where ultra-low latency, mandatory data locality, and privacy-preservation all apply at once. Heavy vehicle logistics and infrastructure maintenance were identified as the most commercially viable entry points, ahead of smart city applications, which tend to lack the budget and scale needed to sustain a platform.
Open innovation
The November 2025 hackathon produced five use case concepts: Federated Safety Analytics, Blind Spot Sharing, Smart Parking, Real-Time Synthetic Data for Smart Mobility, and Life Atlas. Social media coverage of the event generated over 1,100 engagements at an average engagement rate of 16.4%, indicating genuine interest in the problem the project addressed.
CoTwin advanced the platform from TRL 2 (concept) to TRL 6 (demonstrated in a real environment). The technical architecture is validated, the regulatory questions have formal answers, and the commercial path is mapped. Follow-on work can now focus on scaling the platform to multiple vehicles and infrastructure nodes, extending the regulatory guidance to additional data types and cross-border scenarios, improving the AI models with larger datasets, and contributing to EU-level standardisation work on confidential computing and vehicle data APIs.
Do you want to know more?
