TEACHING · Project

AI Safety Toolkit That Helps Self-Driving Cars and Aircraft Respond to Human Behavior

transportPilotedTRL 5

teaching-h2020.eu

Imagine a self-driving car that not only follows the road but actually reads your stress level and adjusts how it drives — slowing down if you're nervous, or taking over smoothly if you're drowsy. That's what this project built: an AI system that keeps humans in the loop for safety-critical machines like autonomous vehicles and aircraft. Instead of replacing people, it makes the machine and the human work together, with the AI learning from human feedback to get better and safer over time. They tested it in real autonomous driving and aviation scenarios across 5 European countries.

By the numbers

consortium partners across Europe

countries involved (AT, DE, EL, FR, IT)

pilot domains tested (autonomous driving and aviation)

industry partners in the consortium

55%

industry participation ratio

total project deliverables produced

The business problem

What needed solving

Companies building autonomous vehicles and aircraft face a critical gap: their AI systems make decisions without understanding the human in the seat. When a driver is stressed, distracted, or a pilot is fatigued, current systems either ignore it or handle it crudely. This leads to safety incidents, regulatory pushback, and consumer distrust that slows adoption of autonomous technology worth billions in potential market value.

The solution

What was built

The project built a complete computing platform with AI-as-a-Service capabilities for human-aware autonomous systems. Key deliverables include: a final-release computing and communication platform evaluated against real use cases, an AI-as-a-Service system (also final release), engineering methods and architecture patterns for dependable cyber-physical systems, and datasets from pilot deployments in autonomous driving and aviation.

Audience

Who needs this

Autonomous vehicle developers needing human-state-aware AI for ADAS and self-driving systemsAvionics companies building next-generation cockpit automation with pilot-aware AIEdge computing platform manufacturers targeting safety-critical automotive and aerospace marketsAutomotive Tier-1 suppliers developing AI-based driver monitoring and intervention systemsCybersecurity firms specializing in securing autonomous vehicle and aircraft communication systems

Business applications

Who can put this to work

Automotive & Autonomous Vehicles

enterprise

Target: Autonomous vehicle developers and Tier-1 automotive suppliers

If you are an automotive company developing advanced driver-assistance or autonomous driving systems and struggling with the handoff between AI control and human intervention — this project developed a computing platform with AI-as-a-Service that monitors human state and adapts vehicle behavior accordingly. It was piloted in autonomous driving scenarios and comes with engineering methods for building dependable systems that regulators can actually certify.

Aviation & Aerospace

enterprise

Target: Avionics system integrators and aircraft manufacturers

If you are an aviation company dealing with cockpit automation complexity where pilots must seamlessly interact with increasingly autonomous flight systems — this project developed human-aware AI that adapts to pilot feedback in real time, tested in aviation pilot scenarios. The dependable AI architecture runs on edge computing, meaning faster decisions without relying on cloud connectivity, critical for in-flight operations.

Industrial Edge Computing

mid-size

Target: Edge computing platform providers and IoT system integrators

If you are a technology company building edge computing hardware or platforms for safety-critical applications and need AI capabilities that are energy-efficient and secure — this project delivered a final-release computing and communication platform with integrated cybersecurity support. With 6 industry partners in the consortium and a ready AI-as-a-Service architecture, this is a tested integration path for adding human-aware AI to edge devices.

Frequently asked

Quick answers

What would it cost to license or adopt this technology?

The project was a publicly funded Research and Innovation Action, so core results are likely available through academic licensing from Università di Pisa or the 6 industry partners. Specific licensing terms would need to be negotiated directly with the consortium. No commercial pricing has been published.

Can this work at industrial scale in real vehicles or aircraft?

The project demonstrated its platform in 2 pilot scenarios: autonomous driving and aviation. The final release of the computing platform was evaluated with real use case integration. However, scaling from pilot to mass production would require additional engineering and certification work specific to each deployment context.

Who owns the intellectual property?

IP from EU-funded RIA projects is typically retained by the consortium partners who generated it. With 6 industry partners (55% of the consortium), significant IP likely sits with commercial entities. Specific licensing arrangements would need to be discussed with the coordinator at Università di Pisa.

Does this meet automotive or aviation safety regulations?

The project specifically addressed dependability and safety for autonomous safety-critical systems, with dedicated deliverables on engineering methods and architecture patterns for dependable systems. Based on the project objectives, regulatory acceptability was a core design goal, though formal certification would depend on the specific application domain.

How long would integration take?

The project ran from January 2020 to June 2023 and produced 22 deliverables including a final-release computing platform and AI-as-a-Service system. Integration timelines would depend on your existing systems, but the platform was designed for deployment of adaptive applications, suggesting it was built with integration in mind.

Does it work with existing automotive or avionics systems?

The platform was designed as a computing toolkit for building autonomous applications, with a distributed architecture leveraging edge computing. The use case deployment reports and integration evaluations suggest it was tested against real-world system interfaces. Specific compatibility would need to be verified with the consortium.

Is there ongoing support or development?

The project closed in June 2023. With 11 partners across 5 countries and strong industry participation, some partners may continue development commercially. The 2 SMEs in the consortium are likely candidates for offering commercial support or spin-off products based on the results.

Consortium

Who built it

This is a strong, industry-heavy consortium with 11 partners across 5 European countries. The 55% industry ratio (6 out of 11 partners) is notable — it means the technology was developed with commercial application in mind from day one, not just academic research. The consortium includes 2 SMEs that could serve as agile commercialization vehicles, backed by 3 universities (including coordinator Università di Pisa) providing deep AI and computing research, and 2 research organizations. The geographic spread across Austria, Germany, Greece, France, and Italy covers major European automotive and aerospace markets. For a business looking to adopt this technology, the mix of industry partners who understand production realities and academic partners who built the core AI gives confidence that results are both scientifically rigorous and practically applicable.

UNIVERSITA DI PISACoordinator · IT
AIT AUSTRIAN INSTITUTE OF TECHNOLOGY GMBHparticipant · AT
TECHNISCHE UNIVERSITAET GRAZparticipant · AT
IDEAS & MOTION SRLparticipant · IT
CHAROKOPEIO PANEPISTIMIOparticipant · EL
INFORMATION TECHNOLOGY FOR MARKET LEADERSHIPparticipant · EL
INFINEON TECHNOLOGIES AGparticipant · DE
CONSIGLIO NAZIONALE DELLE RICERCHEparticipant · IT
AVL LIST GMBHparticipant · AT
THALESparticipant · FR
MARELLI EUROPE SPAparticipant · IT

How to reach the team

Coordinator is Università di Pisa (Italy). Contact their technology transfer office or the project PI through the university website.

Order a report on this consortium See UNIVERSITA DI PISA profile →

Next steps

Talk to the team behind this work.

Want an introduction to the TEACHING consortium for licensing discussions or technical evaluation? SciTransfer can arrange a direct meeting with the right partner for your use case.

Order a report on this topic More in Transport & Mobility