When we talk about Security and Defence, we are talking about one of the most demanding domains for any technology: unpredictable environments, critical real-time decisions, zero tolerance for error. It is precisely in that context that AIR Institute has stepped forward.
We are pleased to announce our participation in ROCCA (Operative Robotics for Quality Control, Inspection and Assisted Maintenance for Defence), a Cervera Strategic Project funded by the Spanish Ministry of Science and Innovation through the CDTI, launching in 2026 with an investment of nearly €6 million and a working horizon extending to 2029.
What is ROCCA?
ROCCA is a collaborative initiative bringing together seven Spanish technology centres — AIJU (coordinator), AIN, AITIIP, AIR Institute, CETEM, INESCOP and VICOMTECH — around a shared goal: developing advanced digital technologies in robotics and artificial intelligence to address the new challenges facing the Security and Defence sector.
The project works across three major pillars: robotic inspection, intelligent quality control and assisted predictive maintenance, with direct application in critical infrastructures, complex industrial environments and aerospace defence platforms.
It sits within the line of digital technologies associated with Industry 5.0, where human-machine collaboration, operational autonomy and system trustworthiness are defining factors.
Our contribution to the consortium is built around the areas where we are a recognised reference: intelligent autonomous systems, post-quantum cybersecurity and explainable AI.
Autonomous agents that learn and collaborate
We contribute Agentic AI architectures based on Multi-Agent Deep Reinforcement Learning (MA-DRL): systems where multiple intelligent agents operate in a decentralised manner, coordinating with one another to tackle missions in dynamic environments. We complement this with autonomous inter-agent communication capabilities through LLMs and imitation learning, enabling heterogeneous platforms to interoperate naturally.
Cybersecurity built for the future
In defence environments, security cannot be an afterthought. That is why we integrate post-quantum encryption and AI-based intrusion detection, ensuring that the systems developed are resilient against current threats as well as those that quantum computers will pose in the near horizon.
AI that explains its decisions
In critical contexts, it is not enough for AI to get things right — it must be able to justify its decisions. Our work in Explainable AI (XAI) ensures that predictive maintenance and computer vision models are interpretable, auditable and trustworthy for human operators.
Lightweight and efficient computer vision
We develop hybrid CNN-transformer architectures with interpretable attention mechanisms, optimised to run on edge devices, supporting visual inspection and quality control tasks under field conditions.
A track record that backs our commitment
This project does not come out of nowhere. AIR Institute's participation in ROCCA draws on a solid track record in cutting-edge projects: NOUS and dAIEDGE (Horizon Europe), QUANTUMSHIELD (industrial post-quantum cybersecurity), DIGIS3 (the smart digitalisation EDIH that we coordinate ourselves) or CIBERIA (cross-border digital resilience), among others.
With over 100 specialist professionals and more than 90 R&D projects to our name, joining a Cervera consortium of this calibre is also a recognition of the technological maturity we have built as an organisation.
Why this project matters
ROCCA is not just another R&D project. It is a signal of the direction European industry is heading: defence and security systems that incorporate AI in a rigorous, ethical and trustworthy way. Systems that are not only intelligent, but can explain what they do, withstand sophisticated attacks and operate autonomously when circumstances demand it.
At AIR Institute, we have spent years preparing for this moment. Now it is time to build it.
Want to know more about our work in autonomous systems, post-quantum cybersecurity or explainable AI? Get in touch at comunicacion@air-institute.com