If you are a factory operator dealing with production delays caused by sending sensor data to the cloud and waiting for responses — this project developed a self-learning edge computing node that processes data right on the shop floor. With 32 partners across 7 countries building and testing prototypes, the platform is designed to scale from simple sensor nodes to high-performance edge servers, cutting response times for time-critical operations.
Smart Edge Computing Nodes That Learn and Adapt for Industrial IoT
Imagine your factory machines could think for themselves right where they sit, instead of sending every bit of data to a distant cloud server and waiting for instructions. FRACTAL built a kind of smart building block — a computing node that sits at the "edge" near your equipment, learns from its surroundings, and makes decisions on the spot. These blocks can be stacked and connected like LEGO pieces, from tiny sensors to powerful local servers, forming a network that gets smarter as it grows. The result is faster responses, better security, and systems that keep running even when the internet connection drops.
What needed solving
Industrial IoT systems today face a painful tradeoff: send all sensor data to the cloud (slow, expensive, insecure) or process it locally with dumb devices that cannot learn or adapt. Time-critical operations — robotic assembly, grid management, autonomous vehicles — cannot afford the latency of cloud round-trips. Companies need computing power at the edge that is smart enough to make decisions locally, secure enough for industrial environments, and flexible enough to scale from a single sensor to an entire facility.
What was built
The project built prototypes of scalable cognitive edge computing nodes — hardware platforms that serve as building blocks for industrial IoT. These nodes can learn from their environment, predict their own performance, and adapt to changing conditions. The 38 deliverables include the reference hardware designs and the fractal network architecture that lets nodes self-organize from low-power sensors to high-performance edge servers.
Who needs this
Who can put this to work
If you are a utility company dealing with unreliable connectivity at remote substations and wind farms — this project developed a secure, low-power computing node that can monitor equipment, predict failures, and act autonomously without a constant cloud connection. The cognitive capabilities mean each node learns local operating patterns and adapts, reducing downtime across your distributed infrastructure.
If you are a building automation company dealing with the complexity of connecting hundreds of sensors across HVAC, lighting, and security systems — this project developed a fractal network architecture where edge nodes self-organize and share intelligence. With 22 industry partners validating the design, the open-standard platform lets you scale from a single building to an entire campus without redesigning your system.
Quick answers
What would it cost to implement this edge computing platform?
The project budget details are not available in the dataset. However, the platform is built on open-standard hardware and software (OSS), which typically reduces licensing costs compared to proprietary edge solutions. Contact the coordinator for pricing of prototype hardware and integration support.
Can this scale to hundreds or thousands of nodes across our operations?
Scalability is a core design goal. The fractal architecture means nodes are identical building blocks that scale from low-computing sensor nodes to high-computing edge servers. The consortium of 32 partners across 7 countries tested this scalability principle across multiple use cases.
What is the IP situation — can we license or buy this technology?
The project was funded as a Research and Innovation Action (RIA) under ECSEL. IP is typically shared among consortium partners. With 22 industry partners and 13 SMEs involved, commercial licensing paths likely exist. Contact IKERLAN S. COOP (coordinator) for specific licensing terms.
How does this handle security for industrial environments?
Security is one of the project's core pillars alongside reliability and low power. The platform was designed as a 'Safe-Reliable' hardware node with cognitive capabilities to detect and adapt to threats at the edge. Based on available project data, security was validated through the prototype reference designs.
How does this integrate with our existing cloud and IT infrastructure?
The platform is designed as a computing layer between physical devices and the cloud, not a replacement for either. It uses open standards (OSS) for interoperability. The 69% industry ratio in the consortium suggests strong focus on real-world integration requirements.
What is the timeline from evaluation to deployment?
The project ran from September 2020 to August 2023 and produced 38 deliverables including working prototypes. As a closed project, the technology is available now for evaluation. Deployment timeline would depend on your specific use case and integration complexity.
Who built it
This is a heavyweight industrial consortium with 32 partners across 7 European countries (Austria, Switzerland, Germany, Spain, Finland, France, Italy). The 69% industry ratio — with 22 industry partners including 13 SMEs — signals that this was built with commercial deployment in mind, not just academic research. Coordinator IKERLAN is a well-known Spanish technology center specializing in embedded systems and industrial electronics. The mix of 7 universities and 3 research organizations provided the scientific backbone, while the dominant industry presence ensured the prototypes address real-world constraints like power consumption, reliability, and security in production environments.
- IKERLAN S. COOPCoordinator · ES
- UNIVERSITAET SIEGENparticipant · DE
- PROINTEC SAparticipant · ES
- TIETOEVRY FINLAND OYparticipant · FI
- OULUN YLIOPISTOparticipant · FI
- AITEK SPAparticipant · IT
- SIEMENS AKTIENGESELLSCHAFT OESTERREICHparticipant · AT
- VIRTUAL VEHICLE RESEARCH GMBHparticipant · AT
- INDRA SISTEMAS SAthirdparty · ES
- UNIVERSITA DEGLI STUDI DELL'AQUILAparticipant · IT
- EIDGENOESSISCHE TECHNISCHE HOCHSCHULE ZUERICHparticipant · CH
- RO TECHNOLOGY SRLparticipant · IT
- AVL LIST GMBHparticipant · AT
- AKKODIS ITALY SRLparticipant · IT
- INDRA FACTORIA TECNOLOGICA SLthirdparty · ES
- RULEX INNOVATION LABS SRLparticipant · IT
- CAF SIGNALLING S.Lparticipant · ES
- UNIVERSITAT POLITECNICA DE VALENCIAparticipant · ES
- UNIVERSITA DEGLI STUDI DI GENOVAparticipant · IT
- THALESparticipant · FR
- HALTIAN OYparticipant · FI
- UNIVERSITA DEGLI STUDI DI MODENA E REGGIO EMILIAparticipant · IT
- BARCELONA SUPERCOMPUTING CENTER CENTRO NACIONAL DE SUPERCOMPUTACIONparticipant · ES
IKERLAN S. COOP is a technology center in the Basque Country, Spain — reach their edge computing or embedded systems division for licensing and partnership inquiries.
Talk to the team behind this work.
Want a one-page brief matching FRACTAL's edge computing platform to your specific industrial use case? SciTransfer can arrange a direct introduction to the right technical contact in the consortium.