UCoCoS · Project

Software Tools to Predict and Control Complex Networked Systems Like Grids and Robot Fleets

digitalPrototypeTRL 3Thin data (2/5)

Imagine you have thousands of devices — robots, sensors, power stations — all talking to each other, and the whole system behaves in ways none of the individual parts would suggest. It's like a traffic jam: no single driver causes it, but the interactions create gridlock. UCoCoS built mathematical methods and software to understand these emergent behaviors and, crucially, to design controls that steer the whole network toward what you actually want. They validated their approach on real mobile robots and sensor networks.

By the numbers

academic partners across 3 countries

total project deliverables

industry partner ratio in consortium

demo deliverables including software package and robot experiments

The business problem

What needed solving

Companies running interconnected systems — robot fleets, sensor networks, power grids, logistics chains — face a growing problem: the interactions between components create unpredictable collective behaviors that no single-component analysis can explain or fix. Traditional engineering approaches that reduce a system to its parts fail when the global dynamics are driven by connections, not components. This leads to cascading failures, coordination breakdowns, and systems that resist optimization.

The solution

What was built

The project delivered a publicly available software package integrating control-oriented tools for complex networked systems (hosted at KU Leuven). They also completed experimental validations: controller design tools tested on a mobile robotics platform at ECL-Inria, and robot-sensor network experiments integrated in the Inria Equipex FIT IoT-Lab. In total, 12 deliverables were produced across the consortium.

Audience

Who needs this

Industrial robot fleet operators needing coordination algorithms for multi-robot systemsPower grid operators managing stability across interconnected networksIoT platform companies deploying large-scale sensor networksAutonomous vehicle companies solving multi-agent coordinationLogistics companies optimizing interconnected warehouse and delivery systems

Business applications

Who can put this to work

Industrial Automation & Robotics

enterprise

Target: Manufacturers operating fleets of collaborative robots on production lines

If you are a factory operator struggling to coordinate fleets of collaborative robots that behave unpredictably when working together — this project developed controller design tools validated on mobile robot platforms that help predict and manage the collective behavior of interconnected machines. The software was experimentally tested at Inria's IoT-Lab facility on real robot-sensor network scenarios.

Energy & Utilities

enterprise

Target: Power grid operators and energy network managers

If you are a grid operator dealing with cascading failures or instability in interconnected power networks — this project created control methods specifically aimed at understanding how interactions between network components drive global dynamics. Their publicly available software package can help model and stabilize complex grid behavior before problems cascade.

Telecommunications & IoT

mid-size

Target: Companies managing large-scale sensor networks or IoT deployments

If you are an IoT platform provider whose sensor networks produce emergent behaviors that are hard to predict or manage — this project built and tested tools on a real robot-sensor network scenario integrated in the Inria Equipex FIT IoT-Lab. Their software package offers methods to design network structures and policies that keep large interconnected systems under control.

Frequently asked

Quick answers

What would it cost to license or use this software?

The project produced a publicly available software package integrating tools developed across the consortium. As an output of a publicly funded Marie Curie training network hosted at KU Leuven, the software is likely available under open or academic licensing terms. Contact KU Leuven directly for commercial licensing details.

Can these tools work at industrial scale — thousands of nodes?

The methods were designed for complex networked systems like the internet and power grids, which inherently operate at large scale. However, experimental validation was performed on mobile robot and sensor network scenarios at lab scale (Inria's FIT IoT-Lab). Scaling to full industrial deployment would require additional engineering and testing.

Who owns the intellectual property?

IP is held by the 3 academic partners: KU Leuven (Belgium), and partners in France and the Netherlands. As an MSCA-ITN training network, the primary outputs are trained researchers, publications, and open software. Commercial IP terms would need to be negotiated with KU Leuven as coordinator.

Is this ready to deploy in a production environment?

Not yet. The project delivered experimental validation on robot platforms and a software package, but this was a research training network that ran from 2016 to 2020. The tools would need significant engineering to move from lab-validated methods to production-grade software.

What exactly was tested and where?

Two experimental validations are documented: controller design tools tested on a robotics platform at ECL-Inria (France), and a robot-sensor network scenario integrated in the Inria Equipex FIT IoT-Lab. These demonstrate proof-of-concept, not industrial deployment.

Does this comply with industry standards for control systems?

Based on available project data, the tools focus on mathematical control methods for complex networks. No specific industry certification or compliance testing (e.g., IEC 61508 for functional safety) is documented in the deliverables. Industrial adoption would require standards alignment.

Consortium

Who built it

This is a purely academic consortium — 3 universities in Belgium, France, and the Netherlands with zero industry partners. While the academic expertise in control theory, computer science, and mechanical engineering is strong, the absence of any industrial participant means the results have not been stress-tested against real business requirements. The project was funded as a Marie Curie training network (MSCA-ITN-EJD), meaning its primary purpose was training early-stage researchers, not developing commercial products. Any company interested in these tools should expect to invest in adapting the research outputs for their specific industrial context.

KATHOLIEKE UNIVERSITEIT LEUVENCoordinator · BE
CENTRALE LILLE INSTITUTparticipant · FR
TECHNISCHE UNIVERSITEIT EINDHOVENparticipant · NL

How to reach the team

KU Leuven, Belgium — reach out to the Department of Computer Science or the NUMA research group for commercial inquiries about the software package.

Order a report on this consortium See KATHOLIEKE UNIVERSITEIT LEUVEN profile →

Next steps

Talk to the team behind this work.

SciTransfer can help you evaluate whether UCoCoS methods fit your network control challenges and connect you with the right researchers at KU Leuven.

Order a report on this topic More in Digital & ICT