TIMESTORM · Project

Giving Robots a Sense of Time So They Work Better With People

digitalPrototypeTRL 4Thin data (2/5)

Imagine a robot that has no sense of whether a second or an hour has passed — it can't tell if you're waiting impatiently or if a task is running late. TIMESTORM studied how the human brain tracks time and then built software models that give robots a similar ability. They installed these time-awareness models on a humanoid robot called ARMAR and tested whether it could plan actions, manage knowledge, and interact with people in a more natural, time-sensitive way. Think of it as giving a robot an internal clock that doesn't just count seconds, but actually understands timing the way we do.

By the numbers

EUR 2,892,500

EU funding for temporal cognition research

research partners in consortium

countries represented (DE, EL, FR, NL, UK)

prototype demonstrations delivered

total project deliverables

The business problem

What needed solving

Robots and AI systems today have no real understanding of time — they execute commands but cannot judge whether something is taking too long, whether a human co-worker is waiting, or how to pace their actions naturally. This makes human-robot collaboration clunky and limits robots to pre-programmed sequences rather than adaptive, real-time teamwork.

The solution

What was built

The team built 8 prototype cognitive and neurocomputational models covering time perception, timely action planning, social self-awareness, and knowledge management over time. The complete model was physically installed on the ARMAR humanoid robot and demonstrated in real-world conditions.

Audience

Who needs this

Collaborative robot (cobot) manufacturers looking to improve human-robot task coordinationAssistive and social robotics companies building robots for elder care or customer serviceAI companies developing autonomous agents that must operate in time-critical environmentsVideo game and simulation companies needing more realistic NPC timing behaviorResearch labs and universities studying cognitive robotics and embodied AI

Business applications

Who can put this to work

Collaborative Robotics

mid-size

Target: Companies building collaborative robots (cobots) for factory floors or logistics

If you are a cobot manufacturer struggling with robots that misjudge human pace and timing during shared tasks — this project developed neurocomputational models of timely action planning that were physically instantiated on the ARMAR humanoid. These models let robots anticipate human timing and coordinate actions more naturally, which could reduce task completion delays in human-robot teams.

Elder Care & Assistive Robotics

SME

Target: Assistive robotics companies or care home operators deploying social robots

If you are a care robotics provider where robots must interact with elderly users who move and respond at varying speeds — this project built cognitive models of social self and temporal cognition tested in real-world conditions. Robots equipped with this technology can adapt their pace and responses to individual users, making interactions feel less mechanical and more patient.

AI Software & Autonomous Systems

any

Target: AI companies developing autonomous agents for dynamic environments

If you are an AI company whose autonomous agents fail in time-critical environments because they lack temporal reasoning — this project produced 8 prototype cognitive models covering time perception, action planning, and knowledge management over time. These models could be integrated into your AI stack to give agents a genuine sense of timing in unpredictable, real-world settings.

Frequently asked

Quick answers

How much would it cost to license or access this technology?

No pricing or licensing terms are publicly available. The project was funded with EUR 2,892,500 in EU contribution under a Research and Innovation Action, so the IP is likely held by the 6 academic and research partners. Licensing terms would need to be negotiated directly with the coordinator.

Can this work at industrial scale, for example on production lines with many robots?

Based on available project data, the models were demonstrated on a single ARMAR humanoid robot in real-world conditions. Scaling to multi-robot industrial environments was not part of the project scope. Significant engineering work would be needed to move from one research humanoid to fleet deployment.

Who owns the intellectual property and can I get a license?

The consortium of 6 partners across 5 countries (DE, EL, FR, NL, UK) likely shares IP under the Horizon 2020 grant agreement. The coordinator, IDRYMA TECHNOLOGIAS KAI EREVNAS in Greece, would be the first point of contact for IP discussions.

What exactly was demonstrated and how mature is it?

The team delivered 8 prototype demonstrations including cognitive models of time perception, action planning, social self, and knowledge management over time. The full model was physically instantiated on the ARMAR humanoid and assessed in real-world conditions. These are research prototypes, not production-ready systems.

How would this integrate with our existing robot control systems?

The project produced computational models based on human brain mechanisms, not plug-and-play software modules. Integration with commercial robot platforms would require adaptation from the neurocomputational model format used with ARMAR to your specific control architecture. Based on available project data, no commercial integration guides or APIs were produced.

Are there any regulatory considerations for robots with time perception?

Based on available project data, the project did not address regulatory compliance. However, robots with more human-like cognitive abilities may face additional scrutiny under emerging EU AI Act requirements, particularly for systems interacting directly with people in care or workplace settings.

Consortium

Who built it

The TIMESTORM consortium is purely academic — 5 universities and 1 research organization across 5 countries (Germany, Greece, France, Netherlands, UK), with zero industry partners and zero SMEs. This is a red flag for near-term commercialization: there was no company at the table to pull the technology toward a real product. The coordinator is a Greek research foundation. For a business looking to adopt this, expect to invest significant effort in technology transfer, as none of the 6 partners are set up for commercial delivery. The multi-country academic spread does mean the research is well-validated across different labs, but the path from lab prototype to your factory floor is entirely unbridged.

IDRYMA TECHNOLOGIAS KAI EREVNASCoordinator · EL
IMPERIAL COLLEGE OF SCIENCE TECHNOLOGY AND MEDICINEparticipant · UK
THE UNIVERSITY OF SUSSEXparticipant · UK
KARLSRUHER INSTITUT FUER TECHNOLOGIEparticipant · DE
UNIVERSITE CLERMONT AUVERGNEparticipant · FR
RIJKSUNIVERSITEIT GRONINGENparticipant · NL

How to reach the team

Coordinator is IDRYMA TECHNOLOGIAS KAI EREVNAS (FORTH) in Greece — a major research foundation. Contact their technology transfer office for licensing inquiries.

Order a report on this consortium See IDRYMA TECHNOLOGIAS KAI EREVNAS profile →

Next steps

Talk to the team behind this work.

SciTransfer can broker an introduction to the TIMESTORM research team and help you evaluate whether their temporal cognition models fit your robotics or AI product roadmap. We handle the academic-to-business translation so you don't have to.

Order a report on this topic More in Digital & ICT