An unofficial concept study. Enter the passphrase to view.
Keeping a field healthy means putting real research problems in front of the next generation early. I build open, accessible on-ramps that do exactly that.
An international competition I chair that puts open research problems in response robotics in front of high-school students. Now in its second decade, running at the RoboCup world championships.
→ Explore the RMRCOpen-source, 3D-printable robot designs so students, researchers, and educators can build a functioning rescue robot from a common set of parts, then publish their own so others can build on them. All designs, instructions, and source code are released under an open-source licence.
I have been part of RoboCup since 2003 and a Trustee since 2021. Its stated goal, a team of robots able to beat the human World Cup champions by 2050, is deliberately audacious. A grand challenge of that scale pulls an entire research community forward and gives students a problem worth pursuing.
The same instinct runs through the standard test methods: turn a hard problem into a course you can score, and people will come to solve it.
Make the frontier buildable. Open designs, cheap parts, published results.
Teams from multiple countries, every year, building rescue robots small enough to hold.
Beyond the competitions, the goal is a healthier research ecosystem. Open designs and shared results lower the cost of entry, standard test methods make progress measurable, and a steady pipeline of new researchers keeps the field moving. Underneath it all is a consistent argument: without trust, explainability, and rigorous measurement, over-hyped AI risks another "AI winter." Standards and outreach are how that is avoided.