SpatioTemporal is proud to be a Founding Partner of two new initiatives designed to help solve one of the biggest problems facing robotics and Physical AI: proving what robots can actually do in the real world.
The Robot Benchmark will create an independent framework for measuring and comparing robot capability.
The Robot Range will provide the real-world proving ground where those capabilities can be tested.
Together, they are being built to create something the emerging robotics industry badly needs: evidence people can use.
Robotics has a proving problem
Robotics is advancing quickly. Humanoids, mobile manipulators, autonomous systems and increasingly capable Physical AI models are moving from research environments towards warehouses, hospitals, workplaces, homes and public spaces.
But evaluating them remains surprisingly difficult.
Manufacturers have their own demonstrations and specifications. Research benchmarks tend to test narrow capabilities or controlled tasks. Buyers have few reliable ways to compare one system against another — or to understand whether a robot that performs well in a demo can perform reliably in the messy conditions of the real world.
That becomes a commercial problem very quickly.
Robots are substantial capital purchases. Yet buyers are still routinely being asked to accept manufacturers’ claims about what those machines can do, how reliably they can do it, and how they will behave once deployed into a real workplace.
We don’t accept that equation for cars. We don’t accept it for industrial equipment or even household appliances. Performance is tested against standards, independently measured and made comparable.
A robot can cost as much as a car – and operate directly alongside people – yet the buyer is often left to discover for themselves whether it works, whether it works in their environment, and whether it can be trusted to keep working when conditions change.
That isn’t enough evidence on which to build a market.
And increasingly, the important questions are not simply can the robot complete the task?
They are questions like:
- Can it do it repeatedly?
- Can it cope when the environment changes?
- Can it work around people?
- Can it recover when something goes wrong?
- How much human intervention does it require?
- And ultimately: would you trust it to do the job?
Those questions need better answers.
The Robot Benchmark
The Robot Benchmark is being established as an independent, international benchmarking initiative for robotics and Physical AI.
Its purpose is not to declare which robot is “best”. It’s to create a common language for describing and measuring robot performance through three fundamental questions:
Capability
Can it do it?
Does the robot possess the physical, sensory and functional abilities it claims?
Competence
Can it do it well?
Can it perform reliably and repeatedly under variation, interruption and uncertainty?
Conduct
Can we trust how it behaves?
How does it behave around people, other machines and unexpected events?
Together, these Three C’s provide the structure through which robot performance can be tested, measured and compared.
The ambition is to develop benchmarks that are useful beyond the laboratory.
- Manufacturers should be able to demonstrate capability.
- Researchers should be able to measure progress.
- Buyers should be able to compare machines against the jobs and environments they have in mind.
- Insurers, regulators, governments and investors should have stronger evidence on which to make decisions.
The Benchmark is being designed as shared infrastructure for the industry, developed with the organisations that will ultimately need to trust and use it.
The Robot Range
Benchmarks also need somewhere to meet reality. That’s the role of The Robot Range: a physical proving ground where robots can be tested against repeatable tasks, environments, obstacles, human interactions and edge cases.
Rather than a pristine robotics laboratory, the idea is deliberately closer to the environments in which robots will eventually have to operate.
Real floors. Real doors. Real objects. Real people. Changing conditions. Unexpected events.
Some tests may measure straightforward task performance. Others may explore dexterity, navigation, robustness, recovery, human interaction or increasingly complex combinations of capabilities.
The important part is that testing becomes repeatable, observable and comparable.
The Robot Benchmark defines what matters.
The Robot Range helps prove it.
Why SpatioTemporal is involved
At SpatioTemporal, our work is focused on the trust layer for Robotics and Physical AI: helping machines understand human movement, intent and consequence so they can behave more naturally and predictably around people. That makes this challenge particularly important to us.
We believe robotics adoption will depend on more than increasingly impressive capabilities. It will depend on whether manufacturers, workplaces and ultimately the public can trust those capabilities. And trust requires evidence.
Independent testing also creates something enormously valuable for the wider Physical AI ecosystem: a way to understand not just whether new technologies improve a robot, but how much they improve it, under what conditions, and where they still fail.
That is why SpatioTemporal has joined The Robot Benchmark and The Robot Range as a Founding Partner.
Not to build a benchmark around our technology. Quite the opposite.
We want to help create independent infrastructure against which ours – and everybody else’s – can be tested.
This needs to be built together
Neither initiative should belong to one robotics company, one research institution or one technology vendor. For them to matter, they need participation from across the ecosystem.
We are now assembling the group that can help build them. That means:
- Robot manufacturers and Physical AI companies.
- Researchers and universities.
- Simulation, compute and infrastructure partners.
- Testing, standards, safety and assurance organisations.
- Governments and industry bodies.
- And critically, the organisations that will eventually buy, deploy, insure and rely upon robots.
SpatioTemporal is the first Founding Partner to publicly put its name behind the initiatives. We don’t intend to be the last.
If your organisation belongs at that table – or you know somebody we should be talking to – we’d like to hear from you.
The Robot Benchmark will set the standard. The Robot Range will prove it.