At the University of Queensland, a four-legged robot named QiuQiu has begun the quiet, methodical work of asking whether a campus truly belongs to everyone who walks it. Equipped with laser scanners and cameras, the machine traversed a real pedestrian route, measuring slopes and surfaces that determine whether mobility is a right or a privilege. The effort, led by UQ's VisLab, is less about the novelty of the robot than about what it reveals — that accessibility is not a feeling or a policy, but a geometry, and geometry can be measured.
UQ VisLab deploys robot dog to map campus accessibility
Accessibility is measurable. It is spatial. It is something a robot can help us see.
Mark
So they sent a robot dog to walk around campus. What's actually the point of that? Why not just have someone walk it and take notes?
Mimi
Because a robot collects data continuously and consistently. It records its own movement, the slope of the ground, obstacles—all in precise spatial terms. A person taking notes is making judgments about what matters. A robot captures everything.
Luke
But the handheld scanner is also there, right? So they're not replacing human observation. They're comparing two methods.
Mimi
Exactly. The robot gives you automated, continuous measurement. The handheld scanner gives you the ability to focus on specific features. Together, they create a more complete picture than either alone.
Mark
And this matters for accessibility because...?
Mimi
Because accessibility is about the actual physical demands of moving through space. A ramp that's too steep, a slope that's too long—these aren't opinions. They're measurable facts. If you can map them precisely, you can make better decisions about where to build, where to fix things, where people might struggle.
Luke
The route they tested was 790 metres in thirteen minutes. That's one specific path. How much does that actually tell you about the whole campus?
Mimi
It's a proof of concept. They're showing the method works, that you can gather this kind of data reliably. Once you know it works, you can scale it.
Mark
What happens with the data once they have it?
Mimi
It feeds into a Digital Twin—a virtual model of the campus. If you can layer accessibility information into that model, along with infrastructure data, you have a tool for planning that's based on evidence, not guesswork.
Luke
Has that Digital Twin been built yet?
Mimi
Not yet. This is early work. But that's the direction they're heading.
Le Pouls
- A robot dog is walking university footpaths not as a spectacle but as a surveyor, recording the precise gradients that separate passable from impassable for wheelchair users and stroller-pushers alike.
- A single ramp near the route's start rises at a 1:9 gradient — a number that sounds technical until you are the one who cannot climb it.
- Two methods of measurement — robotic motion capture and human-guided LiDAR scanning — are being run in parallel, because neither machine nor person alone sees the full picture.
- The data feeds an ambition larger than one walk: a Digital Twin of the entire campus that could make invisible barriers visible before they are ever built.
- The findings are still being analyzed, but the underlying argument is already made — accessibility is spatial, it is measurable, and ignoring it is a choice, not an oversight.
At the University of Queensland, a four-legged robot named QiuQiu has begun the quiet, methodical work of asking whether a campus truly belongs to everyone who walks it. Equipped with laser scanners and cameras, the machine traversed a real pedestrian route, measuring slopes and surfaces that determine whether mobility is a right or a privilege. The effort, led by UQ's VisLab, is less about the novelty of the robot than about what it reveals — that accessibility is not a feeling or a policy, but a geometry, and geometry can be measured.
At the University of Queensland's St Lucia campus, a quadruped robot named QiuQiu has been put to work on a problem older than robotics: does this path actually work for everyone who needs to use it? The Unitree Go2 machine, fitted with cameras and LiDAR scanners, walked roughly 790 metres between the VisLab and the ferry terminal — a route chosen because real people depend on it daily — recording its own movement, orientation, and a continuous picture of the terrain around it.
The VisLab team, led by Dr Jurij Karlovsek and Assoc Prof Fred Fialho Teixeira alongside several colleagues, designed the trial as a comparison exercise. While QiuQiu walked the route autonomously, a researcher followed on foot carrying a handheld LiDAR scanner, producing a second dataset from the same path. A real dog named Milo, apparently conducting his own assessment, accompanied them. By running the route multiple times in both directions, the team built overlapping records — the same ground measured by different means, each method catching what the other might miss.
What the data revealed was not dramatic, but it was precise. The route descends about seven metres overall, mostly gently — yet near the start, a ramp climbs at roughly a 1:9 gradient, steep enough to present a genuine obstacle for wheelchair users or anyone pushing a heavy load. These are the kinds of details that rarely appear on campus maps but determine, in practice, who can move freely and who cannot.
The work is a step toward something larger. UQ is developing a Digital Twin of the St Lucia campus — a virtual model of the physical environment — and accessibility data could be embedded within it, creating a planning tool that makes mobility barriers visible and addressable before they harden into permanent features of the built environment. For now, QiuQiu's thirteen-minute walk is a proof of principle: that accessibility is not a matter of intention or compliance, but of measurable space — and that measurement is where change begins.
At the University of Queensland's St Lucia campus, a four-legged robot named QiuQiu has taken on the unglamorous work of measuring whether the ground beneath students' feet is actually walkable. The machine, a Unitree Go2 quadruped equipped with cameras and laser scanners, spent thirteen minutes covering roughly 790 metres between the VisLab and the ferry terminal—a route chosen specifically because it matters to real people moving through campus every day.
The VisLab team, led by Dr Jurij Karlovsek and Assoc Prof Fred Fialho Teixeira alongside Dr Qianbin Xu, Dr Thien-Minh Nguyen, Dr Tyson Phillips, Rui Zhang, and Mohan Ren, designed the trial to answer a deceptively simple question: what can a robot tell us about accessibility that we might otherwise miss? QiuQiu walked the route while recording its own pose, orientation, and movement data. A 360-degree camera mounted on its back captured the environment in all directions. A forward-facing camera recorded what lay ahead. Simultaneously, a researcher walked the same path carrying a handheld LiDAR scanner, creating a second dataset to compare against the robot's measurements. A real dog named Milo accompanied them, conducting what the team described as his own survey of the route.
What emerged from the data was a map of the terrain's actual demands. The route proved mostly gentle, with a total descent of about seven metres over the full distance. But gentleness is not uniformity. Near the start, a ramp rose at roughly a 1:9 gradient—steep enough to matter to someone in a wheelchair or pushing a stroller. Beyond that, the path descended gradually, a long slope that accumulates over distance. These are not abstract measurements. They are the physical reality that determines whether a route is passable for everyone or only for some.
The comparison between robotic and handheld scanning revealed something important: each method captures different information. The robot's continuous motion data and automated sensors record what happens as it moves through space. The handheld scanner, guided by human judgment, can focus on specific features and capture details a robot might miss. By running the route multiple times in both directions, the team could build comparable datasets—the same path measured different ways, creating a richer picture than either method alone could provide.
The work points toward a larger ambition. The university is developing a Digital Twin of the St Lucia campus, a virtual model that mirrors the physical space. Accessibility data—the slopes, the ramps, the ground conditions—could be woven into that model alongside infrastructure information, creating an interactive tool for understanding how the campus actually functions for people with different mobility needs. Such a model could support decisions about future construction, renovation, and maintenance. It could identify problem areas before they become barriers. It could provide evidence for planning that goes beyond assumption or anecdote.
For now, QiuQiu's walk between the VisLab and the ferry terminal is a beginning. The data is still being analyzed. But the principle is clear: accessibility is not a matter of good intentions or compliance checklists. It is measurable. It is spatial. It is something a robot can help us see.
Citations marquantes
A particular focus is the relationship between movement and accessibility. As QiuQiu walks, it records its pose, orientation and odometry, while its LiDAR scans the ground ahead and behind, revealing changes in slope and terrain along pedestrian routes.— University of Queensland VisLab research description
Envie de l'histoire complète ?
Lire l'original sur Spatial Source ↗
Education