Short answer: an atrium or canopy glass cleaning robot works indoors when the glass is reachable from a flat, load-rated surface and the machine can be tethered to a fixed anchor. It fails on three things: glass that sits over people with no safe catch, framing that interrupts continuous runs, and dust loads that are lighter than outdoor roofs but far more visible. Indoor canopies are not easier versions of skylights — they are a different job with different rules, and buyers who treat them as the same job usually regret the machine they chose.
What makes indoor atrium glass different from a skylight?
The dust is different, and that drives most of the cost. An outdoor skylight collects rain-washed mineral film, bird residue and industrial fallout, and rain resets the surface every few weeks. An indoor atrium collects carpet fibre, skin cells, printer and kitchen particulate, and construction dust from any fit-out happening above or below. None of it gets rained off. It sits, it settles into the frame rebates, and it keeps arriving.
The result is a lighter but constant load. Machine cleaning times per panel drop, but the frequency goes up. A roof skylight might be on a quarterly cycle; a busy atrium with a food outlet below is often on a fortnightly cycle, and the machine hours per year end up higher than the roof case. Budget the annual cost on cycle count, not on per-panel speed.
The second difference is visibility. Indoor glass is usually below the sightline of the people who pay for the building. A haze that nobody would notice on a roof is a complaint in a lobby, and it shows worst under low-angle afternoon sun through the glazing. Machines that leave a fine film of dirty water — the common failure mode of any spray-and-squeegee pass — look worse indoors than outdoors, because the film dries on glass that never gets rain to wash it off.
The third difference is access to the glass itself. Roof skylights are usually reached from the roof. Atrium glass is reached from the floor below, from a catwalk, or not at all. That single fact decides whether a robot is viable before any specification question is asked.
How does fall protection work on an indoor canopy?
Differently, and this is where most indoor projects go wrong. On a roof, a machine that loses suction falls onto a roof surface, which is bad but survivable. Over an atrium, a machine that loses suction falls into an occupied space. That means the machine needs a secondary retention system — a tether or safety line rated for the drop, anchored to a structural point, not to the glazing frame — and the anchor has to be engineered, inspected and documented. On many atrium projects there is no structural point above or beside the glass at all, only the mullion system, which is not an anchor.
The practical solutions are a horizontal lifeline installed during a maintenance programme, a beam trolley on a permanent rail, or restricting the machine to glass that sits over a catch surface such as a roof podium or a low-level flat glazing band. If none of those is available, the machine is not appropriate no matter how well it cleans.
There is a second question: the people below. Even a tethered machine can drop tools, brushes or spray, and indoor work usually runs during building hours, so the programme needs a drop-zone exclusion under the working area — for a large atrium, part of a lobby or a staircase. Our notes on facade and high-level access planning set out how that exclusion zone is normally sized and who has to sign it off.
The edge cases are the ones that catch projects late: glass that runs out at a wall, a change of plane over a stair void, a canopy that overhangs a security desk. These are where a machine stops and a person has to step in, and our page on crossing obstacles and edges covers the same problem on vertical facades — the geometry is different, the constraint is identical.
What should an indoor robot specification include?
Start with reach and geometry rather than cleaning performance. Measure the longest continuous glass run, the narrowest gap between frame members, the pitch, and the height of the highest panel from the surface the operator will stand on. Most indoor atrium failures are geometry failures: the machine fits the panels the vendor demonstrated on and does not fit the end bays.
Then look at water handling. Indoors there is no drainage, so the machine should be able to run on low water volumes, and the operator needs a plan for the wastewater that runs down the glass. On a roof this is dismissed; in a lobby it is a slip risk and a carpet claim. Machines that recirculate water or wipe dry with a blade in contact are strongly preferred over a finished floor.
Noise matters more than buyers expect. An atrium carries sound. A machine that is acceptable on a roof can be unusable in a quiet lobby at 10am, which pushes the schedule into the night shift and changes the labour cost. Ask for a measured sound level at one metre and test it in the actual space before committing.
Finally, power. Indoor machines are almost always cabled, and the lead must reach a safe outlet without crossing a fire route, a public corridor or a wet floor. Cable runs longer than the supplied lead, daisy-chained extension leads, and leads crossing egress paths are the three most common commissioning blockers on indoor jobs. Our notes on field reviews of skylight robot setups show the same pattern repeating across sites.
Is an indoor canopy robot worth it compared with a manual team?
It depends on height and on continuity. A low-level atrium band two or three metres up is manual work forever — a pole, a squeegee and twenty minutes. No machine pays back there. The crossover opens when the glass is above roughly six to eight metres and the runs are long and continuous, because that is where manual work needs scaffolding, a scissor lift or rope access inside a building, all of which are slow and disruptive.
Compare the annual cycle cost, not the single visit. Take panels, multiply by cycles per year, add setup per visit, and compare against the lift hire or rope access you would otherwise pay for. On a large atrium with a high dust load and a fortnightly cycle, the machine usually wins decisively. On a small canopy with a quarterly cycle, the lift hire is cheaper and the machine sits idle — and idle machines are the ones that fail at the next service because nobody has kept the seals wet and the brushes clean.
For most buildings a split works best: machine on long high runs, manual on short bays. Plan for that from the start rather than treating the machine as a full replacement.
Key Takeaways
- Indoor atrium glass carries a lighter but constant dust load, so cycles are more frequent and annual machine hours are often higher than on a roof.
- A machine over occupied space needs a rated tether and a structural anchor, not the glazing frame. If no anchor exists, the machine is not viable.
- Budget an exclusion zone under the working area, and expect it to close part of a lobby or stair.
- Low water, dry-wipe and measured noise levels matter more indoors than cleaning speed does.
- Below about six to eight metres, manual still wins. Above that, with long continuous runs, the machine pays back — plan for a split programme rather than a full replacement.