Expert Opinions

Why High-Traffic Floors Defeat Most Autonomous Scrubbers, and Which Ones Cope

September 07, 2026

A concourse scrubbed at 02:00 looks untouched by 08:00. Nothing was done badly. The work simply happened at a time that no longer matches when the soil arrives.

In a high-traffic building the binding constraint is not machine power. It is usable minutes. A scrubber rated for 2,000 m² an hour delivers a fraction of that once it is stopping for queues, waiting out crowd surges and working in twenty-minute gaps between peaks.

So the selection question is not which machine is fastest. It is which machine finishes a meaningful amount of floor inside the windows your building actually has, while people are standing on it. Gausium answers that across three site scales with Mira, Marvel and Omnie.

The Gap Between Rated Coverage and Delivered Coverage

Every specification sheet quotes a theoretical maximum measured on an empty floor in a straight line. No commercial building resembles that. Three forces separate the printed figure from the delivered one.

What happens on a busy floor	What it costs	What reduces it
Interruption. The machine pauses for groups, reroutes around queues, waits at temporary barriers	Direct loss of coverage rate, worst at entrances and vertical circulation	Perception that reads crowds early enough to slow rather than stop, plus consistent avoidance behavior
Resoiling behind the machine. The area cleaned twenty minutes ago is dirty again	Uniform routes spend most of their time on floor that did not need attention	Targeted work that treats detected soil instead of repeating a full pass
Wet floor exposure. Scrubbing in opening hours creates a drying window where footfall cannot be controlled	Signage, restricted access and staff supervision, which erase the labor saving	Small localized wet areas and effective water recovery, so signage covers a short span for a short time

A machine suited to high traffic reduces all three. Moving faster reduces none of them.

This is also why area-based buying fails in these buildings. A 40,000 m² warehouse and a 4,000 m² transit concourse can generate the same cleaning workload. What separates them is people per hour, and whether those people are present while the machine works.

One Pass, Because There Is No Time for Two

High-traffic floors carry dry debris and ground-in soil at the same moment. Receipts, packaging, grit and leaf matter arrive with the people. Residue and scuffing build up underneath.

A scrubber alone spreads the debris. A sweeper alone leaves the residue. Running two machines in sequence means two routes, two sets of oversight and two consumable streams for one floor, inside a window that was already too short.

Mira and Marvel sweep and scrub in a single pass. Omnie integrates scrubbing, sweeping and dust mopping in one solution. In an open warehouse the sequential approach is workable. In a concourse at 17:00 it is not.

Why High-Traffic Floors Defeat Most Autonomous Scrubbers, and Which Ones Cope

Matching the Machine to the Building

	Mira	Marvel	Omnie
Site type	Supermarkets, healthcare buildings, light industrial, mid-format retail	Hypermarkets, warehouse retail, large distribution and back-of-house floors	Airports, stations, metro passages, mall common areas, large public buildings
Floor plate	Mid-sized, broken up by fixtures and aisles	Large and open	Large, complex, irregular circulation
Dominant challenge	Aisle width and frequent layout change	Coarse debris volume across distance	Continuous crowd density
Debris profile	Light to moderate, mixed dry and wet	Paper, packaging, thread and bulkier material	Mixed litter, grit, spills, tracked-in moisture
Key capability	Sweep and scrub in one pass; Drop & Go deployment with no professional mapping	Heavy-duty side brush plus disc brush scrubbing in the same pass	Multimodal SLAM with 360-degree 3D LiDAR and panoramic camera for zero-blind-spot awareness
Maintenance load	Internal self-cleaning system flushes the wastewater tank automatically	Same self-cleaning architecture at larger scale	Scrubbing, sweeping and dust mopping in one machine, reducing the number of units on site
Weakest fit	Very large open floor plates	Narrow aisles and tight turning circles	Sites where a simpler machine would do the job

All three are hard-floor machines. A building with carpet at gates, lounges or offices needs the surface-coverage question settled before the scrubber question, because that decision determines whether one machine or two is realistic.

Two deployment references show the pattern in practice. Gausium put Omnie into Milan’s airports ahead of the Milano Cortina 2026 Winter Olympics, working through continuous passenger flow. In retail, Gausium supplied a Greek supermarket chain with 25 Omnie and Phantas robots across its store network, which is the mixed-format version of the same problem.

Five Checks That Decide the Purchase

Check	How to test it	What a pass looks like
Certified for occupied space	Ask which standard the machine is certified against for autonomous operation among people. IEC 63327 is the reference for this equipment class	A named certification, not a general safety statement
Coverage under interruption	Run the demonstration during a real peak, not at 06:00. Compare area completed against the rated figure	You know the ratio for your building before you sign
Behavior, not sensor count	Stand in the route. Watch whether it slows smoothly or brakes hard, and whether people hesitate around it	No congestion forms and no one has to make a deliberate detour
Replenishment between windows	Time a full cycle including water refill, discharge and charging	The machine can start a second run without a person attending it
Response to layout change	Move a display or a barrier and restart the route	It re-plans on its own rather than requiring a technician to remap

The second row is the one most buyers skip. A demonstration on an empty floor tells you almost nothing about a building that is never empty.

Scheduling Against Peaks Instead of Shifts

High-traffic buildings rarely have one cleaning window. They have several troughs, and the troughs differ by zone.

Log a full week of footfall zone by zone, then mark the usable duration of each trough as measured rather than estimated. The real figure is almost always shorter than the assumed one, because arrival and clearance time sit inside it.

Then split the work. Assign full-coverage routes to the deepest trough, usually overnight or pre-opening. Assign targeted work to the shallow daytime troughs, where the objective is to stop soil spreading rather than to treat the whole floor.

Entrance, concourse, retail frontage and back-of-house will not dip at the same time. A single site-wide schedule always leaves some zone uncovered, which is the most common reason a deployment underperforms despite a capable machine.

Gausium’s public transport solutions page sets out how this splits across station and terminal environments.

Where the Answer Is Not a Scrubber

Floors that are mostly carpet or matting. Settle surface coverage first. A scrubber specified into a carpeted terminal will sit idle.

Spaces under constant reconfiguration. Event build-outs, seasonal installs and refits make routes unstable, and automated operation is better paused than fought.

Buildings whose deepest trough is a few minutes long. If no window is long enough to complete anything, the constraint is the operating pattern, not the equipment.

Acoustically sensitive rooms inside otherwise busy buildings. Assess noise in the specific space. The same machine can suit a concourse and fail in a library reading room.

To size any of this against a specific building, Gausium arranges site assessments through the contact page.