Service Robot Site Survey and Facility Readiness Assessment

Why the Site Survey Decides the Project

Procurement teams spend weeks comparing payload, battery chemistry and sensor suites, then send one person to the site for a ten-minute glance. That ordering is backwards. The robot specification tells you what a unit can do; the site survey tells you whether this building will let it. In our own deployment history the single most common cause of a stalled rollout is a physical constraint that was visible on day one and never recorded: a fire door that closes under spring tension, a carpet lip at a threshold, a mirrored lobby wall that breaks a lidar return.

A proper survey is not a sales visit. It is a measurement exercise with a pass/fail sheet, run before the purchase order, by someone who will not be blamed for the robot. Treat the output as an engineering document: raw numbers, photographs with a scale object, and a scored conclusion. If a vendor will not run this survey before you sign, that is itself a finding.

This matters commercially because a site that fails readiness after delivery does not fail cheaply. Docked units still incur the RaaS subscription, the depot and spare-unit ratio is already committed, and staff who were trained on a promise lose confidence that is slow to rebuild. For the cost arithmetic on a stalled rollout, see the downtime and OEE cost model.

The Measurements to Take and Why Each One Matters

Bring a laser distance meter, a tape, a phone camera, a spirit level app and a floor-tile marking pen. Record every number against a named location, not "the main corridor". The table below is the minimum set for a floor-cleaning or delivery deployment.

Measurement	How to take it	Typical minimum	Why it fails the robot
Clear doorway width	Narrowest point between jamb and any door-stop or closer arm	Unit width + 150 mm each side	Door closers and fire-door arms reduce clear width by 40-90 mm
Corridor width	Wall to wall at the tightest pinch point	Unit width + 300 mm for passing	Below this the unit must single-lane and loses throughput
Threshold height	Vertical step at every floor-material change	Under 15 mm ramped, under 8 mm square	Square lips over 15 mm stall a caster; 20 mm+ needs a ramp
Floor transition type	Photograph with a coin for scale	Hard, fixed, no loose mats	Loose anti-fatigue mats and curled vinyl defeat wheel odometry
Lift car depth	Inside car, door to back wall	Unit footprint + 200 mm	A delivery robot with open shelves needs turning clearance to back out
Ramp gradient	Rise over run, as a percentage	Under 8% for laden units	Dock ramps over 10% drain battery and risk load shift
Turning-circle pinch	Radius available at the tightest junction	Unit length + 100 mm	Determines whether the route needs a reverse manoeuvre

Take each door measurement with the door in its closed resting position, not held open. A fire door on a spring closer presents a narrower clear opening than the frame suggests, and a robot running a night route will meet it closed. The same applies to the lift: measure with the doors open at their widest travel, because a partially open car door can clip a side shelf on a wide unit.

Floor and Surface Checks

Floor condition drives more real-world failures than any sensor specification. Run a three-part check:

1. Slip and grip. A scrubber recovers the water it dispenses through a squeegee. On a highly polished stone floor with a tight squeegee angle, recovered water can pool momentarily and reduce traction on the drive wheel. Note any area with a visible sheen and test the unit's default water volume there before committing to a schedule.
2. Surface continuity. Walk the intended route and photograph every change of material: tile to carpet, carpet to stone, stone to metal grating. Each transition is a potential stall point. Loose-lay vinyl and curling cable covers are the worst offenders.
3. Obstruction density. Count freestanding obstacles per 100 m² — display stands, planters, bin stations, stanchions. A dense layout does not disqualify a robot, but it changes the route plan from sweeping passes to a tighter grid, which raises cycle time. Feed the count into the throughput calculation rather than discovering it on day one.

For buildings where the floor is also the highest-value asset, the same survey feeds the preventive schedule; the interval arithmetic is set out in the preventive maintenance schedule.

Network, Power and the Infrastructure Sign-Off

Physical readiness is half the survey. The other half is the infrastructure the robot's fleet software runs on, and it is usually owned by a different department than the one buying the robot. Capture it in the same visit so nobody discovers the gap later.

• Wi-Fi coverage. Walk the route with a phone on the facility SSID and record signal strength in dBm at the ten weakest points. Anything below −70 dBm is a candidate dead zone for teleoperation and telemetry. Robots can navigate offline, but a robot that cannot be remotely supervised during an incident will be parked. The full IT checklist is in the Wi-Fi and network connectivity guide.
• Charging point availability. Confirm a cabled supply at the planned dock, sized for the fleet, not the first unit. The electrical arithmetic is in the fleet charging infrastructure plan.
• Lift and door integration. Establish who controls the lift interface and whether a handover protocol already exists. Multi-floor routes live or die on this; see multi-floor deployment navigation layers.
• Security and data. Identify the network the fleet controller will sit on, and whether the security team will accept it. Flag the review early rather than at commissioning.

The Readiness Scoring Sheet

Turn the raw measurements into a single decision. Score each dimension pass, conditional or fail, and set the rule that any fail blocks the order until resolved. This converts an informal site visit into a document both sides can sign.

Dimension	Pass	Conditional	Fail
Doorways and corridors	All clear widths above minimum	1-2 pinch points, ramps or route changes fixable	Core route physically impassable
Floors and thresholds	Continuous, fixed, under threshold limits	Isolated lips needing ramps	Widespread loose surfaces or steps over 20 mm
Vertical transport	Lift handover agreed or single floor	Interface in progress with a named owner	No agreement and no owner
Network	SSID covers route above −70 dBm	1-2 dead zones with a remediable fix	No usable network on the route
Power	Dock supply sized for the fleet	Upgrade scheduled before delivery	No supply available at any viable dock
Stakeholder sign-off	Facilities, IT and security all engaged	One function not yet consulted	Security rejects the fleet controller outright

Weight the sheet to your own risk: a single-floor site can ignore vertical transport entirely, while a hospital or a multi-tenant tower should treat the lift interface as a hard gate. The point is not the specific weights; it is that the decision is made on recorded numbers before the order, not on optimism after it.

What a Good Survey Deliverable Looks Like

The survey is finished when you can hand over a single document containing: a marked-up floor plan with the proposed route; a photograph log with scale references for every doorway, transition and dock; a raw measurements table; the filled scoring sheet; and a short list of remediation items with an owner and a target date for each. If the sheet says conditional, no purchase order should be raised until every remediation item is closed.

Run the survey in daylight and, if the site will operate at night, run it once after dark as well. Lighting levels, reflective floors under artificial light, and closed fire doors change the picture materially. A survey run only at 11 a.m. in a bright lobby tells you nothing about the 2 a.m. reality the robot will actually work in.

Get a Quote

Tell us about your site and workload and we will size the fleet, the accessories and the service plan around your actual operation. Request a quote or email [email protected].