Airport Robots Usually Stop at the Security Zone Line

The Three Zones an Airport Robot Has to Live In

A US commercial airport is divided into zones with different access rules, and a robot has to be approved for one of them. Landside is everything before the checkpoint: kerbside, check-in halls, baggage claim, car parks. The sterile area is the departure side past screening, where passengers wait at gates. Airside covers the ramp and aircraft movement areas, most of which the airport designates a Security Identification Display Area (SIDA) under 49 CFR Part 1542.

Those rules were written for people, and that is the awkward part of a robot project. A worker gets unescorted SIDA access only after a fingerprint-based criminal history check, a security threat assessment and SIDA training. The airport operator administers all three under Transportation Security Administration requirements. A machine cannot hold a badge or sit the training. An airside robot is therefore approved as equipment inside an operator's own security programme. That means a named person accountable for it, plus a written procedure for what happens when it stops.

This is the real gate on the project. Two robots with identical sensors and software can be months apart on delivery purely because one crosses the checkpoint. Ask which zone a quoted timeline assumes. A landside pilot tells you almost nothing about an airside one.

Floor Cleaning at Terminal Scale

Terminal cleaning is the most established robot job in aviation, because the floor area is enormous and the route repeats. Cincinnati/Northern Kentucky International Airport was the first US airport to run the Avidbots Neo scrubber. Avidbots' own CVG account describes cleaning plans mapped across up to 750,000 square feet, with roughly 200,000 square feet cleaned per week during the pilot. The machine runs about six hours between charges.

Work out the shift arithmetic before the square footage impresses anyone. A scrubber covering 200,000 square feet a week is doing the walking part of a cleaner's job. Toilets, seating, glass and spills still need a person, so the saving lands as fewer hours pushing a manual scrubber rather than fewer people on the contract.

Airports break cleaning robots in one specific way. There is no quiet window. A shopping centre closes and a hospital ward calms at night, while a hub terminal runs red-eye departures and early arrival banks. The machine works among moving passengers for most of its duty cycle and yields constantly. Ask for productivity figures measured in an occupied terminal, because a number from an empty hall after midnight will not survive a 5am queue. The machine specifications themselves are covered in the cleaning robots guide.

Wayfinding and Passenger Information

Passenger information is the job airports pilot most visibly. The current example is at San José Mineta International Airport, which launched a four-month pilot of José on 1 April 2026. The humanoid, built by IntBot, stands in Terminal B near Gate 24. It answers questions, gives directions and reads out live terminal information. The airport says it works in more than 50 languages. The Robot Report notes the pilot is timed around the 2026 FIFA World Cup, when the airport expects heavy international arrivals.

Note where that robot stands. Terminal B near a gate sits inside the sterile area, past screening, where passengers are already checked in and waiting. That is the easiest audience a wayfinding machine will get, and the placement most likely to produce big interaction counts. Put the same robot kerbside and it meets people who are late and walking with purpose.

Language coverage separates aviation from every other public building, and the wider front-of-house case is set out in the customer service robots guide. A shopping centre serves a local population and a hospital serves a catchment area. An international terminal serves whoever landed that morning. A robot handling a handful of languages therefore answers a smaller share of questions than the same robot would in a mall. Treat multilingual support as a threshold rather than a bonus.

Security Patrol and Perimeter Monitoring

Patrol robots in aviation cover ground that is expensive to walk and dull to watch. Knightscope sells autonomous patrol machines that record video, read number plates and flag people in areas that should be empty. That fits long-stay car parks, service corridors and perimeter roads. The saving is coverage rather than headcount, because one operator watches several feeds instead of driving the loop.

The perimeter is where airports differ from a warehouse fence. An airfield boundary runs for miles across grass and service roads, beside movement areas where anything straying onto a taxiway becomes an air safety incident. Perimeter deployments therefore stay outside the movement area, or run under a procedure agreed with airport operations. Ask any vendor promising autonomous runway inspection to name the airport that approved it.

Treat patrol robots as a supplement to the security programme an airport already runs. Screening, access control and the officers on duty are all mandated. A patrol machine adds recorded coverage of places nobody was watching, and the wider category is compared in the security robots guide.

Baggage, Cargo and Ramp Equipment

The airside work with real commercial traction is baggage and cargo movement. Aurrigo builds the Auto-DollyTug, an electric autonomous vehicle combining a baggage tractor and a dolly. It drives sideways to position itself against a stand and loads unit load devices with robotic arms. It has been trialled at several airports, including work with Changi Airport Group in Singapore and the Royal Schiphol Group in Amsterdam.

Baggage suits automation because the route is fixed and the shift is hard to staff. A tug drives the same loop between the sorting hall and the stand hundreds of times a day, outdoors, in whatever the weather is doing. Removing the driver takes a person off a repetitive outdoor task. Nobody loses a customer-facing role to it.

The approval path is why this category moves slowly. An autonomous vehicle on the ramp shares the surface with pushback tugs, fuel bowsers, catering trucks and aircraft under power. It needs the ground handler's operating procedures rewritten and the airport's sign-off before wheels turn. Autonomous pushback and de-icing sit a step further out again. Both put a machine in direct contact with an aircraft, and no airport has published a routine unsupervised deployment of either.

• Baggage and cargo tugs: trialled and moving toward commercial use
• Terminal cleaning: in routine service at many large airports
• Wayfinding and information: pilot stage, usually in the sterile area
• Perimeter and car park patrol: in service outside aircraft movement areas
• Autonomous pushback and de-icing: demonstrated, not routinely deployed unsupervised

Why Airport Procurement Is Slower Than It Looks

Four constraints make robots in aviation harder to buy than the same machines elsewhere. Security zoning splits the estate into areas with different rules. The terminal never closes, so there is no overnight maintenance window. Passengers arrive speaking dozens of languages. Anything moving airside needs a sign-off a facilities manager cannot grant alone.

Charging is the constraint that catches new buyers. A cleaning robot needs a dock with power somewhere that is neither a fire escape route nor a retail unit paying rent per square foot. Terminal floor space is allocated years ahead. What we see readers get wrong most often is treating the dock as an installation detail. In a live terminal it is a property negotiation, and it can outlast the pilot.

The upside is that the upside is scale, and it is why airports repay these machines when smaller buildings do not. Floor areas run to hundreds of thousands of square feet, staffing is round the clock, and the same questions get asked at the same points every day. Where a machine can legally operate, the volume pays for it.

Who Should Not Buy an Airport Robot

A regional airport with one terminal and a handful of daily departures should not start here. Every case above depends on volume, and a building that empties by mid-afternoon will not repay a cleaning fleet or an information humanoid. Spend the money on self-service check-in and better signage, which work with nobody on shift.

Ground handlers without their own airside operating certificate should also wait. If your staff work under another operator's security programme, you cannot get an autonomous vehicle approved on that ramp on your own timetable. The pilot then stalls at sign-off after the hardware is paid for. Confirm who holds the approval before signing anything.

Two developments would change this advice. A recognised approval route for unmanned ground equipment airside, of the kind that already exists for badged workers, would cut months out of every ramp project. Cleaning robots that hit their productivity figures in a crowded concourse rather than an empty one would move terminal cleaning from a shift-shortening tool to a headcount decision.