Pittsburgh International Airport Plans Up to Four Robot Dogs for Air Monitoring

Passengers at Pittsburgh International Airport could soon share the Airside Terminal with up to four autonomous robot dogs. The airport is preparing a six-month trial in which the four-legged machines will begin by collecting indoor air-quality readings, turning environmental sensors into a mobile monitoring system rather than relying only on measurements from fixed locations.

The trial with Pittsburgh-based Skild AI also calls for two charging stations. Each robot will be assigned a defined operating area and required to observe designated keep-out zones. Skild will install and map the machines, supervise and monitor their operation, and train airport employees to use them.

Air monitoring comes first

For the initial assignment, an environmentally equipped robot will travel through its mapped area and collect readings tied to specific locations. Its system will compare those measurements with predetermined thresholds and alert airport employees when a reading falls outside acceptable parameters.

That arrangement matters because mobility changes what an environmental sensor can cover. A fixed monitor samples conditions at one installation point, while a robot can reposition the same type of sensing capability across a larger operational area. The trial must establish whether that additional coverage produces timely, useful information without creating excessive supervision or navigation demands.

Cameras and other sensors will support navigation and anomaly detection, while Skild’s adaptive autonomy software will govern how the robot moves through the terminal. Information will be relayed to airport operations staff, leaving people responsible for interpreting alerts and deciding what action, if any, is needed.

The machines use wheels at the ends of articulated legs, according to a demonstration and trial briefing reported by TribLive. Wheels improve rolling efficiency on smooth terminal floors, while the leg joints let the platform change posture. That configuration is well matched to an indoor airport, but it does not remove the harder autonomy problem: moving predictably around passengers, baggage and employees whose paths can change without warning.

A terminal is not a controlled test floor

Airports place unusual demands on autonomous movement. Passenger density can rise rapidly around gates, queues alter walking routes, luggage introduces low obstacles, and people may stop or turn abruptly. A route that is open when a robot begins moving may be blocked seconds later.

Defined operating areas and keep-out zones therefore act as more than administrative boundaries. They reduce the navigation problem to spaces that can be mapped and monitored, while excluding locations where the robot should not travel. Continued Skild supervision and employee training add human control around the autonomy rather than treating the machines as unsupervised terminal workers.

The robots’ usefulness will depend on the complete chain from sensing to response: a sensor must produce a credible reading, the robot must associate it with the correct location, software must recognize that a threshold has been crossed, and the alert must reach the appropriate employee. Mobility adds coverage, but it also introduces changing position, charging requirements and interactions with the public.

Spills and full bins remain later tests

If the first phase progresses, Pittsburgh International Airport may evaluate whether the robots can identify spills, full or nearly full rubbish bins, and areas of passenger congestion. These are proposed second-phase applications, not capabilities already proven in terminal service.

They would also require different sensing and decision logic from air monitoring. A threshold-based environmental alert comes from a measured value, while recognizing a spill or judging whether a bin is nearly full depends on visual perception and classification. Congestion monitoring adds another layer because crowds are dynamic and an observation must be converted into information that staff can use.

More ambitious possibilities include carrying BioFlyte airborne-threat sensors and autonomously delivering food, drinks or employee supplies. BioFlyte began testing its BioTOF z200 bioaerosol technology at the airport in 2023, and the two organizations signed a five-year agreement in 2025. However, that does not establish that BioFlyte’s equipment has been integrated with a Skild robot or validated as a mobile system. Robot-mounted threat detection and terminal delivery remain possible later-stage work.

The project is being conducted through the airport’s xBridge program, which gives technology companies access to an operating airport rather than a closed laboratory. That setting is the point of the trial and its central constraint: up to four robots must show that they can collect useful information while staying inside mapped areas, respecting keep-out zones and operating under human oversight amid the constant movement of a passenger terminal.

By Jonathan Barrett — Editor for AMI’s future mobility and autonomous systems section, with two decades covering robotics, e-mobility, drone-vehicle convergence, and transport mechanical systems.