Define the service, not the technology
An autonomous ground vehicle supports a service that already has an owner, a deadline and a failure consequence. “Move baggage” is too broad a requirement. A more useful mission statement identifies a collection point, a destination, an approved route, the load configuration and the person who accepts completion. It also says what the system must not do. A tractor permitted on one service road is not thereby permitted to approach every aircraft stand.
Begin with an ordinary shift rather than a promotional demonstration. Record how loads become ready, how often destinations change and where drivers currently negotiate informal exceptions. Those small interventions are part of the existing service. Removing the driver without replacing that coordination can expose a hidden dependency rather than create a reliable automated workflow.
Separate baggage transport from baggage handling
Baggage transport is an attractive candidate because some journeys repeat between sorting facilities and staging areas. Yet the journey is only one part of the chain. Loading, reconciliation, cart coupling, restraint, inspection and aircraft-side acceptance may remain manual. A successful arrival does not establish that the right bags reached the right flight.
For a towing mission, examine the entire train. Trailer geometry, turning envelope, braking behaviour, coupling integrity and load distribution matter as much as the tractor’s sensing equipment. Route clearance must account for the last trailer, not only the vehicle body. A route that works empty may not work with the permitted load and cart combination.
British Airways and IAG reporting about Heathrow provides a useful starting point for researching baggage automation trials. Treat the reported trial scope as the evidence boundary. Do not translate a successful test into a claim that all baggage journeys at Heathrow are autonomous or that the same operating model transfers unchanged to another airport.
Compare cargo and service logistics
Cargo transfer can offer stable origin-and-destination pairs, especially between controlled facilities. However, unit load devices, freight dollies and warehouse interfaces introduce their own handling constraints. Confirm load dimensions, securement, permissible acceleration and the receiving team’s ability to accept an early or late arrival. Dangerous goods require separate controls; autonomous movement does not remove the underlying transport obligations.
Service logistics may include moving supplies, empty equipment or maintenance materials. These missions can be less tightly coupled to an aircraft departure than outbound baggage, making a limited initial deployment easier to isolate. The apparent simplicity is not universal: catering, fuel and aircraft servicing involve specialised hazards and interfaces. A generic autonomous platform is not automatically suitable for each task.
Recognise when people change the mission
A staff or passenger shuttle adds boarding, accessibility and emergency-response responsibilities. A route that is acceptable for unattended freight may not provide suitable evacuation points for passengers. The operator must consider mobility assistance, clear communication, a failed door, a medical event and a passenger who boards the wrong service.
Distinguish landside passenger shuttles from airside logistics. They may use similar technology but operate under different access controls, traffic rules and service expectations. The SAE J3016 terminology is useful for discussing driving automation roles, but a road-vehicle automation label does not itself approve an airside transport service.
Draw the operating boundary
The operational design domain describes the conditions in which the automated function is intended to operate. For airport planning, make that description tangible: mapped roads, designated crossings, speed limits, lighting, weather, surface condition, traffic interactions and permitted payloads. Include exclusions, such as construction zones, active runway crossings or particular aircraft-proximity areas.
A narrow boundary is not a weakness if it matches a useful task. It can make performance easier to measure and recovery easier to organise. Conversely, an ambitious route that depends on frequent manual assistance may transfer work from drivers to supervisors without reducing operational fragility. Compare complete service outcomes rather than vehicle utilisation alone.
Choose a measurable first mission
Select a mission with a clear customer, manageable interfaces and a credible fallback. Establish the current manual baseline: completed loads, waiting time, damage, missed windows and staff effort. Then decide what evidence would justify continuation, revision or stopping the trial. Do not define success only as distance driven without an incident.
IATA’s Airport Handling Manual and Ground Operations Manual help frame handling responsibilities, while airport-specific procedures define the local implementation. Changi Airport Group and Schiphol publish material on autonomous airside activity that can inform questions, but their reported conditions are not a substitute for your own task analysis. The right first mission is one whose boundaries can be explained clearly to everyone sharing the apron.