Ask a baggage manager what breaks automation projects and the answer is rarely the standard suitcase. It is the soft backpack that slumps, the odd-shaped case that will not sit square, the bag whose strap catches everything. The 80/10/10 model of airport baggage handling automation starts from that reality instead of fighting it: automate the predictable majority with dedicated transfer robots, give the irregular minority to a flexible humanoid, and keep human handlers for the true exceptions. This article explains the three layers and how they work as one system in Booster's airport baggage transfer solution.
The 80 Percent: Standard Bags, Machine Rhythm
Most checked baggage is boring, and boring is automatable. Standard suitcases arrive in a known size range, present flat faces to a gripper and tolerate a fixed rhythm. The DawnAnt baggage transfer robot was built for exactly this flow: a 6-DoF collaborative arm with a 2,000 mm reach and a suction-cup and gripper composite end effector identifies each bag, grips it securely and loads it at up to 180 pieces per hour, running 7x24 through departure peaks and holiday surges. One robot's shift never ends, never tires at hour six, and never varies its cycle - which is what a transfer link feeding a make-up area actually needs.
The 10 Percent: Irregular Bags, Flexible Hands
Backpacks, handbags, soft holdalls and oddly strapped cases defeat fixed grippers not because they are heavy but because they are unpredictable - they slump, swing and present no reliable face. The second layer answers with a wheeled dual-arm humanoid robot: two 7-DoF arms that can cradle, re-grip and steady a soft bag the way a person does, guided by dual LiDAR and depth-vision navigation in a live hall. It is slower than the transfer robot and deliberately so; its job is the bags that would otherwise stop the line, handled at whatever pace keeps them moving.
The Final 10 Percent: People, Where People Are Best
Some bags should stay manual: the leaking case, the damaged bag needing repacking, the item flagged for security handling, the live-animal crate. The model reserves people for judgement work rather than spending them on ten thousand routine lifts. That allocation also answers the workforce question honestly - automation absorbs the volume growth and the injury-prone repetitive lifting, while the human role concentrates on exceptions, supervision and recovery when something in the hall goes wrong.
One System, Not Three Purchases
The layers only pay off when they are scheduled as one flow. A multi-robot scheduling layer assigns work across transfer robots and the humanoid, while full-process digital traceability records every bag's journey through the system - which machine handled it, when, and where it went next. When a bag is queried at the desk, the answer comes from the record, not from a walk down the hall. Airports evaluating the model should therefore specify the software layer with the same rigour as the robots: scheduling, traceability and the operator's view of exceptions are what turn three machines into a system.
Piloting Without Rebuilding the Hall
The model's practical virtue is that it starts small. DawnAnt works within existing terminal processes - its rail-based layout adapts to the hall rather than demanding a new one - so a pilot can automate a single transfer link, measure pieces per hour, exception rates and traceability completeness against the manual baseline, and extend link by link. The 80/10/10 split is a design centre, not a promise: each airport's baggage mix moves the percentages, and the pilot exists to find the local numbers before the wider commitment.
Running a terminal baggage operation or planning its upgrade? Compare the relevant product range and deployment scenarios to take the next step.
