BISON, PHOENIX, ATLAS 2: Training the Transition Before the Vehicle Arrives
Our pushback and towing simulator was built with Goldhofer and carries all three of their tractors. Any move between them can be trained in advance, on your own stands, with your own operators, before a single new machine touches the apron.
Ground handling fleets are moving between tractor types more than they used to, and not in one fixed direction. BISON to PHOENIX. PHOENIX to ATLAS 2. BISON straight to ATLAS 2. Some operations run two types side by side for years and never fully converge.
Every one of those moves creates the same practical problem. The operator is qualified on the machine they have been driving, the new machine changes something about how the job is done, and the window in which both machines exist is short, busy and expensive.
The simulator exists to take that problem off the apron. We developed it together with Goldhofer, who were involved from the start of the project through to validating each of the vehicle models. The BISON, the PHOENIX AST-2E and the ATLAS 2 are all in the same system, and they behave the way the real machines behave because the manufacturer said so. An operator can run a pushback on one, then repeat it on the next, then on the third, back to back in a single session, and feel the difference rather than be briefed on it.
They do it on their own airport. We replicate the customer's aerodrome in full, and more than one if the operation covers several stations, so the stands in the simulator are the stands the operator will actually work: the same layout, the same approach lines, the same tight corners and the same poor sightlines.
Four ways operations use it
Pre-training, before delivery. The transition can be trained while the new tractor is still being built. Operators arrive at the handover already familiar with the machine, the pick-up sequence and the way it behaves on their own stands. The delivery date stops being the start of the learning curve.
Blended with live apron time. The simulator is not a replacement for supervised line training and is not sold as one. It works best as the layer underneath: repetitions, mistakes and non-normal situations happen in the simulator, and live apron time is spent consolidating rather than introducing. Operators reach their first real pushback on the new machine having already done it many times.
Vehicle familiarisation. Controls, instrumentation, sightlines, the pick-up sequence, the feel of the combination under load. The things an operator normally absorbs over weeks of live work, compressed and repeatable.
Stand and procedure familiarisation. The new machine changes how a stand is worked: approach, positioning, connection, the manoeuvre itself, disconnection, and what the operator is watching at each step. Because the airport is modelled in full, those procedures are rehearsed stand by stand on the real layout rather than on a generic apron, which matters most on the positions where the margins are tightest.
And through the mixed-fleet period, when operators have to stay current on two types at once, both types stay available in the simulator regardless of what is happening to the real fleet.
What the training has to cover
The content of a transition syllabus depends on which move is being made. Two examples, using Goldhofer's own descriptions of the vehicles.
BISON to PHOENIX AST-2E
The towbar disappears as an intermediate element. The nose landing gear is picked up and cradled, so the combination behaves as one unit rather than as an articulated pair: the operator's steering input goes straight to the nose gear instead of through a hinge, and the geometry they have spent years reading no longer applies. The pre-pushback sequence changes with it, since there is no towbar to bring out, match to the type and connect.
The aircraft range widens too. Goldhofer describes the PHOENIX E as covering 70 percent of the passenger and cargo aircraft on the market today, up to 352 t maximum take-off weight, so type-specific limits stop being background knowledge and become a daily decision.
Towing changes in scale rather than in kind. Conventional tractors tow as well as push, but without a type-specific towbar to fetch and fit, moves to a remote stand, to the hangar or to a de-icing pad get easier to schedule. Where an operation responds by doing more of them, the syllabus has to reach past the stand into the manoeuvring area.
What the operator practises in the simulator:
Approaching and aligning to the aircraft without a towbar, and running the pick-up and cradle sequence until it is routine rather than deliberate
Steering a rigid combination: watching where the nose gear goes rather than where the tractor goes, and rebuilding the turn judgement that the hinge used to provide
Finding new visual reference points, since the towbar itself used to be one of them
Pushback geometry on constrained stands, including the awkward ones on their own apron
Braking and stopping behaviour of the loaded combination, and how it changes on wet or contaminated surfaces
Retiming communication with the flight deck and the headset operator around a pre-pushback sequence that is now shorter
Disconnection and clearing the aircraft
Selecting the aircraft type and working to type-specific limits across a much wider range of aircraft than before
Towing over distance: taxiway routing, speed discipline and how the combination behaves away from the stand
Non-normal situations that cannot be staged on a live aircraft: an obstruction in the push path, a stop and reposition mid-manoeuvre, loss of communication
PHOENIX AST-2E to ATLAS 2
This one is easy to underestimate, because the ATLAS 2 is built on the proven PHOENIX platform and carries over key elements including the pick-up system. Shared lineage is a genuine advantage, and it also means the transition can present as a like-for-like refresh.
Look at what Goldhofer says the vehicle does. The pick-up process for the nose landing gear is guided: the machine actively supports correct pick-up and verifies the aircraft type. It regulates permissible braking and towing forces according to the aircraft being handled, reducing load on the nose gear. There is a single-lever emergency switch-off, an automatic lock against accidental movement, and 360° all-round vision for manoeuvring. Goldhofer's product management team describes it as a simple guided operating concept giving active protection against operator errors.
Those are safety gains, and several of them remove failure modes that used to sit entirely with the person in the cab. What they change for training is the operator's role: part of the job that used to be done unaided is now shared with the machine, so the operator needs to know what the system manages, what remains theirs, and what normal looks like well enough to notice when it is not.
What the operator practises in the simulator:
The routine gate pushback with the guidance working as intended, repeated until the operator knows exactly what normal looks and feels like
The guided pick-up itself: what the machine is showing, what it is verifying, and how to respond when the verification does not agree with what the operator expected
Working alongside force regulation: which decisions the system is now making, which remain the operator's, and where the line sits
Recognising an abnormal system state early, under the time pressure of a real departure bank
The emergency switch-off and the automatic lock, so both are muscle memory rather than something read about
A scan pattern built around 360° vision rather than around the mirrors and blind spots the operator used to compensate for
Multi-gate working: moving between stands and fitting the machine into a different allocation pattern
Planning charge state across a shift, alongside the departures the operator is committed to
What you get out of it
Two more things are worth practising whichever direction the transition runs. The first is a paired run: the same aircraft, the same stand, the same weather, performed first on the machine the operator knows and then immediately on the one they are moving to. Everything except the vehicle is held constant, so the only thing that changes is how the machine behaves, and the operator feels that difference instead of being told about it. This matters because the real risk in a transition is the reflexes carried over from the previous tractor, and reflexes are not changed by a briefing. The second is the mixed-fleet reality of switching between two types within a single shift, which is the situation the changeover period actually produces.
Because every operator runs the same scenarios on the same stands, the results are comparable rather than anecdotal: like for like, per operator, per station and per machine. Session recording and scoring, including export into your own training records, can be built in on request.
That makes a transition something you can plan and evidence rather than something you get through.
The three vehicles are already built. If your operation is moving between any two of them, or running two side by side, the training for it can start now.
Talk to us
We are at GSE Expo Europe in Lisbon, 15 to 17 September 2026, stand D57, with the simulator on the stand. Come and run a pushback yourself, on any of the three tractors. If you would rather talk first, write to us and we will walk you through what a transition programme would look like for your fleet and your stands.
Sources: Goldhofer press release on the »ATLAS« 2 world premiere at the International GSE Expo, 17 September 2025; »ATLAS« 2 product brochure (08/2025); »BISON« Family brochure (08/2025); Goldhofer review of GSE Expo Europe 2024.
