Autonomous terminal vehicles

Autonomy for
the working port

An electric container carrier and fleet software, engineered for SAE Level 4 operation within defined terminal areas.

SAE Level 4
Target, within defined terminal operating areas
Battery-electric
Regenerative braking, built for round-the-clock moves
All-wheel steering
Tight turns and precise container handoff
Fleet software
Dispatch, routing and monitoring across the terminal
01 The problem

The world's ports still run on drivers.

Container volumes keep climbing, but most terminals move every box with a person in a cab, alongside heavy machinery, around the clock. It is costly, hard to staff and among the most dangerous work in logistics.

TEU 858M1

Containers handled by the world's ports in 2023, up a third in a decade.

of terminals 7%2

Main container terminals that are fully or semi-automated; the rest still run on drivers.

fatality rate 5×3

US marine-terminal workers die on the job at about five times the all-worker rate.

CAGR 26%4

Forecast annual growth in automated guided vehicles deployed at seaports to 2030.

02 The vehicle

One purpose-built machine for the container move.

A battery-electric carrier designed around quay-to-yard transfers: redundant sensing on every side, automotive-grade compute and a safety channel that can stop the vehicle on its own.

Concept visualisation: the General Axis vehicle on a wet quay at night, a container just set on its deck by a ship-to-shore crane
Autonomy
SAE Level 4 target, within a mapped terminal operating domain
Drivetrain
Battery-electric with regenerative braking
Steering
All-wheel, for tight turns and precise handoff
Sensing
360° cameras, lidar, 77 GHz radar, RTK-GNSS and IMU
Safety
Independent safety monitor and hardware emergency stop
Fleet
Dispatch, routing and monitoring across the terminal

Development targets, subject to engineering validation and site conditions.

Full platform and sensor suite
Concept visualisation: dense sea fog at night over a container quay, cranes lit by blue floodlights, and the headlights of a General Axis vehicle far down the lane
Concept visualisation: the General Axis vehicle carrying a rust-red container along a wet quay at night in fog, headlights on
Concept visualisation: the vehicle seen from behind at night in thick fog, with a forklift in silhouette further along the quay
Concept visualisation: the vehicle stopped on a wet quay at night in fog, a wide clear gap between it and a forklift driving away, its driver in a hi-vis vest and hard hat
Concept visualisation: General Axis vehicles carrying containers along separate lanes of a terminal just before dawn, cranes and a container ship behind them
03 Technology

Trained in the cloud, proven in simulation, driven on the vehicle.

Our development pipeline follows NVIDIA's three-computer approach to physical AI: models are trained on data-centre GPUs, tested against a digital twin of the terminal, then deployed to the vehicle's own computer.5

01 · Train

Learn from the terminal

Perception and planning models are trained on GPU clusters from real terminal driving data, then optimised for the vehicle with TensorRT.

GPU trainingPyTorchTensorRT
02 · Simulate

Prove it in a digital twin

A physically based twin of the terminal in NVIDIA Omniverse replays real drives and generates rare, dangerous scenarios on demand with Cosmos world models.

OmniverseOpenUSDCosmosIsaac Sim
03 · Drive

Run it on the vehicle

The same software runs on the bench and in the vehicle, on automotive-grade compute, with an independent safety channel watching every command.

Automotive-grade computeSafety monitorDrive-by-wire
How we approach Level 4
04 Safety

Engineered to fail safe.

At Level 4 there is no driver to hand back to, so the vehicle itself must reach a safe state whenever something goes wrong.6 An isolated monitor checks every command and can stop the vehicle without the main computer, backed by a hardware emergency stop.

Redundancy
Independent sensing, compute, power and braking paths
Operating domain
Geofenced to mapped terminal zones, speed bounded per zone
Rollout
Supervised by trained operators until safety is demonstrated
Standards
Engineered toward ISO 26262, ISO 21448 and ISO 3691-4
Safety and compliance
Concept visualisation: the General Axis vehicle at night with a container seated on its deck as the crane spreader releases it
05 Economics

The cost case for an autonomous electric fleet.

Electric drive cuts the running cost of a yard tractor by around two thirds, and autonomy removes the largest line item of all: a driver in every cab, on every shift.

Operating cost per hour $36 → $117

Diesel against electric yard tractor, excluding driver and electricity, in a public fleet comparison.

Per driver, per year $243k8

Average earnings of a full-time registered longshore worker on the US West Coast in 2024.

Electric tractor premium ≈ $150k9

Approximate extra purchase cost of an electric terminal tractor over diesel, charging included.

Market outlook and cost benchmarks
06 Deployment

From site survey to autonomous operation.

A staged path that keeps people in the loop until the system has earned trust on your terminal.

01

Site assessment

Survey quay lanes, staging zones and handoff points, and build a high-fidelity digital map of the terminal.

02

Digital twin

Model the site in simulation to validate routes, traffic rules and edge cases before a vehicle arrives.

03

Supervised autonomy

Staged rollout with operators in the loop, validating safety envelopes against real operating conditions.

04

Autonomous operation

Continuous operation within the approved operating domain, with 24/7 fleet monitoring and failover.

Contact

Bring autonomy to your terminal.

For pilot deployments, collaboration and integration planning across port and wharf environments. We typically respond within two business days.

Prefer email? Write to contact@generalaxisrobotics.com.