Robots Atlas>ROBOTS ATLAS
Robotics

Vehicle Marshalling

ActivePublished
Key innovation
Frames the in-facility moving, sorting, parking and staging of vehicles as a distinct logistics process that can be automated with autonomous transport (AVM) instead of human drivers.
Category
Robotics
Abstraction level
System
Operation level
SystemRobot control
Use cases
Vehicle drive-off from the conveyor belt at end-of-lineTest tracksTest rigs and measurement stationsParking lots and vehicle logistics centresStaging vehicles from end-of-line to dispatchAutomotive OEM plantsPorts and yards

How it works

In the manual form, drivers ferry vehicles between successive stations (drive-off from the conveyor, test tracks, measurement rigs, parking, dispatch). In the automated form (AVM) an infrastructure-based autonomy model is used: the intelligence sits in the facility rather than in the vehicle, with no onboard sensors. LiDARs mounted in the infrastructure localise each vehicle with centimetre-level accuracy and continuously detect static and dynamic objects. An offboard autonomy stack performs perception, behaviour prediction (learning-based AI models) and trajectory planning; trajectories are transmitted wirelessly to the vehicle about 10 times per second in packets of around 150 bytes. A fleet management system assigns missions and optimises throughput.

Problem solved

Manually ferrying every vehicle between stations inside a facility is labour-intensive, costly, hard to scale to 24/7 operation and prone to collisions and damage. Vehicle marshalling organises this flow, and automating it (AVM) cuts driver costs and stabilises throughput.