Vehicle automation technology developed by CSIRO ICT Centre scientists is being trialled at Rio Tinto Aluminium’s Bell Bay Smelter in Tasmania.
CSIRO has been developing automation technology for hot metal carriers (HMCs) since 2005.
HMCs are large forklifts used in aluminium smelters to move giant crucibles containing molten aluminium from pot lines to casting machines. The crucibles weigh about two tonnes and hold up to eight tonnes of molten metal. Pot lines in smelters can be up to two kilometres long.
Autonomous vehicles have been operating in some factories for over a decade. However, heavy industrial settings, such as aluminium smelters, offer a greater challenge. Vehicles such as hot metal carriers operate in the smelter sheds near high temperature sources, as well as outdoors, in shifting weather conditions.
Our goal is to establish reliable technology for autonomous vehicles operating in these different environments.
The metals production industry is interested in automation for safety, consistency of operation and minimisation of maintenance and repairs, yet automated HMCs are not yet used anywhere in the world.
Productivity and safety
We are developing technologies that allow driverless vehicles to operate around industrial worksites.
Our goal is to have autonomous vehicles conducting operations that are repetitive, hazardous or expensive for people to undertake.
Autonomous vehicle technology can:
-
improve productivity
-
improve safety of site operations
-
reduce maintenance costs
-
provide more consistent and predictable service than human-conducted operations
-
provide better situational awareness of site and vehicle operations.
Successful demonstrations
Data from these sensors, along with a map of the site, are used to locate the vehicle to within 20 centimetres.
In the first phase of the project we automated a HMC by adding programmable controls and sensors for onboard systems monitoring and vehicle control. We also added scanning laser rangefinders and pan-tilt-zoom cameras for navigation, obstacle detection and crucible docking.
Since then we have developed higher level systems for autonomous operation. This vehicle has conducted hundreds of hours of successful demonstrations and tests at our site in Brisbane, Queensland, Australia.
Localisation solutions
One of the challenges facing the team has been the development of an accurate and robust localisation system that tells the vehicle where it is at any instant.
Our localisation system is based on laser scanners, a steering encoder and vehicle odometry. Data from these sensors, along with a map of the site, are used to locate the vehicle to within 20 centimetres at all times.
Component testing on manned HMCs
Recently, we began testing several non-control technology components on a manned vehicle operating at Rio Tinto Aluminium’s site in Bell Bay, Tasmania, Australia.
The technology we are testing provides the foundation for a vehicle to navigate autonomously around an industrial site, in the same way that the fully autonomous version does at the CSIRO site.
What's next?
We have demonstrated reliable and accurate navigation, localisation, obstacle detection, hardware, low- and middle-level software.
We are now investigating higher level components to allow dependable operations under various environmental and operational conditions. This research will involve pre-failure detection of hardware and software systems and intelligent decision making in the event of failure or degraded operations.
The outcome of this phase of the project will be an autonomous industrial vehicle capable of highly reliable, consistent, predictable and safe operations, and with the ability to monitor and manage its own systems.
There has been much interest from the commercial sector in our autonomous HMC technology, especially the various technology components of the automation systems.
Collaborators
CSIRO is working with Rio Tinto Aluminium on the HMC.
Read more about CSIRO's research on Robotics.
Site Map | Legal Notice and Disclaimer | Privacy | Copyright | Last updated: 16 May 2008
