Materials Handling
We are developing technology to enable autonomous vehicles to manipulate and interact with the environment rather than simply navigating through it. Typical tasks that are ripe for automation include the movement and stacking of containers, and the clearing of debris after a natural disaster. In order to perform such tasks, an autonomous vehicle must be capable of mapping the ever-changing environment and localising itself at a macro level (sufficient for navigation) within that environment. It must also be capable of recognising and tracking objects and performing fine-grained localisation to enable the pick-up and accurate placement of payloads. The vehicle must be capable of both short and long term task planning, as well as reacting to externally-imposed changes to the environment.
Autonomous Hot Metal Carrier
Our technology for automation of ground vehicles has been extensively tested on a Hot Metal Carrier (HMC). This is a 20 tonne forklift-style vehicle used in smelters to pick up and deliver giant crucibles of molten metal.
In a typical smelter operation, multiple HMC operate continuously both inside and outside of sheds in the presence of other site vehicles and people. The HMC operators must consider how to conduct their current tasks in this highly dynamic environment so as to ensure safe and efficient operations. Additionally, in order to achieve highly efficient operations, vehicle schedulers continually assess the location of the HMC fleet to ensure that their operations keep pace with the output of the smelter.
- Our prototype Hot Metal Carrier has demonstrated many hours of continuous autonomous operation.
Experience has shown that automated fleets offer the potential for consistent product delivery along with increased efficiency in operation and maintenance costs. CSIRO has automated an HMC that is capable of conducting all the operations of a manned vehicle, including navigation, detecting and reacting to obstructions in the vehicle's path, and monitoring the vehicle's health. The prototype autonomous HMC has been in operation since 2005 at CSIRO's industrial worksite at the Queensland Centre for Advanced Technology, where the vehicle has conducted operations over many continuous hours in all weather conditions. Concurrently, field trials of the core technologies are also being undertaken at a local smelter.
This project is being undertaken as part of CSIRO's Light Metals Flagship program. For more on this project, visit "a smarter way of moving materials".
Task-based Robotics
- Our sensor-equipped tractor has been used to demonstrate autonomous golf ball collection and mowing, as well as visual road following, target following, and parallel parking.
The Autonomous Systems Laboratory maintains a fleet of ground vehicles that are used as research platforms to test and demonstrate autonomous operation as well as cooperative behaviour. In addition to the Hot Metal Carrier, the fleet includes a bobcat, an electric utility farm vehicle, and a tractor. The different vehicles are all fitted with the same basic automation systems so that new sensors and automation technologies can be easily installed and tested on multiple vehicles.
Our primary platform for demonstrating task-based robotics in mining and industrial applications is a Bobcat S185 skid-steer loader. The Bob-Tach™ mounting system enables us to switch between various implements according to the task at hand: we use a combination bucket and a blade dozer for mining applications and a pallet fork attachment for pick-and-place operations typical of an industrial environment. The vehicle is instrumented with an onboard computer, a sensor suite, and a communication link that support various levels of automation, from remote control to supervised autonomy.
- This Bobcat S185 skid-steer loader is our primary platform for demonstrating task-based robotics in mining and industrial applications.
In mining, our research focuses on automating tasks that will enable the deployment of an autonomous earth moving system. The main tasks under investigation include excavation, loading, haulage and dumping. A first technology demonstrator is currently under development and uses the Bobcat for excavating material at a point A, then hauling and dumping it at a point B. In parallel, we are responding to the needs of the mining industry by investigating the integration of human and robotic controllers for these specific tasks. The challenges in our research stem from the need to operate day and night, in an unstructured dynamic environment, in dusty conditions and inclement weather, and possibly co-ordinate with other manned or unmanned equipment. (Read more about the applications of our research in mining here.)
Another important thread of research is the automation of pick-and-place tasks in an industrial environment. Building on the success of the autonomous hot metal carrier project, we are looking at developing an autonomous forklift system, capable of handling palletized material. Operating in a semi-structured industrial environment raises many of the challenges mentioned above. In addition, it may be impractical to restrict personal movement around the vehicle, forcing the system to negotiate strategies for sharing the workspace with humans.
Autonomous Reversing of a Tractor-Trailer System
- Our autonomous tractor-trailer reversing system performed as well as a moderately skilled human operator.
Reversing a tractor-trailer system is an notoriously difficult task. Our work has shown how a tractor can successfully reverse with a trailer using a trajectory following control system that at a low-level controls the hitch-angle of the tractor-trailer rather than the steering angle of the tractor. This control method was implemented on our experimental tractor vehicle, and its performance compared against that of a number of human drivers. The control system performed as well as a moderately skilled human operator, and therefore offers potential for driver-assistance in applications such as caravaning or boating, as well as potential for full automation in industrial applications such as transportation, cargo handling, or agriculture.
Contacts
Publications
- Navid Nourani-Vatani, Jonathan Roberts, and Srinivasan, M. "IMU aided 3D visual odometry for car-like vehicles". In: Kim, J. and Mahony, R., eds. Proceedings of the 2008 Australasian Conference on Robotics & Automation (ACRA 2008); Canberra, ACT. ARAA; 2008. 8 pp. ISBN: 9780646506432.
- Stephen Nuske, Jonathan Roberts, and Gordon Wyeth. "Visual localisation in outdoor industrial building environments". IEEE International Conference on Robotics and Automation (ICRA 2008); Pasadena, Calif. IEEE; 2008: 544-550. ISBN: 9781424416462.
- Cedric Pradalier. "Vision-based handling tasks for an autonomous outdoor forklift". In: Laugier, C. and Siegwart, R., eds. Field and Service Robots: Results of the 6th International Conference (FSR '07); Chamonix, France. Springer; 2008: 61-70. (Springer tracts in advanced robotics. v. 42). ISBN: 9783540754039.
- Cedric Pradalier, Ashley Tews, and Jonathan Roberts. "Vision-based operations of a large industrial vehicle: autonomous hot metal carrier'. Journal of Field Robotics. 2008; 25(4-5):243-267.
- Cedric Pradalier and Kane Usher. "Experiments in autonomous reversing of a tractor-trailer system". In: Laugier, C. and Siegwart, R., eds. Field and Service Robots: Results of the 6th International Conference (FSR '07); Chamonix, France. Springer; 2008: 475-484. (Springer tracts in advanced robotics. v. 42).
- Cedric Pradalier, and Kane Usher. "Robust trajectory tracking for a reversing tractor trailer". J. Field Robotics. 2008; 25(6-7):378-399.
- Jonathan Roberts, Ashley Tews, and Stephen Nuske. "Redundant sensing for localisation in outdoor industrial environments" [presentation]. 6th IARP/IEEE-RAS/EURON Joint Workshop on Technical Challenges for Robot Dependability in Human Environments (IARP '08); Pasadena, Calif. 2008.
- Navid Nourani-Vatani, Mike Bosse, Jonathan Roberts, and Matthew Dunbabin. "Coverage algorithms for an underactuated tractor in an uncertain environment". IEEE International Conference on Robotics and Automation (ICRA '07); Rome, Italy. IEEE; 2007: 698-703.
- Jesus Nuevo, Cedric Pradalier and Luis Bergasa. "Model-based load localisation for an autonomous Hot Metal Carrier". IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2007); San Diego, Calif. IEEE; 2007: 247-252.
- Cedric Pradalier and Kane Usher. "A simple and efficient control scheme to reverse a tractor-trailer system on a trajectory". IEEE International Conference on Robotics and Automation (ICRA '07); Rome, Italy. IEEE; 2007: 2208-2214.
- Jonathan Roberts, Cedric Pradalier, and Ashley Tews. "Autonomous hot metal carrier". In: Grandfield. J. F. and Taylor, J. A., eds. Aluminium Cast House Technology : 10th Australasian Conference & Exhibition; Sydney, NSW. CSIRO Publ.; 2007: 247-256.
- Jonathan Roberts, Ashley Tews and Cedric Pradalier. "Experiments and experiences with dependability for a large autonomous industrial vehicle". 5th IARP/IEEE-RAS/EURON International Workshop on Technical Challenges for Dependable Robots in Human Environments; Rome, Italy. 2007.
- Jonathan Roberts, Ashley Tews, Cedric Pradalier and Kane Usher. "Autonomous hot metal carrier: navigation and manipulation with a 20 tonne industrial vehicle". IEEE International Conference on Robotics and Automation (ICRA '07); Rome, Italy. IEEE; 2007: 2770-2771.
- Ashley Tews, Cedric Pradalier and Jonathan Roberts. "Reliable autonomous industrial vehicle operations". IEEE International Conference on Robotics and Automation (ICRA '07); Rome, Italy. IEEE; 2007: 1176-1182.
- Kane Usher and Matthew Dunbabin. "Automated excavating and loading". Aust. Mining. 2007; 99(7):44.
