There are very few robot tasks that require as complicated an interaction with the environment as the autonomous excavation problem for fragmented rock. Not only does the mobile robot have to regulate its motion so as to effectively penetrate the rock pile, but it must also actuate its boom (manipulator) and bucket (end effector) so as to fill the bucket at every pass. Fragmented rock is not homogeneous in size and composition, can be wet and sticky—or dry, depending on local factors—and it inherently moves and exposes new hidden material that is not visible prior to executing the excavation process. Despite a large body of research and engineering work in robotic excavation, spanning decades, commercially viable systems are still in their infancy.
Our research in robotic excavation has focused on automated loading of fragmented rock for underground load-haul-dump (LHD) machines, although we have also conducted experiments using small surface wheel loaders. We have devised an admittance-based excavation control strategy that, in extensive field trials, has proven to be an effective approach to autonomous excavation. More recently, we have worked with Epiroc AB (Sweden) in the development of a learning-based admittance control scheme that adapts to changing pile conditions and can automatically track a desired bucket payload.
A. Lopez-Pacheco, Fully loaded: Atlas Copco feels its way into autonomous loading for LHDs, CIM Magazine, December 2016-January 2017.
A. Craig, Researchers rock out with robots, Queen’s Gazette, May 12, 2015.
H. Fernando and J. A. Marshall. What lies beneath: Material classification for autonomous excavators using proprioceptive force sensing and machine learning. In Automation in Construction, vol. 119, November, 2020. DOI: 10.1016/j.autcon.2020.103374
U. Artan and J. A. Marshall. Towards automatic classification of fragmented rock piles via proprioceptive sensing and wavelet analysis. To appear in Proceedings of the 2020 IEEE Conference on Multisensor Fusion and Integration, Karlsruhe, Germany, September 2020.
H. Fernando, J. A. Marshall, and J. Larsson. Iterative learning-based admittance control for autonomous excavation. In Journal of Intelligent & Robotic Systems, vol. 96, no. 3-4, December 2019. DOI: 10.1007/s10846-019-00994-3
H. Fernando, J. A. Marshall, H. Almqvist, and J. Larsson. Towards controlling bucket fill factor in robotic excavation by learning admittance control setpoints. In Proceedings of the 11th Conference on Field and Service Robotics (FSR 2017), Zürich, Switzerland, September 2017. DOI: 10.1007/978-3-319-67361-5
A. A. Dobson, J. A. Marshall, and J. Larsson. Admittance control for robotic loading: Design and experiments with a 1-tonne loader and a 14-tonne load-haul-dump machine. Invited paper in the special issue on Field and Service Robotics of the Journal of Field Robotics, vol. 34, no. 1, pp. 123-150, January 2017. DOI: 10.1002/rob.21654
A. A. Dobson, J. A. Marshall, and J. Larsson. Admittance control for robotic loading: Underground field trials with an LHD. In Proceedings of the 10th Conference on Field and Service Robotics (FSR 2015), Toronto, ON, June 2015. Conference best paper award! DOI: 10.1007/978-3-319-27702-8_32
A. A. Dobson. Autonomous Loading of Fragmented Rock: Admittance Control for Robotic Digging. Ph.D. Thesis, The Robert M. Buchan Department of Mining, Queen’s University, January 2015
J. A. Marshall, P. F. Murphy, and L. K. Daneshmend. Toward autonomous excavation: Full-scale experiments. IEEE Transactions on Automation Science and Engineering, vol. 5, no. 3, pp. 562-566, July 2008.
J. A. Marshall. Towards Autonomous Excavation of Fragmented Rock: Experiments, Modelling, Identification and Control. M.Sc.(Eng.) Thesis, Department of Mechanical Engineering, Queen’s University, Kingston, ON, August 2001.
Offroad Robotic Loader
Offroad’s robotic R520S loader is autonomous and equipped as a result of generous support—both financial and technical—from CFI, NSERC, and Clearpath Robotics.