Skip to content. Skip to navigation
CIM Menus
 

CIM Robotics Seminar

Control of Complex Systems: Inverse Problems in Collaborative Robotics


Matthew Harker
Institute for Automation, Robotics and Artificial Vision University of Leoben
Austria

October 10, 2019 at  2:00 PM
McConnell Engineering Room 437

Abstract:

Groups of robots and drones can be used in a collaborative manner to accomplish tasks that may be unruly or impossible with a single robot, and can thereby enhance the efficiency and versatility of robotic mechanical systems. Each actor requires individual control, but the system as a whole requires intra-system control for coordination and collision avoidance. This problem is combinatorial in nature, and is best addressed by treating control as an inverse problem, that is, by critical reasoning in the direction from effects to causes. As an example, if multiple drones are used to carry a single object along a desired path, the inverse problem approach can be used to determine the controls required to track the path of each drone in an optimal manner while ensuring that the actual paths are physically realizable. Further, the same approach can be applied to maintain a specified distance between the vehicles to within a given tolerance where, as a trade off, a sacrifice in the path tracking accuracy arises. Applications of this type of collaborative control are physio-therapeutic assistive devices, agri-mechatronics, and adaptive manufacturing. The inverse problem approach to the control of complex systems is a fruitful one, as it can also be used to treat the problems of functional safety, system identification, and optimal control, to name but a few.

Bio:

Matthew Harker obtained his bachelor degree in mechanical engineering with a specialization in mechatronics from McGill University. Subsequently, he obtained his Ph.D. in Austria at the University of Leoben, working at the Institute for Automation. He has obtained his habilitation (a prerequisite for European professorship) in Automation Engineering, and teaches courses in the area of mechatronics at the University of Leoben. His research interests lie in inverse problems in mechatronics, such as measurement and control for heavy machinery and robotic mechanical systems.