Hi, my name is Mathias Hauan Arbo and I am the second PhD student that will be working with SFI Manufacturing. The Bayesian approach to robotics is what I will employ in my PhD.
I come from the department of engineering cybernetics and work mainly with sensor fusion and robotics. My master thesis was on sensor fusion of delayed displacement measurements. The Bayesian formulation of how to handle that delay was my main topic. Describing the general underlying probability based algorithm for handling delayed displacement measurements in systems doing localization. This Bayesian approach to robotics is what I will employ in my PhD. I will be focusing on using a sensor fusion approach to robots interacting with flexible materials and objects.
Robots and manufacturing processes
Rigid-body physics allows us to create highly accurate models for the motion of objects that obey the rigid-body assumptions. And with modern techniques for grasping, computer vision, and sensor fusion, industrial robots are used in a variety of manufacturing processes.
Unfortunately, most objects do not always obey those rules. Metals bend, cloth folds, and wood breaks. This means that one either has to rely on human workers, or dedicated tools and hardware for each task. This limits a robot’s adaptability, and subsequently limits the industry.
Yet, mechanically, a robot should be capable of doing anything a human can do. To grasp an object so as to minimize the strain it will feel when moved is a matter of accurate modeling and planning. To move an object without it bending is a matter of sensor fusion with assumptions on flexibility. And to do this in an environment for manufacturing requires the combination of these things.
Robots interacting with flexible materials
The focus in this PhD will be on robots interacting with flexible materials. More specifically on how to use sensor fusion methods for tracking flexible objects, grasping strategies, and manipulation of flexible materials with industrial robots.
A probabilistic approach to tracking and localization of a robot manipulator when interacting with flexible materials may allow us to handle unmodeled vibrations and flexing as positional uncertainties. This may be useful for path planning, and ensuring safety in human-robot interactions. Right now, the goal is only a faint shimmer in the distance. And it is time to start figuring out the lay of the land, so that I can find the best path there.
I come from the department of engineering cybernetics and work mainly with sensor fusion and robotics. My master thesis was on sensor fusion of delayed displacement measurements. The Bayesian formulation of how to handle that delay was my main topic. Describing the general underlying probability based algorithm for handling delayed displacement measurements in systems doing localization. This Bayesian approach to robotics is what I will employ in my PhD. I will be focusing on using a sensor fusion approach to robots interacting with flexible materials and objects.
Robots and manufacturing processes
Rigid-body physics allows us to create highly accurate models for the motion of objects that obey the rigid-body assumptions. And with modern techniques for grasping, computer vision, and sensor fusion, industrial robots are used in a variety of manufacturing processes.
Unfortunately, most objects do not always obey those rules. Metals bend, cloth folds, and wood breaks. This means that one either has to rely on human workers, or dedicated tools and hardware for each task. This limits a robot’s adaptability, and subsequently limits the industry.
Yet, mechanically, a robot should be capable of doing anything a human can do. To grasp an object so as to minimize the strain it will feel when moved is a matter of accurate modeling and planning. To move an object without it bending is a matter of sensor fusion with assumptions on flexibility. And to do this in an environment for manufacturing requires the combination of these things.
Robots interacting with flexible materials
The focus in this PhD will be on robots interacting with flexible materials. More specifically on how to use sensor fusion methods for tracking flexible objects, grasping strategies, and manipulation of flexible materials with industrial robots.
A probabilistic approach to tracking and localization of a robot manipulator when interacting with flexible materials may allow us to handle unmodeled vibrations and flexing as positional uncertainties. This may be useful for path planning, and ensuring safety in human-robot interactions. Right now, the goal is only a faint shimmer in the distance. And it is time to start figuring out the lay of the land, so that I can find the best path there.
RSS Feed
