Vision-based control with applications to robotic systems
An increasing number of applications in robotics rely on vision-based control methods. As we enter into the era of the fourth industrial revolution, we have seen huge advances in image analysis using methods from machine learning, artificial intelligence, and more recently deep learning. These advances have led to a completely new type of control systems in robotics that use images to generate reference signals for mechanical systems. This has led to a revolution in the way robots interact with the real world and with each other.
This project will develop a completely new class in vision-based control of robots, also called visual servoing. Even though this is at the very forefront of research, the algorithms are relatively simple to implement and understand. The class can therefore be taught at master or even bachelor level and will provide the students with extremely powerful tools to develop mechanical systems that can interact with the real world.
This is a collaborative project between NMBU and PUC-Rio where the main objective is to develop a completely new class in visual servoing, including supervision of master students. In close cooperation with the industrial partners Umoe Bioenergy and Saga Robotics we will also arrange workshops for students in Norway and Brazil where the students come together and work on related problems, with focus of implementation and gaining practical experience. This practical approach will give the students the necessary experience to implement these algorithms on real systems.