Online Machine Learning for Robot Control

Abstract
The foundations of robotics and robot control were established at a time when there was very limited computational power available. Hand-crafted analytic model-based approaches have remained the prevalent robot control method for over half a century. But nowadays we have at our disposal huge computational resources and powerful machine learning methods, so why not use them to reinvent robot control? The goal of this PhD project is to develop radically new approaches for controlling robots based on Online Machine Learning that allow self-learning of robot’s own model on-the-fly while adapting to changes in both the robot and the environment.

How to Apply 
Please contact Dr Petar Kormushev if you are interested in applying for this project. 

For more information on submitting a PhD application, please consult our application pages here. When you submit your application, please cite this project title in the 'how you are planning to fund your PhD' section.