Abstract

This talk will address uncertainty phenomena in ground and aerial vehicles' control, navigation and perception problems. First, in order to handle the uncertainty problem in robot control, learning control methods will be discussed. Second, this talk will elaborate on uncertainties due to fast and agile flight applications. In particular, neuromorphic cameras will be discussed to deal with visual degradation such as motion blur or high dynamic range effect. Furthermore, due to the cost associated with data collection and training, sim2real transfer learning will also be mentioned to transfer knowledge between aerial robots and thereby increase the efficiency of their control. Not but not least, some state-of-the-art drone applications, e.g. autonomous drone racing, will also be elaborated.

Biography:

Prof. Kayacan's research and contributions lie within the areas of  control systems with applications in guidance, navigation, and automation of unmanned systems, machine learning, uncertainty-based and active learning.