Robotics Engineering Colloquium Series: Jana Pavlasek | Think Graphical, Act Local: Distributed Inference for Robot Perception, Planning, and Education

Wednesday, September 27, 2023
12:00 pm to 1:00 pm
Floor/Room #

Robotics Engineering Colloquium Series

Jana Pavlasek

Think Graphical, Act Local: Distributed Inference for Robot Perception, Planning, and Education

Wednesday, September 27th, 2023
12:00 PM - 1:00 PM
Unity Hall Rm 520

Abstract: The field of robotics has seen impressive growth over the past decades, but the promise of ubiquitous, versatile robot assistants has yet to be fulfilled. To get robots out of the lab and into the real world, we need to scale our solutions to complex and uncertain environments. My work focuses on distributed solutions to robotics problems, taking a divide-and-conquer approach to solving these complex problems by considering multiple smaller subproblems which can be efficiently solved in parallel. In particular, my research focuses on problems which can be represented as graphs and solved via graphical probabilistic inference. These distributed inference algorithms enable scalable, robust operation under the uncertainty inherent in real-world environments which arises from variability, partial observability, sensor noise, and other agents. In the first part of this talk, I will discuss my work on using belief propagation to solve two challenging robotics problems: highly cluttered articulated object pose estimation and multi-robot coordination. I will demonstrate how nonparametric representations of uncertainty can yield diverse solutions using differentiable algorithms. In the last part of the talk, I will show how lessons on distributed inference can be applied to education to meet the growing demand for robotics curricula across universities.

Bio: Jana Pavlasek is a PhD candidate in the Robotics Department at the University of Michigan advised by Professor Chad Jenkins. Her research interests include robotic perception and planning under uncertainty in challenging, real-world environments. Particularly, her work focuses on scalable distributed inference algorithms, including perception in clutter and multi-robot coordination. Jana is the developer and instructor for a core course on introductory robotics for the Robotics undergraduate degree at the University of Michigan, which has also been taught across several other universities as part of a distributed teaching initiative. Prior to her PhD, Jana received a Bachelor’s degree in Honours Electrical Engineering at McGill University. She is the recipient of the NSERC Postgraduate Scholarship and has interned at NVIDIA Research.



Robotics Engineering