Physics of Living Systems (PoLS) Seminar - Prof. Floris van Breugel

Speaker: Prof. Floris van Breugel, Univ.  Nevada Reno.

Host: Prof. Simon Sponberg

Title: : Information gathering as a guiding principle for animal (and robot) movement

Abstract: 

I study how organisms integrate sensory information from multiple modalities across time and space to make decisions in complex naturalistic environments. My end goal is twofold: to understand how brains process sensory information, and to generate new, bioinspired, algorithms for engineered systems that enable the kind of resilience characteristic of biology. In my talk I will describe recent work in my lab that leverages optogenetics in freely flying fruit flies to remotely activate their sense of smell. Using this approach, we discovered a novel behavior they localize odor sources in still air. We also show that flying flies are capable of estimating the presence and direction of ambient wind. To understand how they might achieve this my group developed new control-theoretic tools for empirically assessing the nonlinear observability of individual states—that is, what sensor combinations and movement motifs are required such that wind direction can be estimated. Finally, I will describe our preliminary efforts to design nonlinear observers for wind direction, and describe a framework for how this approach could lead to estimation strategies that are resilient to unanticipated measurement anomalies.

Bio: Floris van Breugel is an assistant professor at the University of Nevada, Reno in the Mechanical Engineering Department, with affiliations with the Integrative Neuroscience and Ecology and Evolution programs. Floris earned his BS in Biological Engineering at Cornell, where he worked with Hod Lipson on bio-inspired flapping machines. He earned his PhD in Control and Dynamical Systems at Caltech under Michael H Dickinson and Richard Murray, with support from Hertz and NSF fellowships. After continuing as a postdoc with Michael, he did a brief postdoc at the University of Washington with Jeff Riffell, Nathan Kutz, and Bing Brunton with support from a Moore/Sloan Data Science Fellowship. He started his lab at UNR in 2019. His research brings together control theory, neuroscience, behavior, and bio-inspired robotics.