BEGIN:VCALENDAR VERSION:2.0 PRODID:-//Memento EPFL// BEGIN:VEVENT SUMMARY:From Insects to Robots and Back DTSTART:20190311T140000 DTEND:20190311T160000 DTSTAMP:20260508T083610Z UID:3bf154589d762ee4563548fc2aa9e05678bd317f6d774daad704462a CATEGORIES:Conferences - Seminars DESCRIPTION:Holger Krapp\, Stéphane Viollet\, Emily Baird\, Pavan Ramdya\ nA thematic workshop at the intersection of biorobotics\, neuroscience\, c omputer vision\, and behavioural science with 4 speakers in 2 hours\n\nFl ies and other autonomous systems. The benefits of interdisciplinary appro aches to study sensorimotor control\nHolger Krapp\nImperial College Lond on\n\nAbstract: Inspiration from biology to improve the performance of au tonomous robotic systems is mostly associated with the translation of des ign principles of living systems into novel technical applications. But t he joint venture between biological research and engineering also enables us to apply engineering methodologies to study functional mechanisms in biology. In my presentation I will review earlier work on biological matc hed filters for optic flow and its implementation of aerial robotic syst ems. I will also present two robotic platforms for studies on sensory sys tems that flies use to control their gaze and flight: (i) A fly(brain)-ro botic interface (FRI) where the electrical activity of an identified mot ion-sensitive cell steers a 2-wheeled robot clear of collisions with the walls of an experimental arena\; (ii) A robotized gonimetric recording se tup that has been developed based on the parts of the FRI system that he lps to further study general functional principles of biological optic fl ow processing. Finally\, I will discuss sensorimotor control architecture s that suggest how known locomotor activity helps to derive meaningful s teering commands from ambiguous sensory signals.\nBio: Holger G Krapp is Professor of Systems Neuroscience in the Department of Bioengineeing at I mperial College London. He was awarded his Diploma in Biology and his PhD for work on the neuronal basis of optic flow processing in flies at the Max-Planck Institute for Biological Cybernetics in Tübingen\, Germany. A fterwards he spent one year as a postdoctoral scholar in the Division of Biology at Caltech\, USA\, studying neuronal circuits underlying collisi on avoidance in locust. He then returned to Germany working on insect vis on at Bielefeld University for three years before accepting a lectureship at the University of Cambridge\, UK\, in 2000. In 2005 he accepted a pos ition at Imperial College London where he has expanded his research towar ds the understanding of biological design principles supporting sensorimo tor control in insects with potential applications in autonomous robotic systems. He published >70 peer-reviewed papers and was invited to give > 100 conference and seminar talks\, including public/keynote lectures. He is PLoS ONE academic editorial board member\, assesses manuscripts for > 35 different scientific journals\, grant proposals for ten UK/internati onal funding agencies and was expert reviewer for the European Commission on an interdisciplinary robotics project.\n\nFrom insect perception to r obots and vice versa\nStéphane Viollet\nAix-Marseille University\n\nAbs tract: The Biorobotic approach is a meeting point where robotics and neu roscience are used to try to explain the behaviour of animals\, especially  winged insects (fly\, bee\, wasp...) and to model the processing of the sensory modalities at work in these outstanding animals. The neurophysiol ogy is also used to better understand the sensorimotor reflexes at work i n insects. The robots are a kind of embodiment of this insect-based knowl edge to validate our models. Recent studies carried out at our laborator y focused on the graviception in fly\, i.e.\, the ability of the animal t o assess its orientation with respect to gravity\, and on ant-inspired na vigation strategies by means of a celestial compass. Would it be possibl e that a fly\, able to achieve exquisite manoeuvres\, could not have any clue of its body tilt with respect to gravity during flight? Could future robotic applications take a great benefit of the skylight polarization? Several bio-inspired visuals sensors\, as well as bio-inspired robots\, will be presented in this talk.\nBio: Stéphane Viollet is a CNRS resea rch director at the Institute of Movement Sciences\, Aix-Marseille Univer sité and head of the Biorobotics team since 2008. He received the master ’s degree in control engineering from the University of Bordeaux 1\, F rance\, and the Ph.D. degree from the National Polytechnic Institute\, Gr enoble\, France\, in September 2001 under the supervision of Dr. Nicolas  Franceschini. He obtained a CNRS permanent position in 2003. His interest s include sensory-motor reflexes in flies\, retinal micro- movements and  bio-inspired control strategies for aerial robots. He is the leader of the development of autonomous robots and innovative visual sensors for r obotics (artificial compound eye\, hyperacute sensors\, and artificial ret inas) and involved in several national and European projects on these to pics. He is the author of more than 60 publications\, 8 patents\, and rec ipient of several best paper awards and nominations.\n\nWhat does a bee s ee? Towards understanding the link between visual morphology and visual flight control\nEmily Baird\nStockholm University\n\nAbstract: Insects suc h as bees rely heavily on vision to control their flight. The visual info rmation that they can use for this depends upon the morphology of the eye s\, which varies significantly across insects\, and even between closely- related individuals of the same species. To better understand the relatio nship between visual morphology and flight control behaviour in different insects\, we have developed a method that uses a combination of X-ray mi cro-computed tomography and ray-tracing. Here\, I will present some result s from this approach and discuss how it may be useful to understand visua l flight control strategies.\nBio: Emily Baird is an Associate Professor in the Department of Zoology at Stockholm University. She did her PhD at the Australian National University with Mandyam Srinivasan and Shaowu Zha ng before doing a brief postdoc with Martin Egelhaaf in the Department o f Neurobiology at the University of Bielefeld. She then moved to the Lund Vision group where she became an Assistant Professor before recently tak ing up a Senior Lecturer position at Stockholm University. Emily's resear ch combines morphological analyses with behaviour to better understand ho w animals use visual information to guide them through their world. \n\n Reverse-engineering Drosophila limb control circuits\nPavan Ramdya\nEPFL\, Brain-Mind Institute and Bioengineering Institute\n\nAbstract: A shared g oal of neuroscience and robotics is to understand how systems can be built to move effectively through the world. However\, state-of-the-art algorit hms for selecting and executing limbed behaviors in robots are still quite primitive compared with those used by animals. To inform robotic control approaches\, we are investigating how the fly\, Drosophila melanogaster\ , walks\, grooms\, and moves its limbs. I will discuss how we are combinin g 2-photon imaging of the ventral nerve cord in behaving Drosophila with physics-based simulations and neural network modeling to uncover how flie s achieve flexible limb control.\n Bio: Pavan Ramdya is interested in how animal behavior diversifies\, evolves\, and can be artificially engineere d. He received his PhD in Neurobiology from Harvard University in 2009 and then performed Postdoctoral work at EPFL\, UNIL\, and Caltech in Robotic s\, Neurogenetics\, and Integrative Neurobiology\, respectively. His labor atory at EPFL in Lausanne\, Switzerland\, is leveraging computational\, en gineering\, and optical microscopy approaches to investigate how neural po pulation dynamics\, gene expression\, and experience sculpt behavior in  Drosophila melanogaster.\n  LOCATION:MED 2 1522 https://plan.epfl.ch/?room=MED21522 STATUS:CONFIRMED END:VEVENT END:VCALENDAR