EESS talk on "Effective dynamic environment monitoring with a swarm of buoys"
Event details
| Date | 12.03.2024 |
| Hour | 12:15 › 13:15 |
| Speaker | Prof. Roland Bouffanais, Department of Computer Science, University of Geneva |
| Location | Online |
| Category | Conferences - Seminars |
| Event Language | English |
Abstract:
Swarm Robotics offers a promising approach to the pervasive monitoring of marine environments. Traditional monitoring techniques rely on either a single autonomous robot—autonomous surface vehicle—or a fixed network of sensors. Neither existing technology is suitable or efficacious for the robust monitoring and tracking of dynamic environmental features at the surface of aqueous environments. There is a pressing need for small, low-cost and rapidly deployable autonomous buoys. One powerful source of inspiration comes from the process of self-organization and swarming, observed throughout the natural world. We present the design, construction, and testing of a relatively large swarm of buoys. This multi-robot system has been tested with up to 50 units dynamically deployed over large surface areas of an uncontrolled open-water environment without any supporting infrastructure. Multi-agent reinforcement learning (MARL) approaches have been considered to optimize the effectiveness of the swarming approach.
Biography:
Swarm Robotics offers a promising approach to the pervasive monitoring of marine environments. Traditional monitoring techniques rely on either a single autonomous robot—autonomous surface vehicle—or a fixed network of sensors. Neither existing technology is suitable or efficacious for the robust monitoring and tracking of dynamic environmental features at the surface of aqueous environments. There is a pressing need for small, low-cost and rapidly deployable autonomous buoys. One powerful source of inspiration comes from the process of self-organization and swarming, observed throughout the natural world. We present the design, construction, and testing of a relatively large swarm of buoys. This multi-robot system has been tested with up to 50 units dynamically deployed over large surface areas of an uncontrolled open-water environment without any supporting infrastructure. Multi-agent reinforcement learning (MARL) approaches have been considered to optimize the effectiveness of the swarming approach.
Biography:
Roland Bouffanais is an Associate Professor in the Department of Computer Science at the University of Geneva. His research focuses on interdisciplinary applications at the intersections of complexity, network science, control theory, machine learning, and multi-agent systems. He has published over 100 peer-reviewed papers in leading scientific journals and conference proceedings. He authored Design and Control of Swarm Dynamics (2016)—the only full-length book on the subject—in Springer’s Complexity Series. He received his Ph.D. from EPFL (Lausanne, Switzerland) in computational science for which he was awarded the prestigious IBM Research Prize in Computational Sciences (2008), and the ERCOFTAC Da Vinci Award Silver Medal (2007). He was a postdoctoral fellow and associate at the Massachusetts Institute of Technology (MIT) and remains a research associate with the Department of Mechanical Engineering at MIT.
Practical information
- General public
- Free
- This event is internal
Organizer
- EESS - IIE
Contact
- Prof. Alcherio Martinoli, DISAL