| On 2022-11-09 at 11:00:00 (Brussels Time) |
Abstract
Swarm robot controllers can be created in manifold ways. I will present two different approaches: an evolutionary swarm robotics method with a task-independent reward and a ROS 2 package for robot swarms.
The first part of the talk focuses on evolving swarm robot controllers using a task-independent reward. In my so-called minimize surprise method, robot swarms are rewarded for prediction accuracy during the evolutionary process instead of for a specific swarm behavior. Easy-to-predict swarm behaviors emerge as a byproduct resulting in behavioral diversity. I will introduce the method and show results ranging from simple simulations to real-world experiments with Thymio II robots.
In the second part of the talk, I will present my ROS 2 package ROS2swarm. ROS2swarm is an easy-to-extend framework for and library of swarm behaviors that allows the quick setup of swarm robot experiments. Swarm behaviors are implemented as basic and combined patterns that can be combined to create complex swarm behaviors. With ROS2swarm, I want to facilitate the use of swarm robotics in research, education, and industry. Short Bio of the Speaker
Tanja Katharina Kaiser is a postdoctoral researcher in the Cyber-Physical Systems Group headed by Prof. Dr.-Ing. Heiko Hamann at the University of Konstanz, Germany. In August 2022, Tanja successfully defended her doctoral thesis at the University of Lübeck, Germany. She holds an M.Sc. degree in computer science from the Technical University of Berlin, Germany, and a B.Sc. degree in computer science from the Baden-Wuerttemberg Cooperative State University Stuttgart, Germany. Her main research interests include evolutionary swarm robotics, open-ended evolution, intrinsic motivation systems, and the application of ROS 2 for robot swarms.
