Evolutionary robotics is a field of developing robots using the principles of evolution. From the principles of evolution in nature, it is possible to derive a number of concepts and strategies for solving learning tasks and developing optimization strategies for artificial intelligent systems. The purpose of this talk is to give a survey of applications where the principles of evolution are used to solve real practical problems. In the first part of the presentation, the application of evolutionary reinforcement learning (a general framework for evolving the policy or value function to be learned) in solving robotic learning tasks will be presented. To illustrate the concept further, an example of learning to walk forward by a robotic insect and learning to balance double poles mounted on a single cart will be presented. In the second part of the talk, evolutionary methods that can evolve both the morphology and control of robotic agents will be discussed. Different methods from literature that evolved the morphology and control of agents will be explained.
VeranstaltungsortRaum Seminarraum 117, Robert-Hooke-Str. 5 in Bremen
In der Regel sind die Vorträge Teil von Lehrveranstaltungsreihen der Universität Bremen und nicht frei zugänglich. Bei Interesse wird um Rücksprache mit dem Sekretariat unter sek-ric(at)dfki.de gebeten.