Neurorobotics brings together interdisciplinary research in machine learning, robotics and bio-inspired artificial intelligence. Embodying neural models in a physical agent allows interacting with and learning from a complex and multimodal environment. This makes robots an ideal platform for research on topics like multimodal and active sensing, complex object manipulation and navigation, neural reinforcement learning or verbal and non-verbal social interaction. For research in machine learning, robotic control provides a complex and challenging test environment where neural approaches can be evaluated and improved with regard to their ability to adapt, generalize and cope with noisy input. For neuro- and cognitive science research, neurorobotics provides an environment to implement, evaluate and refine models of biological and human information processing.
In the session on Neurorobotics, we want to bring together researchers with diverse backgrounds and expertise to promote collaboration and knowledge sharing in the field of Neurorobotics. The topics include neural control and learning in robotics, sensor fusion, multimodal and active sensing in robots, bio-inspired and developmental robotics, human-robot interaction and collaboration.
A System-Level Brain Model for Enactive Haptic Perception in a Humanoid Robot
Kristin Osk Ingvarsdottir, Birger Johansson, Trond A. Tjøstheim, Christian Balkenius
Clarifying the Half Full or Half Empty Question: Multimodal Container Classification
Josua Spisak, Matthias Kerzel, Stefan Wermter
CycleIK: Neuro-inspired Inverse Kinematics
Jan-Gerrit Habekost, Erik Strahl, Philipp Allgeuer, Matthias Kerzel, Stefan Wermter
Robot at the mirror: learning to imitate via associating self-supervised models
Andrej Lúčny, Kristína Malinovská, Igor Farkaš
Safe Reinforcement Learning in a Simulated Robotic Arm
Luka Kovac, Igor Farkas