Reproducible Human robot interaction and Natural Interaction Interfaces
Currently robots are developed to join us in our homes and become part of the family. To be able to overcome the novelty effect, we need to have a strategy to allow endusers to teach robots new tasks and capabilities.
In a human environment robots need natural interaction interfaces on the one hand and are expected to be reliable and non-invasive on the other hand.
I will talk about features to generate a smooth teaching environment for robots and humans interacting with each other. The special needs for isolated user groups [ASD] will be discussed, as well as deep learning strategies for action description (language) and object learning
Katrin Lohan joined the school of Mathematical and Computer Sciences at Heriot-Watt University as an assistant Professor in 2013. She is deputy director of the Robotics Lab. She became SICSA team leader in the Cyber Physical Systems research theme in 2016. She was General Chair for the European Robotics Forum 2017. She is hired under the Global Platform Recruitment for Research Leaders and part of the Edinburgh Centre for robotics. Previously, she was working at the Istituto Italiano di Tecnologia (IIT) as a Post Doc in the RobotDoc project funded by the Marie Curie Fellowship. She obtained her Ph.D. in Engineering from Bielefeld University, Germany in 2012, where she was associated with the ITALK Project. Her main research interests are in understanding the learning mechanisms between parents and infants, between adults and adults, and between humans and robots in order to create a natural interaction with a robot. Furthermore, she is interested in deep learning of semantic objects, both through vision and speech.