ANC Seminar - Claudio Castinelli


Title: Human-machine {interaction,interfaces}: is it really about machine learning?


Speaker: Claudio Castinelli



After introducing and motivating the talk, I will sketch the current achievements of natural “myocontrol” via adaptive biological HMIs, going into some detail about how to model the relationship between biological signals, the user’s intent to move and control signals for a robotic artefact (in teleoperation, prosthetics, virtual reality). I’ll then define my personal ideal human-machine interface for the disabled, revolving around the concept of natural control and sensori-motor schemes. Optimising the interaction between a user and a machine is not about machine learning, although we cannot do without it, nor it has to do with big data. Rather, it is a matter of collecting good data from the user and/or pushing the user to produce good data, monitoring and exploiting reciprocal adaptation and building appropriate interaction strategies. Lastly, I’ll throw some open question and issues to the audience, hoping to stimulate an articulated discussion.



Claudio is a senior researcher and team leader in rehabilitation and assistive robotics, human-machine interfaces and interaction and applied machine learning. Claudio obtained a degree in Electronic (Biomedical) Engineerings in 1998 from the University of Genoa and a Ph.D. in Artificial Intelligence (Mathematical Logic) in 2005 from the School of Informatics of the University of Edinburgh. He then turned his attention to robotics for the disabled, and after spending 4.5 years as a post-doctoral fellow at the Advanced Robotics Laboratory of the University of Genoa, he landed at the German Aerospace Center where he currently works. As of now, he has (co)authored short of 120 scientific papers, is involved in a few research projects and has served, or is currently serving, for some international editorial boards and committees.





Oct 27 2020 -

ANC Seminar - Claudio Castinelli

Tuesday, 27th October 2020