23 March 2022
by Theodoros Stouraitis, University of Edinburgh
I’m Theo, a research associate at the Statistical Machine Learning and Motor Control (SLMC) group at the University of Edinburgh and I am about to complete a three-month research visit at Interactive Robotics Group (IRG), at the Massachusetts Institute of Technology (MIT), Cambridge, MA, USA. During my research visit, Prof. Julie A. Shah who leads the IRG, Shen Li, a PhD student at IRG, and I jointly worked on a project that we started remotely during the pandemic.
Our joint project is focused on modelling the uncertainty of humans’ behaviours when interacting with robots. Equipping robots with the ability to estimate the uncertain state of the human and predict the uncertain motion of the human is key to guaranteeing human safety during interaction. We work on methods that describe the uncertainty with sets and we develop uncertainty-aware algorithms for estimation and prediction of the human behaviour as well as robot motion generation.
As an example, let’s consider a robot helping a human getting dressed. During dressing the robot cannot directly see the position of the human’s elbow due to the cloth, neither can predict where the human arm will be next. This in turn complicates the robot motion generation as the robot might be overconfident in an erroneous human state that might render the robot’s behaviour unsafe. Modelling uncertain human behaviour allows robots to decide when to move conservatively and when efficiently such that the human partner is always safe.
My three-month visit started in January 2022, and it will be completed at the end of March 2022. In late December 2021 and during the Omicron wave I packed my suitcase and left cold Edinburgh for the even colder Boston. The first couple of weeks were about adjustment to the new town and weather, hunting for a room and getting familiar with MIT and the IRG, while the following weeks were full of focused work, meeting students, academics and fellow roboticists at MIT, as well as exploring the particularly cold (for a Greek) and vibrant Boston. My visit has been extremely stimulating and productive and I would definitely recommend it to any researcher, especially if the visit takes place in another continent. After the completion of my visit, our collaboration between the IRG, the SLMC group, and the Honda Research Institute Europe will continue towards developing further human uncertainty models.
Finally, I would like to thank SICSA for funding this research visit via the SICSA Postdoctoral and Early Career Researcher Exchanges (PECE) award, which greatly supported me during this fruitful research visit!