Research
The SocRob team is mainly driven by research on the following topics:
Perception and sensor fusion, that is, the problem of extracting meaningful information from one or more sensors, and how to fuse together information from several sources, taking into account their associated uncertainty. We take a probabilistic approach on quantitatively representing uncertainty, as well as performing inference, on the solid theoretical framework provided by Bayesian statistics;
Decision making under uncertainty, since both perception and action affects on the real world are inherently uncertain, we take a decision-theoretic approach on the maximization of the expected utility;
Human-robot interaction, once robots operate in human-inhabited environments, robots need to interact with humans; in this context we pursue multiple communication modalities, including speech, web-based remote interface, and onboard touch interface;
Manipulation and grasping, that is, the problem of robots being capable of grasping and manipulating real world objects in human; here we are interested on topics ranging from finger design, up to motion planning;
Publications