Figure 9.

iCub solves a “Table of Hanoi” task, moving the small, medium and big object, from the Middle location to the Left location, after learning the rules that a larger object cannot go to a location where there is already a smaller object. These rules are used in a standard planner to allow the system to solve new instances of the task small-s, medium-m, big-b, middle-M, left-L, right-R. Intially all objects are at M, and the goal is to move them to L. From Petit et al. (2016).