Robot learning to walk: An architectural problem for intelligent controllers

The authors analyze the design premises of a new control architecture based on the principle of nested hierarchical control which has the feature of learning from the already programmed skills of a teleoperated system. Different architectural principles are explored and compared. It is shown how high level languages can be built from low level knowledge by using cooperation of the already existing skills, new expert knowledge (teaching new heuristics), and self-experience (example based) learning.<
>