Framework for Learning and Adaptation of Humanoid Robot Skills to Task Constraints

Abstract

Humanoid robots are expected to work and collaborate with humans performing in changing environments. Developing this kind of robots requires them to display intelligent behaviors. For behaviours to be considered as intelligent they must at least present the ability to learn skills, represent skill’s knowledge, and adapt and generate new skills. In this work a framework is proposed for the generation and adaptation of learned models of robot skills for complying with task constraints. The proposed framework is meant to allow: for an operator to teach and demonstrate to the robot the motion of a task skill it must reproduce; to build a knowledge base of the learned skills knowledge allowing for its storage, classification and retrieval; to adapt and generate learned models of a skill, to new context, for compliance with the current task constraints. A learning from demonstration approach is employ to learn robot skill by means of probabilistic methods, encoding the motion dynamics in a Gaussian Mixture Model. We propose that this models of the skill can be operate and combine to represent and adapt the robot skills.