From Sensorimotor Primitives to Manipulation and Imitation Strategies in Humanoid Robots

Abstract

Regardless of the application area, one of the common problems tackled in humanoid robotics is the investigation towards understanding of human-like information processing and the underlying mechanisms of the human brain in dealing with the real world. Therefore, we are building humanoid robots with complex and rich sensorimotor capabilities as the most suitable experimental platform for studying cognitive information processing. The target system is supposed to interact and function together with humans. It is meant to be able to cooperate and to enter a dialogue with them. Therefore it needs to understand both what it perceives and what it does. To achieve this goal we are following a holistic approach and investigating how different partial results, achieved in sub-disciplines related to the development of humanoids, fit together to achieve complete processing models and integrative system architectures, and how to evaluate results at system level rather than focusing on the performance of component algorithms. In this paper we present our recent and current progress in this direction. For doing so, we present our work on the development of humanoid platforms, the programming of grasping and manipulation tasks, the multi-sensory object exploration as well as the learning of motor knowledge from human observation.