Programming-by-Demonstration of Robot Motions Alexander Skoglund, Boyko Iliev,  and Rainer Palm  Orebro University, Sweden Abstract. In this paper a novel approach to skill acquisition from hu- man demonstration is presented. Usually the morphology of a robot ma- nipulator is very different from the human arm and cannot simply copy a human motion. Instead the robot has to execute its own version of the skill demonstrated by the operator. Once a skill has been acquired by the robot it must also be able to generalize to other similar skills without starting a new learning process. By using a motion planner that oper- ates in an object-related world-frame called hand-state, we show that this representation simplifies a skill reconstruction and preserves the essential parts of the skill. Keywords: Programming-by-Demonstration, Motion Planner, Hand-state, Fuzzy Modeling, Correspondence Problem. 1 Introduction This article presents a method for imitation learning based on fuzzy model- ing and a next-state-planner in a Programming-by-Demonstration (PbD) frame- work. For a recent comprehensive overview of PbD, (also called Learning from Demostration) see [1]. PbD refers to a variety of methods where the robot learns how to perform a task by observing a human teacher, which greatly simpli- fies the programming process [2], [3], [4] and [5]. One major scientific challenge in PbD is how to make the robot capable of imitating a human demonstra- tion. Although the idea of copying human motion trajectories using a simple teaching-playback method seems straightforward, it is not realistic for several reasons. Firstly, there is a significant difference in morphology between the hu- man and the robot, known as the correspondence problem in imitation [6]. The difference in the location of the human demonstrator and the robot might force the robot into unreachable parts of the workspace or singular arm configurations even if the demonstration is perfectly feasible from human viewpoint. Secondly,  Boyko Iliev and Alexander Skoglund are with the AASS, Dept. of Technology Orebro University SE-70182 Orebro, Sweden, Emails: {alexander.skoglund,boyko.iliev}@aass.oru.se  Rainer Palm is adjunct professor at the AASS, Dept. of Technology Orebro Univer- sity SE-70182 Orebro, Sweden, Email: rub.palm@t-online.de