A comparison of learning-by-demonstration methods for force-based robot skills

Abstract

Conventional programming techniques for robotic operations in unstructured environments are not suitable for human-like robot executions. In this paper we present a comparison of three different learning-by-demonstration techniques. A force-based representation of the learned data is presented and analyzed. The action to verify the experimental results can be related to Peg-in-Hole action. The results show which method is the most suitable to use for the specified task.