Robot Errors and Human Teachers: The Effects of Personality and Patience During Learning Tasks

Robots are expected to become a ubiquitous technology in the near future, where different people from a wide variety of backgrounds may have daily interactions with robots. These co-present humans will expect to be able to customize robot behavior to suit their preferences and needs through intuitive methods, such as learning from demonstration (LfD). However, these interactions will vary depending on the personal qualities of the user, and to the best of our knowledge, no work has explored how these qualities affect the teaching of a robot during an LfD task. This paper introduces a novel experimental design for testing patience during LfD instruction, and explores the relationship between personality and patience while teaching an autonomous DARwIn-OP robot. Another contribution of this paper is the introduction of systematic error simulation during LfD.