Performative Body Mapping for Designing Expressive Robots


This paper explores the challenges and opportunities of skill acquisition for creative robotics, where the required knowledge is highly embodied. We present Performative Body Mapping (PBM) as a suitable methodology for harnessing the movement expertise of trained professionals. We describe the results from a series of workshops to design and train a non-humanlike robot through embodied explorations of possible forms and movements. In addition to the PBM methodology, we propose a method for evaluating expressive robot performers by adapting the Godspeed questionnaire, commonly used in social robotics, which gathers audience feedback on the perception of five properties of interest in creative robotics; anthropomorphism, affective agency, intelligibility, perceived intelligence, and perceived originality. We report on some preliminary results from a first audience study of an early prototype of our robot and discuss the implications for our research.

Proceedings of the 9th International Conference on Computational Creativity (ICCC 2018), 25–29 June, Salamanca, Spain, 280–287