A fuzzy data-based model for Human-Robot Proxemics

Paper i proceeding, 2016

This work aims at bringing empirical knowledge on Human-Robot Interaction obtained from user studies closer to being integrated into the capabilities of robots currently available on the market. The Takagi-Sugeno-Kang method and results of a user study conducted with thirty two participants were used to build a fuzzy data-based model for Human-Robot Proxemics. The experiment investigated the effect of robot approach distance and angle on perceived human comfort. The proposed model, consisting of a set of rules, fuzzy sets and their parameters, can be used by the robotics community thanks to their formal form. It can also be directly translated into natural language statements. Results of model cross-validation are reported.