Trajectory generation based on human attention for a bartender robot

Abstract

The bartender robot system is designed to mix and serve drinks to customers. However, the motion of a bartender robot can be designed to be more engaging, thus attracting visual attention of a customer during the process. The robot arm trajectory can be created based on the visual-attention aspect in addition to the primary task objective of moving a glass of drink from the initial to the goal position. In order to attract attention of the specific observer, the arm trajectory can be generated from different motion primitives based on the observer's viewing position. This paper proposed the method for generating a trajectory for a bartender robot that is not only based on a basic task requirement of serving drinks but also from human-robot interaction aspect.