How Could Robots Successfully Elicit Help from Humans?
With the rapid growth of robot usages, more attention has been focused on the robots’ hardware and software to improve functionality. However, the interaction between humans and robots can be key to enhancing robots’ performance. Autonomous mobile robots may more successfully complete their tasks if they effectively solicit help from humans when they reach their limits. Hence, in this thesis, we examine three different manners, friendly, polite, and indirect, combining different interaction modalities, like speech and gestures. We conducted an experiment (involving 45 subjects) where the NAO robot addresses bystanders with the aforementioned three manners to elicit help with opening a public door. The aim of the research is to investigate strategies for robots to successfully request help from humans. We investigate whether passers-by humans respond differently to various manners. We attempt to establish procedures to evaluate and compare manners by which robots could ask humans efficiently for help. These evaluations are based on the responses of bystanders to the robot and their answers for a related questionnaire. In addition, we investigate the effects of bystanders’ source of control assumptions, focusing on analyzing the situational awareness desire declared by the bystanders experiencing the robot request.