Sustaining Emotional Communication when Interacting with an Android Robot

The more human-like a robot appears and acts, the more users will have the belief of communicating with a human partner rather than with an artificial entity. However, current robotic technology displays limitations on the design of the facial interface and on the design of believable Human-Robot Interaction (HRI), therefore when directly interacting with such robots discomfort might be created in recognizing either the robotic expression of emotion, and/or the probable following action of the robot. Failure of the roboticist to meet the expectations that rise from the anthropomorphic appearance of an android related to its actions, perception and intelligence, or failure to identify which robot type is qualified to perform a specific task, might lead to disruption of HRI. This study is concerned with the problem of sustaining emotional communication when interacting with an android social robot, and consists of a number of rather diverse contributions to this field of research. These contributions include new results and methods in relation to the perception of robots both prior to and after HRI, the evaluation and assessment of robotic platforms and robot properties in relation to specific tasks, the importance of designing android robotic interfaces after actual humans, and the ethics of such a persuasive technology social robots are. The teleoperated android Geminoid-DK, whose appearance resembles a specific actual human, was used as a robotic platform for the conducted experiments (including user studies, field studies, laboratory studies, online surveys, and pre/post surveys).

In the first part of this dissertation, an overview over the purpose and the goals of this study is given, accompanied by background knowledge of the related fields of knowledge, and a reflection on the theories and methods followed. During this first part, the contribution of the dissertation is summarized, and positioned in relation to the nine research papers that follow in the second part. The second part of this dissertation contains: (a) a classification of robots based on the dimensions of Intelligence (Autonomy-Control), and Perspective (Tool-Medium), (b) connections made between social robots and persuasive technology, and between robots and the user’s sense of place attachment, (c) an introduction to the Geminoid Reality including the advances in geminoid technology, (d) a methodology of mapping and evaluating genuine human facial expressions of emotion to androids, (e) a case of designing android faces after specific actual persons who portray facial features that are familiar to the users, and also relevant to the notion of the robotic task, in order to increase the chance of sustaining emotional interaction, (f) an open-ended evaluation method pertaining to the interpretation of Android facial expressions (g) a study on how users’ perception and attitude can change after direct interaction with a robot, (h) a study on how androids can maintain the focus of attention during short-term dyadic interactions, and (i) a state-of-the-art report on android hands.

Throughout the dissertation, the main focus has been to understand the underlying problems that could cause disruption in human-robot communication, and to provide meaningful insights on how to prevent them from happening. In
conclusion, the eight main steps for sustaining emotional communication with an android robot are: (1) Prior to interactions, evaluate the properties of the robot, (2) Prior to interactions, assess the attitude of the users towards the robot, (3) Prior to interactions, know what tasks the robot can satisfy according to its appearance, morphology, and abilities, (4) Robotic speech, lips synchronization, facial expressions and movements need to be aligned, (5) The robot should avoid abrupt movements towards the user, (6) The robot needs to be as capable as it appears to be, (7) Robot appearance matters less if the situation is engaging, (8) The gender of the robot affects the interactions.

Even though the presented eight recommendations need further investigation, and validation, which will happen soon as more android robots are produced and put into actual use in real life situations, I believe that by following them communication with androids will become more meaningful.