This paper describes an integrated approach to recognizing and generating affect in a humanoid robot, as it interacts with a human user. Robbel et al. present a method for detecting basic affect signals in the user’s speech input and generate appropriately chosen responses on a robot platform. Generally, the emotional responses are prerecorded, but the proposed system seems to allow the generation of custom messages on the fly. The authors’ particular contribution is an integrated mechanism for generating gestures and emotional speech in a humanoid robot, in a dialog setting where the emotional state of the user is being tracked.
The robot is a mobile unit on a self-balancing base and is equipped with two arms, a flexible torso, and a face with a high degree of freedom, with a variety of sensors that support the display of a range of facial expressions. Responses are selected both in terms of content and emotional quality of the voice. Additionally, this approach synthesizes the robot’s gestures and facial expressions to magnify the effect of the emotional state it conveys.
According to the authors, giving physical agents the ability to detect and display emotion allows them to be effectively used in novel application areas--such as child and elderly care, healthcare, education, and beyond; this is the guiding principle of the work. One of the research results demonstrates that meaningful responses were generated with high accuracy for two affective states of the user--excited and bored.
In general, this research work is easy to read and comprehend, and is likely to be useful to anyone interested in robotics research, particularly in human-robot interaction.