Quantitative analysis of robot gesticulation behavior
This addresses the need for robust quantitative evaluation methods in social robotics to improve gesture generation, though it is incremental as it adapts existing metrics to a specific domain.
The paper tackles the problem of quantitatively comparing generative adversarial network (GAN)-based approaches for robot gesture generation, proposing a new Fréchet Gesture Distance and using Principal Coordinate Analysis and procrustes statistics to assess fidelity and originality of gestures.
Social robot capabilities, such as talking gestures, are best produced using data driven approaches to avoid being repetitive and to show trustworthiness. However, there is a lack of robust quantitative methods that allow to compare such methods beyond visual evaluation. In this paper a quantitative analysis is performed that compares two Generative Adversarial Networks based gesture generation approaches. The aim is to measure characteristics such as fidelity to the original training data, but at the same time keep track of the degree of originality of the produced gestures. Principal Coordinate Analysis and procrustes statistics are performed and a new Fréchet Gesture Distance is proposed by adapting the Fréchet Inception Distance to gestures. These three techniques are taken together to asses the fidelity/originality of the generated gestures.