Un'architettura robotica per l'honest signalling nell'interazione uomo-umanoide

This paper describes the architecture of a detection and classification humanoid integrated system, which is designed to specify one class of biologically salient signals that may contribute to making the human humanoid interaction natural-like and reliable. First the biological theory of the honest signalling is presented. Then the extension of this theory to the sociometrics is outlined. Finally, the paper describes the modules of the architecture that is designed as one generator of models of social perception, which are testable for a defined class of social signals.