Atti del VIII° Convegno dell’Associazione Italiana Scienze della Voce AN ITALIAN EVENT-BASED ASR-TTS SYSTEM FOR THE NAO ROBOT Piero Cosi*, Giulio Paci*, Giacomo Sommavilla*, Fabio Tesser* Marco Nalin**, Ilaria Baroni** *Institute of Cognitive Sciences and Technologies, ISTC, C.N.R., UOS Padova, Italy ** Hospital San Raffaele, Milano, Italy *[piero.cosi, giulio.paci, giacomo.sommavilla, fabio.tesser ]@pd.istc.cnr.it **[ nalin.marco, baroni.ilaria ]@hsr.it 1. ABSTRACT This paper describes an event-based integration approach for building a human-robot spoken interaction system using the NAO robot platform with the URBI middleware within the ALIZ-E project. The ALIZ-E integrated system includes various components but we mainly concentrate on the Automatic Speech Recognition (ASR) and the Text To Speech (TTS) synthesis modules while the following Natural Language Understanding (NLU), Di- alog Management (DM) and Natural Language Generation (NLG) ones will be only briefly introduced. We describe these components and how we adapted and extended them for use in the system. We discuss several options that we have considered for the implementation of the interfaces and the integration mechanism and present the event-based approach we have chosen. We describe its implementation using the URBI middleware. The system has been be used for HRI experiments with young Italian users since April 2011. 2. INTRODUCTION Conversational systems play an important role in scenarios without a keyboard, e.g., when talking to a robot. Communication in human-robot interaction ultimately involves a combination of verbal and non-verbal (gesture) inputs and outputs. Systems capable of such an interaction thus must process verbal and non-verbal observations in parallel, as well as execute verbal and non-verbal actions accordingly in order to exhibit synchronized behav- iors. The development of such systems involves the integration of potentially many compo- nents and ensuring a complex interaction and synchronization between them. NAO (Aldebaran Robotics, 2012; Gouaillier et alii, 2008) is one of the most often used humanoid robots for academic purposes worldwide and it is complete with a user-friendly programming environment. From simple visual programming to elaborate embedded mod- ules, the versatility of NAO and its programming environment enables users to explore a wide variety of subjects at whatever level of programming complexity and experience.