In real world applications a robotic system should be able to interact with humans. In particular it should be friendly and try to understand persons affected by diseases that can not permit them to speak fluently. In this paper a receptionist robot capable to vocally interact with humans that suffer of two types of pathologies, namely deafness and spasticity, is presented. It is composed by a novel auditory system for enhancing the pathological vocal messages and by an audio-video facial animation system which communicates the answer of the receptionist robot to the user through audio-visual modality. The auditory system is composed by a microphone array, a beamformer, an enhancement module (composed by a time scale modification module and emphasis module), and a speech recognizer. Some preliminary results of speech recognition performed with and without the speech enhancement module are reported.
展开▼
