Experimental Evaluation of Speech Recognition Technologies for Voice-based Home Automation Control in a Smart Home

Michel Vacher (1) , Benjamin Lecouteux (1) , Dan Istrate (2) , Thierry Joubert (3) , François Portet (1) , Mohamed Sehili (2) , Pedro Chahuara (1)

1 GETALP - Groupe d’Étude en Traduction Automatique/Traitement Automatisé des Langues et de la Parole
2 ESIGETEL - ESIGETEL
3 THEORIS

This paper presents an audio-based interaction technology that lets the user have full control over her home environment and at detecting distress situations for the elderly and frail population. We introduce the PATSH framework which performs real-time recognition of voice commands anywhere in the home and de- tail its architecture and the state-of-the-art processing technologies it employs. This system was evaluated in a realistic Smart Home with three user groups: seniors, visually impaired people and people with no special needs. Results showed the validity of the PATSH approach and shed light on its usability for people with special needs.

Mots clés: Real-time audio analysis; experimental in-situ evaluation; Smart Home; Ambient Assisted Living

Domaines: Intelligence artificielle [cs.AI]

Michel Vacher, Benjamin Lecouteux, Dan Istrate, Thierry Joubert, François Portet, et al.. Experimental Evaluation of Speech Recognition Technologies for Voice-based Home Automation Control in a Smart Home. 4th Workshop on Speech and Language Processing for Assistive Technologies, Aug 2013, Grenoble, France. pp.99-105. ⟨hal-00953244⟩