Description of the ANR VocADom project


Robust voice command adapted to the user and to the context for AAL

ABSTRACT: The industrial research project VOCADOM aims to define, in conjunction with end users, the features of a distant voice controlled home automation system that will adapt to the user and that can be used in real conditions (noise, presence of several people). To solve the challenges related to distant voice control and associated decision, robust integration of all available information (audio and home sensors) is needed. The partnership combines three research teams who are experts in speech and audio processing, artificial intelligence and methodology for user studies. In addition, a company in the field of embedded computing, and one subcontractor – a user society connected with elderly or visually impaired people – are involved to assure the socio-economic impact of the project. VOCADOM will dispose of one dedicated smart home in order to validate the experimental proof of concept with end users, in complement, some experiments will be made at personal home of such volunteer users.

Keywords: Personalized voice interaction in context; Robust voice command recognition in noisy environments; Home automation assistance for home autonomy; Decision-making in context; Automatic adaptation to the person’s capacities and needs; Pre-study, development and validation in collaboration with users

Project leader: Michel VACHER (2017-2019), François PORTET (2019-2021).

logo_ANR new_logo-cnrs    logo_lig logo_getalp-h100    logo_INRIA     logo-GRePS-Vectoriel   logo_TEORIS