Speech and Speaker Recognition for Home Automation: Preliminary Results

International audienceIn voice controlled multi-room smart homes ASR and speaker identification systems face distance speech conditionswhich have a significant impact on performance. Regarding voice command recognition, this paper presents an approach whichselects dynamically the best channel and adapts models to the environmental conditions. The method has been tested on datarecorded with 11 elderly and visually impaired participants in a real smart home. The voice command recognition error ratewas 3.2% in off-line condition and of 13.2% in online condition. For speaker identification, the performances were below veryspeaker dependant. However, we show a high correlation between performance and training size. The main difficulty was the tooshort utterance duration in comparison to state of the art studies. Moreover, speaker identification performance depends on the sizeof the adapting corpus and then users must record enough data before using the system