This project employs a user interface that is fully controllable by voice commands. For this, it implements a voice-to-text interface powered by the PocketSphinx library (see https://cmusphinx.github.io/) that uses both keyword-spotting and a grammar search in order to recognize and interpret commands given to it.
We use the default US-English language model that comes with the PocketSphinx distribution. This means that any commands spoken should be pronounced with a US-English accent in order to be accurately understood. To change this, one can load a different language model aimed at different languages and accents. Various models are available via the PocketSphinx website, and it is possible to make one's own language model with some effort.
