We believe that voice will play a huge role in making robots more useful.
When designing Aido, we wanted users to talk to Aido, and Aido reply back in a conversational & more personal manner:
- Aido has a 6 mic array set around a circular base with a dedicated hardware chip for listening to human voice, even in noisy environments. It is the same design that many smart TVs use to allow users to control functions using voice.
- Aido uses a leading 3rd party speech engine based on a custom deep neural net architecture for continuous speech recognition.
- This allows Aido to listen to commands and parse them into instructions that Aido understands.
Aido can listen to voice commands and reply as well, perfect when you are busy with something else
While there has been great progress in voice recognition and response for smart devices, many of the interactions still come off with the device sounding robotic. Voice is a key part of personality, and we wanted Aido to sound as friendly as it is:
- We broke down most commonly used sentences and phrases and created a library for these
- We will have a voice artist render these, along with commonly used words
- Aido will store a library of the words, phrases, sentences as well as the audio files
- Aido will use these snippet in real time to talk back to the user
- We expect we can cover most of the daily interactions this way. For the bits that we don’t have voice snippets for, we will use a standard text-to-speech engine.