You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When a 'Hey Vector' user command phrase is not recognized or matched with existing intent, no event gets propagated throug the API at all, so all unrecognized events are dropped at the robot backend and clients do not even recognize that a robot command has been invoked, and ended ussuccessfully.
From the current implementation, it seems that the logic of passing the 'unrecognized' intents further to the API has been started, but never finished.
Suggested Resolution
To finish this implementation and pass the transcription of the microphone audio, from the voice recognition service to the intent response, for both intent api methods
WakeWordEnd - when the robot is not under behaviour control
UserIntent - when the robot is under behaviour control
With the actual voice recognition transcription available to api clients through this api endpoints, it would be possible to react to any phrase or keyword at a wish, and to process the whole parsing logic on the clients themselves, regardless of whether the phrase has been successully matched with a known backend intent or not (this will actually skyrocket the extension possibilites).
The text was updated successfully, but these errors were encountered:
Issue description
When a 'Hey Vector' user command phrase is not recognized or matched with existing intent, no event gets propagated throug the API at all, so all unrecognized events are dropped at the robot backend and clients do not even recognize that a robot command has been invoked, and ended ussuccessfully.
From the current implementation, it seems that the logic of passing the 'unrecognized' intents further to the API has been started, but never finished.
Suggested Resolution
To finish this implementation and pass the transcription of the microphone audio, from the voice recognition service to the intent response, for both intent api methods
With the actual voice recognition transcription available to api clients through this api endpoints, it would be possible to react to any phrase or keyword at a wish, and to process the whole parsing logic on the clients themselves, regardless of whether the phrase has been successully matched with a known backend intent or not (this will actually skyrocket the extension possibilites).
The text was updated successfully, but these errors were encountered: