Announcement: Update for Google's Dialogflow-v2 and Speech-To-Text API for ROS

General
, ,
1

Hi everyone,

Last year I wrote a package for Google’s STT API and NLP platform Dialogflow. The package worked well but had two issues:

  1. It required you to use Google’s STT API to get text and then send it to Dialogflow.
  2. It can only be run locally.

With this updated package you can:

  1. Send audio directly to Dialogflow through your mic.
  2. Send/receive audio through a server that you can run on another computer/robot on your network so that you don’t have to run everything on one machine.
  3. Exposed access to contexts, intents, query text, confidence, and cleaner message format.

There is no need to ROS-ify the Google STT API anymore since Dialogflow does most of the work. However, if you’re interested in using the beta/ML features for more fine tuned speech recognition, then I left a script that can be used if you’re interested.

Documentation can be found here: https://wiki.ros.org/dialogflow_ros
Repo here: https://github.com/piraka9011/dialogflow_ros

The installation instructions are the same and there is an install.sh script you can run which takes care of most of the work, except for setting up your Google credentials (instructions in docs).

Again, I’d love to get some feedback on how to enhance this and what features to add.

Best,
Anas