Speech Recognition

Overview

Speech Recognition is run through a node called speech_recognition_core.py in the cart_endpoints/scripts folder. It is in charge of listening to the microphone and translating what it hears into a string that is then either passed to networking for selecting a location or used to execute a variety of commands. THIS NODE REQUIRES AN INTERNET CONNECTION, as it uses google services to do its speech to text conversion.

Dependencies

Pip Installables

• pip2.7 install --user
  • pyaudio
  • wave
  • SpeechRecognition
  • python-vlc
  • gtts

Launching Speech Recognition

• rosrun cart_endpoints speech_recognition_core.py
• It will then be listening for you to say something, it uses the default microphone for the computer, typically any external microphone then if none available the internal microphone.
• The keyword to activate commands is "Auto Cart" like how you say "hey google" or "alexa"
• Current commands include "hello", "help", "emergency", "resume", "cancel", etc.