This subsystem enables wake word detection, speech recognition using Azure Cognitive Services, and speech synthesis for Pickasso. Commands are published directly to ROS topics for real-time robot control.
- ✅ Wake Word Detection:
"Hey Turtle"(Offline using Vosk) - ✅ Speech Recognition: Azure Cognitive Speech API
- ✅ Speech Synthesis: Azure Text-to-Speech
- ✅ ROS Integration: Publishes commands to
/speech_commands - ✅ Lightweight and Edge-Friendly
| Component | Purpose | Integration Notes |
|---|---|---|
| Microphone | Audio Input | USB Microphone / Respeaker 4-Mic USB Array |
| Speakers | Audio Output | USB/Bluetooth Speaker |
| Compute | Processing Speech | Raspberry Pi 4, Jetson Nano, or Onboard PC |
pip install azure-cognitiveservices-speech pyaudio vosk rospy std_msgsDownload and extract Vosk model (small English model):
wget https://alphacephei.com/vosk/models/vosk-model-small-en-us-0.15.zip
unzip vosk-model-small-en-us-0.15.zip
mv vosk-model-small-en-us-0.15 model- Start ROS:
roscore- Run Wake Word Detector:
python wake_word_detector.py| Topic | Message Type | Purpose |
|---|---|---|
| /speech_commands | std_msgs/String | Publishes recognized commands |
MIT License