ROS Integration

Rero integrates with ROS 2 via a dedicated package.

Before launching ROS nodes, ensure a rero_server instance is running and reachable at the host/port specified in config.ini.

ROS 2: building the `rero_ros` package

1. Install gRPC

Follow the official gRPC C++ quickstart to build and install gRPC to a prefix such as ~/grpc_install. The current version of gRPC that is used for Rero Core and Rero ROS is v1.73.0.

2. Build with `colcon`

In your ROS2 workspace (e.g., ~/ros2_ws):

colcon build --symlink-install \
  --cmake-args -DGRPC_FETCHCONTENT=OFF \
               -DCMAKE_PREFIX_PATH=~/grpc_install

This tells CMake to use your existing gRPC installation rather than downloading it via FetchContent.

Source the workspace after build:

source install/setup.bash

ROS2 nodes and launch files

All commands below assume your workspace has been sourced.

Audio stream publisher

Streams audio from the microphone array into ROS2 topics:

ros2 launch rero_ros audio_stream.launch.xml

Audio stream subscriber

Subscribes to the audio stream and writes audio to .wav files:

ros2 launch rero_ros audio_subscriber.launch.xml

Direction of arrival (DOA)

Estimates the direction of incoming sound sources:

ros2 launch rero_ros doa.launch.xml

NLU node

Converts recognized text into intents:

ros2 launch rero_ros nlu.launch.xml

Speech recognition

Turns microphone audio into text:

ros2 launch rero_ros speech_recognition.launch.xml

Text-to-speech (TTS)

Synthesizes speech from text messages:

ros2 launch rero_ros text_to_speech.launch.xml

Voice activity detection (VAD)

Detects when speech is present in the audio stream:

ros2 launch rero_ros vad.launch.xml

Testing the ROS2 NLU pipeline

In a new terminal, echo the NLU results:
```
ros2 topic echo /rero_ros/nlu_result
```

Publish a test utterance to the speech recognition topic:

ros2 topic pub /rero_ros/speech_recognition std_msgs/String \
  "data: Please play the beatles"

You should see the corresponding NLU result on /rero_ros/nlu_result, verifying the full speech → NLU pipeline.