TTS (Text-to-Speech) ROS 2 Node

This package lets your robot speak text out loud. Publish a string to the /tts ROS 2 topic and the node synthesizes speech and plays it through your speakers.

It uses Coqui TTS (tts_models/en/ljspeech/vits) — a high-quality neural TTS engine that runs completely offline. No internet, no API key, no account needed.

This guide assumes you are running Ubuntu 22.04 with ROS 2 Humble already installed.

---

How It Works

ros2 topic pub /tts  →  tts_topic_node  →  Coqui TTS  →  aplay  →  speakers

Publish "Hello world" to /tts and your speakers say it.

---

Step 1 — Prerequisites

Check ROS 2 is installed:

ros2 --version

Check your ROS 2 workspace exists:

ls ~/final_project_ws/src

If it does not exist:

mkdir -p ~/final_project_ws/src

---

Step 2 — Install System Dependencies

Install aplay (ALSA audio player) used to play the synthesized WAV file:

sudo apt update
sudo apt install alsa-utils

Test it works:

aplay --version

---

Step 3 — Get the Package

Clone this repository into your ROS 2 workspace:

cd ~/final_project_ws/src
git clone https://github.com/Final-Project-ROS2/tts.git

---

Step 4 — Install Python Dependencies

Install Coqui TTS. On Ubuntu 22.04+ with a managed Python environment, use:

pip install coqui-tts --break-system-packages

If you are inside an active virtual environment (tts_venv or similar), use the venv's pip directly to avoid the system restriction:

~/tts_venv/bin/pip install coqui-tts

The first time the node starts, Coqui TTS will automatically download the tts_models/en/ljspeech/vits model (~100 MB). This only happens once; subsequent runs load from cache.

---

Step 5 — Build the Package

Important: Run colcon build from the workspace root (~/final_project_ws), not from inside the package directory.

cd ~/final_project_ws
colcon build --packages-select tts

You should see:

Starting >>> tts
Finished <<< tts

Load the package into your terminal session:

source ~/final_project_ws/install/setup.bash

Add this line to ~/.bashrc to avoid running it every time:

echo "source ~/final_project_ws/install/setup.bash" >> ~/.bashrc

---

Step 6 — Run the Node

ros2 run tts tts_node

You should see:

[tts_topic_node]: Loading Coqui TTS model...
[tts_topic_node]: TTS Topic Node Ready. Listening on /tts

The node is now running and waiting for text to speak.

---

Step 7 — Send It Some Text

Open a second terminal and publish a message:

ros2 topic pub --once /tts std_msgs/msg/String "data: 'Hello from ROS 2'"

Your speakers should say the words out loud. Change the text inside the single quotes to anything you want.

---

Topic Reference

Topic	Type	Direction	Description
/tts	std_msgs/msg/String	Subscribe	Text to be spoken aloud

---

Troubleshooting

No audio output

Check that aplay can play audio:

aplay /usr/share/sounds/alsa/Front_Left.wav

If you hear nothing, check your system's audio output device.

---

aplay: command not found

sudo apt install alsa-utils

---

ModuleNotFoundError: No module named 'TTS'

pip install coqui-tts

---

The node starts but nothing is spoken

Check the node is receiving messages. In a second terminal:

ros2 topic echo /tts

Then publish in a third terminal. You should see the message appear. If not, the topic name may be wrong.

---

source install/setup.bash — command not found

Make sure you are in the right folder:

cd ~/final_project_ws
source install/setup.bash

---

Quick Reference

Task	Command
Install audio player	sudo apt install alsa-utils
Install Coqui TTS	pip install coqui-tts --break-system-packages
Build	cd ~/final_project_ws && colcon build --packages-select tts
Load into terminal	source ~/final_project_ws/install/setup.bash
Start the node	ros2 run tts tts_node
Send text to speak	ros2 topic pub --once /tts std_msgs/msg/String "data: 'your text here'"

---

License

Apache-2.0