ros_voice_assistant

Overview

ros_voice_assistant is a ROS 2 package designed for voice interaction, utilizing Google Text-to-Speech (TTS) and SpeechRecognition to provide a seamless voice assistant experience. It allows users to interact with a robot or system using predefined keywords and phrases while providing customizable responses.

The assistant is configurable, allowing users to change the keyword it listens for and the language it recognizes. Additionally, it comes with built-in support for a variety of commands that users can edit to suit their needs.

Disclaimer

This package uses the Google Text-to-Speech (TTS) service. By using this package, you agree to Google’s Terms of Service and Privacy Policy. Please review these terms before use.

License

The source code is released under a MIT License.

Author: Victor Carvalho

Key Features

• Keyword-based Activation: The assistant activates when it hears the predefined keyword (default: "Chloe"). You can configure the keyword to suit your needs.
• Predefined Phrases: Responds to a set of built-in commands, which are fully customizable.
• Multilingual Support: Supports different languages via the language parameter (default: "en-US").
• Extensibility: Add or modify commands and responses easily within the nlp_node.py.

Applications

The ros_voice_assistant package is versatile and can be used in various voice-interaction scenarios for robotics and automation. Here are some example applications:

1. Personal Assistant for Robots:

   • Add conversational capabilities to robots, enabling them to respond to questions, perform tasks, or provide feedback to users.

2. Voice-Controlled Home Automation:

   • Control smart devices or home appliances through voice commands, using a robot as the central interface.

3. Interactive Conversational Agents:

   • Use the assistant as a chatbot-like interface for entertainment, FAQs, or providing user guidance.

4. Telepresence Robots:

   • Enable natural language interactions during remote communication via telepresence robots.

5. Education and Research:

   • A great tool for teaching robotics and experimenting with voice-enabled human-robot interaction (HRI).

6. Accessibility Tools:

   • Help users with disabilities control robots or devices through voice commands.

7. Voice-Controlled Navigation:

   • Direct mobile robots to specific locations using simple voice commands like "go to the kitchen" or "return to base."

These applications demonstrate the flexibility of the ros_voice_assistant and its potential to enhance human-robot interaction in diverse environments.

Predefined Commands and Responses

Below are the default phrases the assistant understands. These can be edited in the nlp_node.py file.

Command	Default Response
"hello", "hi", "hey"	"Hey there! How's it going today?"
"how are you", "how's it going"	"I'm just a bot, but I'm feeling pretty good! Thanks for asking. How about you?"
"tell me a joke", "make me laugh"	"Why did the robot go on vacation? Because it needed to recharge its batteries!"
"what's the weather"	"I wish I could check the weather for you, but my sensors aren't that advanced yet!"
"what can you do"	"I can chat with you, tell jokes, and maybe even learn new things if you teach me!"
"goodbye", "bye"	"Goodbye! It was nice chatting with you. Come back soon!"
"who are you", "what are you"	"I'm your friendly chatbot! My job is to keep you entertained and assist you."
Unknown Phrase	"Hmm, I'm not sure I understand. Could you repeat that, please?"

Parameters

You can configure the behavior of the assistant by setting the following parameters in the launch file or during runtime.

MicNode Parameters

Parameter	Type	Description	Default
language	string	Language code for speech recognition (en-US, etc.).	"en-US"
keyword	string	The word that activates the assistant.	"Chloe"

SpeakerNode Parameters

Parameter	Type	Description	Default
sync_mode	bool	If True, blocks until the speech finishes.	True

Package Main Files

ros_voice_assistant/
├── voice_assistant/
│   ├── launch/
│   │   └── voice_assistant.launch.py
│   ├── voice_assistant/
│   │   ├── __init__.py
│   │   ├── mic_node.py
│   │   ├── nlp_node.py
│   │   └── speaker_node.py
│   ├── LICENSE
│   ├── package.xml
│   ├── setup.cfg
│   └── setup.py
├── assistant_interfaces/
│   ├── srv/
│   │   └── Text.srv
│   ├── CMakeLists.txt
│   ├── LICENSE
│   └── package.xml
├── LICENSE
├── README.md
└── requirements.txt

Installation and Setup

1. Clone this repository into your ROS 2 workspace:

mkdir -p project_ws/src && cd project_ws  # If you don't have a workspace
git clone https://github.com/victorcarvesk/ros-voice-assistant.git src/ros_voice_assistant

1. Install package system dependencies with rosdep:

rosdep update
rosdep install --from-paths src/ros_voice_assistant -y --ignore-src

1. Create and activate a virtual environment (required on Ubuntu 24.04+):

python3 -m venv .venv --system-site-packages
source .venv/bin/activate

1. Install package Python dependencies:

pip3 install -U -r src/ros_voice_assistant/requirements.txt

1. Build the package with colcon:

colcon build --packages-select assistant_interfaces voice_assistant

1. Launch the voice assistant:

. install/setup.bash
ros2 launch voice_assistant voice_assistant.launch.py

1. Talk to your assistant:

Try commands like:

• "Hello, Chloe!"
• "Chloe, tell me a joke."
• "Chloe, what can you do?"

Customizing the Assistant

Changing the Activation Keyword

Modify the keyword parameter in the launch file:

Node(
    package='voice_assistant',
    executable='mic_node',
    output='screen',
    parameters=[
        {'language': 'en-US'},
        {'keyword': 'Chloe'},  # Change the keyword here
    ],
)

Editing Commands and Responses

The predefined commands and responses are defined in nlp_node.py. You can modify or extend these by editing the nlp_callback method.

Example:

case "play music":
    answer = "Sure! What would you like me to play?"

Feedback and Contributions

Feel free to open issues or submit pull requests to improve this package. Contributions are welcome!