AI Smart Home

AI Smart Home is an open source smart home system developed in Python, featuring an AI-powered voice assistant (powered by GPT4o) and a modular architecture. The system integrates various functionalities such as music playback, diary entry creation and playback, sensor readings (temperature, humidity, battery status), alarm clock management, light control (via Tinytuya), and a web console for remote monitoring and control.

This project was developed as part of a Bachelor’s degree in Computer Science and is primarily designed to run on a Raspberry Pi 4B, though it can be adapted to work with other hardware setups.

---

Table of Contents

• Features
• Technologies
• Installation
• Configuration
• Usage
• File Structure
• Troubleshooting
• Contributing
• License
• Future Improvements
• About

---

Features

• Voice Assistant:

  • Activated with a wake word (default is "assistant").
  • Processes voice commands and integrates with OpenAI GPT4o for intelligent responses.

• Music Player:

  • Streams and plays music from YouTube using yt-dlp and mpv.
  • Example command: "Play music [Song Name]".

• Diary Module:

  • Records diary entries via audio.
  • Converts recordings to text and saves both audio and text files with the current date.
  • Supports playback of diary entries.

• Sensor Module (Hygrometer):

  • Retrieves temperature, humidity, and battery status from a Bluetooth-enabled hygrometers from Govee.
  • Uses asynchronous BLE scanning with the bleak library.

• Alarm Clock:

  • Schedules and manages alarms using the schedule module.
  • Triggers alarms with voice notifications and custom alarm sounds.
  • Commands for setting and deleting alarms.

• Light Control:

  • Controls Tuya-compatible smart lights via the Tuya Cloud API (using tinytuya).
  • Supports adjusting color, brightness, and operating mode.

• Web Console:

  • Provides a web interface for remote system monitoring.

---

Technologies

• Programming Language: Python 3.11
• Libraries:
  • speech_recognition for voice command detection.
  • pygame for audio playback and TTS.
  • gTTS for text-to-speech conversion.
  • yt-dlp and mpv for streaming and playing YouTube audio.
  • bleak for Bluetooth Low Energy (BLE) scanning.
  • tinytuya for controlling Tuya-compatible devices.
  • OpenAI for GPT4o integration.
• Target Hardware:
  • Raspberry Pi 4B (primarily)
  • Razer Seiren Mini Microphone
  • Soundcore Boom 2 Bluetooth Speaker
  • Govee H5075

Installation

Prerequisites

• A Raspberry Pi 4B running Raspberry Pi OS (or compatible hardware).
• Python 3.11 installed.
• Miniforge3 (or an alternative Python environment manager).
• Internet access for API queries and music streaming.

Installation Steps

1. Clone the Repository:

   git clone https://github.com/Luke-Werle-99/AI_Smart_Home.git
   cd AI_Smart_Home

2. Create and Activate a Python Environment:

   conda create -n AI_Smart_Home python=3.11
   conda activate AI_Smart_Home

3. Install Dependencies:

   cd installation
   sh install_dependencies.sh
   sh install_libraries.sh

4. Configure API Keys:

   • Create or edit the .env file in the project root:

     OPENAI_API_KEY=your_openai_api_key_here

   • Also ensure that any other required configurations (e.g., Tuya credentials) in the corresponding modules are correctly set.

---

Configuration

• Environment Variables:
  The .env file contains sensitive data such as the OpenAI API key. Make sure this file is configured correctly.
• Audio Settings:
  • Ensure the correct audio devices (microphone and speakers) are selected.
  • On the Raspberry Pi, additional configuration (e.g., PulseAudio settings) might be necessary.
• Bluetooth & Sensors:
  • Enable Bluetooth and ensure your Govee hygrometer is properly paired.
• Tuya Cloud:
  • Update your Tuya Cloud credentials (DEVICE_ID, ACCESS_ID, ACCESS_SECRET, API_REGION) in the relevant modules.

---

Usage

1. Start the Assistant:

   python main.py

   The voice assistant will start and listen for the wake word (default "assistant").

2. Available Voice Commands:

   • Wake the Assistant: "assistant" (optionally followed by a command).
   • Play Music: "Play music [Song Name]".
   • Diary Functions: "create diary" to start a new diary entry, "read diary" to playback an entry.
   • Sensor Readings: "temperature", "humidity", or "battery" to fetch current sensor values.
   • Alarm Clock: "set alarm" to create an alarm and "delete alarm" to remove an alarm.
   • Light Control: "set light" followed by commands such as "on", "off", color names (e.g., "red", "blue"), or brightness percentages.
   • General Queries: All other queries are forwarded to OpenAI GPT4o.

3. Systemd Service (Optional):
   To have the assistant start automatically on system boot, set up a systemd service as described below.

Example Interaction

1. Say "Assistant" to wake up the system.
2. Use commands like:
   • "Play music Believer by Imagine Dragons."

You also wake the system and input your command in one sentence like: "Assistant Play music Believer by Imagine Dragons"

---

Alternatively, you can configure it to start automatically using a systemd service file (see below).

Systemd Service Setup

To run the assistant on system startup:

1. Create a systemd service file:

   sudo nano /etc/systemd/system/autostart.service

2. Add the following content:

   [Unit]
   Description=Start AI Smart Home Assistant
   After=sound.target bluetooth.target network.target
   
   [Service]
   User=your-username
   WorkingDirectory=/home/your-username/ai-smart-home
   ExecStart=/home/your-username/miniforge3/condabin/conda run -n AI_Smart_Home --no-capture-output python main.py
   Environment="XDG_RUNTIME_DIR=/run/user/1000"
   Environment="DBUS_SESSION_BUS_ADDRESS=unix:path=/run/user/1000/bus"
   Restart=on-failure
   
   [Install]
   WantedBy=multi-user.target

3. Enable and start the service:

   sudo systemctl daemon-reload
   sudo systemctl enable autostart.service
   sudo systemctl start autostart.service

4. Additinally you can set up automatic bluetooth pairing in case you are using a bluetooth speaker like this:

       echo "Connecting to Bluetooth speaker..."
       for i in {1..5}; do
       echo "Attempt $i: Connecting to F4:2B:7D:29:50:41..."
       bluetoothctl -- connect F4:2B:7D:29:50:41 && break
       sleep 5
       done

File Structure

AI_Smart_Home/
├── main.py
├── Modules/
│ ├── init.py
│ ├── music_player.py
│ ├── diary.py
│ ├── hygrometer.py
│ ├── alarm_clock.py
│ ├── light_control.py
│ └── web_console.py
├── diary_entries/
│ ├── YYYY-MM-DD.txt
│ ├── YYYY-MM-DD.wav
├── installation/
│ ├── install_dependencies.sh
│ └── install_libraries.sh
├── .env
└── README.md

Troubleshooting

• Audio Playback:

  • Ensure pygame is properly initialized.
  • If you encounter issues in subprocesses, consider using threads instead of processes to share the Pygame mixer context.

• Microphone Issues:

  • Verify the device index using arecord -l (on Linux) and adjust the settings in the listen() function if needed.
  • Adjust parameters like energy_threshold or dynamic_energy_threshold in the speech_recognition library.

• Bluetooth and Sensor Readings:

  • Ensure Bluetooth is enabled and your hygrometer is paired correctly.
  • Watch for potential conflicts between asynchronous BLE scans (using bleak) and other async operations.

• Bluetooth Speaker Issues:

  • Ensure the speaker is paired and set as the default audio sink using:

    pactl set-default-sink [sink-name]

• OpenAI API:

  • Confirm that your API key is correct and that you have an active network connection.
  • Check logs for any error messages related to API requests.

• General Debugging:

  • Monitor console logs to identify error messages.
  • Review the specific module in the Modules/ directory for module-related issues.

Contributing

Contributions to AI Smart Home are welcome!

To contribute:

1. Fork the repository.
2. Create a new branch for your feature or bug fix.
3. Commit your changes with clear, descriptive messages.
4. Open a Pull Request describing your changes and the benefits they provide.

For any questions or suggestions, please open an issue on the GitHub repository.

License

This project is licensed under the MIT License.

---

About

AI Smart Home was developed by Luke Werle as a Bachelor’s degree project in Computer Science. The aim is to provide a privacy-friendly and modular alternative to commercial smart home systems. Feedback and suggestions for improvements are always welcome!

---

Enjoy your AI Smart Home Assistant!