TurtleBot3 Voice Navigation System

The TurtleBot3 Voice Navigation System is designed to enable users to control a TurtleBot3 robot using voice commands or text input through a Telegram bot interface. The system integrates various components, including a Telegram bot, voice transcription, natural language processing, an AI Agent and ROS2 navigation, to provide an intuitive and user-friendly way to navigate the robot in indoor environments.

System Components

The system consists of the following main components:

• Telegram Bot Interface
• Voice Transcription Service
• Plan-and-Execute AI Agent
• ROS2 Bridge
• TurtleBot3 with ROS2 and Nav2

System Architecture

Detailed Component Descriptions

3.1 Telegram Bot Interface (integrated_bot.py)

• Handles user interactions through the Telegram messaging platform
• Processes both voice messages and text input
• Sends voice messages for transcription and text messages directly to the React Agent
• Manages user sessions and provides a welcome message

3.2 Voice Transcription Service (voice_transcribe.py)

• Uses the OpenAI Whisper API to transcribe voice messages to text
• Handles the upload of voice data and retrieval of transcribed text

3.3 Plan-and-Execute AI Agent (react_agent.py)

• Implements a Plan and Execute agent using LangChain and LLM
• Interprets user commands and generates a plan of action
• Uses custom tools to interact with the ROS2 environment:
  • list_locations: Provides a list of available indoor locations
  • get_coordinates: Retrieves coordinates for predefined indoor locations
  • set_goal: Sets a navigation goal for the robot

3.4 ROS2 Bridge (ros2_bridge.py)

• Acts as an interface between the React Agent and the ROS2 environment
• Converts high-level navigation commands into ROS2 messages
• Interacts with the Nav2 stack to send navigation goals
• Provides feedback on the status of navigation tasks

3.5 TurtleBot3 with ROS2 and Nav2

• The TurtleBot3 robot (simulation) running ROS2 and the Nav2 navigation stack
• Executes the navigation commands received through the ROS2 Bridge