This project is a "smart vacuum" system that includes a mobile app for control and a Python backend for AI-powered features.

• AI-Powered Pathfinding: The system uses a Python backend with a custom fine-tuned Yolo model to analyze a floor plan and calculate an optimal cleaning path.
• Mobile App Control: A SwiftUI-based iOS app provides a user-friendly interface to control and monitor the vacuum.
• Two Cleaning Modes:
  • Manual Mode: An augmented reality view where you can manually drive the vacuum using a joystick. This mode also uses the device's camera to detect "dirt" or "waste" on the ground and visualizes it in the AR scene.
  • SafePath Mode: A simulation view that displays the calculated cleaning path on the uploaded floor plan. You can watch the vacuum's progress and see real-time telemetry data like cleaning coverage and efficiency.
• Floor Plan Analysis: The backend can take an image of a floor plan and convert it into a grid-based map for pathfinding.

The project is divided into two main components:

• api-backend.py: A Python backend built with FastAPI that handles the AI and pathfinding logic.
• SmartVacuum/: An Xcode project for the iOS app, written in Swift and SwiftUI.

The backend provides the following API endpoints:

• POST /predict: Takes an image and returns a list of detected objects with their bounding boxes.
• POST /get_floor_plan: Converts a floor plan image into a binary matrix.
• POST /calculate-path: Calculates a cleaning path based on a floor plan matrix.

The iOS app has the following features:

• Floor Plan Upload: Users can upload a floor plan image from their photo library.
• Connection Simulation: The app simulates connecting to the smart vacuum.
• Mode Selection: Users can choose between "Manual" and "SafePath" modes.
• Manual Control: An AR view with a joystick for manual control and dirt detection.
• Path Simulation: A view that visualizes the cleaning path on the floor plan.

1. Install the required Python packages:

   pip install fastapi "uvicorn[standard]" python-multipart opencv-python-headless ultralytics Pillow

2. Run the backend server:

   python api-backend.py

The server will start on http://0.0.0.0:8000.

1. Open the SmartVacuum.xcodeproj file in Xcode.
2. Select your target device and run the app.

The project also includes Python scripts for visualizing the generated path:

• path_visualization.py: A script to visualize the path from a hardcoded API response using matplotlib.
• visualization.py: Another visualization script.

These scripts are likely for development and debugging purposes.