metadata

title: DeepDrone
emoji: 🚁
colorFrom: green
colorTo: green
sdk: streamlit
sdk_version: 1.41.1
app_file: main.py
pinned: false

DeepDrone

A drone chat agent for drone analytics and operations, built on the smolagents framework with DroneKit integration for real drone control.

Features

Drone Chat: Interact with a drone assistant through a chat interface
Visualizations: Generate flight paths and sensor readings visualizations
Maintenance Recommendations: Get maintenance suggestions based on flight hours
Mission Planning: Generate mission plans for various drone operations
Real Drone Control: Connect to and control real drones using DroneKit
- Take off and land
- Navigate to GPS coordinates
- Return to home
- Execute waypoint missions
- Monitor battery and location

Getting Started

Clone this repository
Copy .env-example to .env and add your Hugging Face API token
Install dependencies: pip install -r requirements.txt
For Python 3.10+ users: Run the compatibility patch: python dronekit_patch.py
Run the application: streamlit run main.py

Using DroneKit Integration

The DroneKit integration allows you to control drones running ArduPilot or PX4 firmware.

Python 3.10+ Compatibility

If you're using Python 3.10 or newer, you need to run the patch script before using DroneKit:

python dronekit_patch.py

This script fixes the "AttributeError: module 'collections' has no attribute 'MutableMapping'" error by patching the DroneKit library to use collections.abc instead of collections.

Simulation

To test the drone control features in simulation:

Install ArduPilot SITL simulator (follow instructions at https://ardupilot.org/dev/docs/setting-up-sitl-on-linux.html)
Start a simulated drone: sim_vehicle.py -v ArduCopter --console --map
Run the example script: python drone_example.py

Note: The simulator must be running before you attempt to connect with DeepDrone.

Real Drone Connection

To connect to a real drone:

Ensure your drone is running ArduPilot or PX4 firmware

Connect using one of these methods:

Via Terminal

# For direct USB connection
python drone_example.py --connect /dev/ttyACM0  # Linux
python drone_example.py --connect COM3  # Windows

# For WiFi/Network connection
python drone_example.py --connect tcp:192.168.1.1:5760

# For telemetry radio connection
python drone_example.py --connect /dev/ttyUSB0

Via Chat Interface

Use natural language commands in the DeepDrone chat:

"Connect to drone at tcp:192.168.1.1:5760"
"Connect to drone using USB at /dev/ttyACM0"
"Connect to drone via telemetry at /dev/ttyUSB0"

Once connected, you can control the drone with commands like:

"Take off to 10 meters"
"Fly to latitude 37.7749, longitude -122.4194, altitude 30 meters"
"Return to home"
"Land now"

Troubleshooting

collections.MutableMapping error: Run python dronekit_patch.py to fix the DroneKit library for Python 3.10+
Connection refused error: Ensure the drone or simulator is powered on and the connection string is correct
Import errors: Verify that DroneKit and PyMAVLink are installed (run pip install dronekit pymavlink)
Permission errors: For USB connections on Linux, you may need to add your user to the 'dialout' group or use sudo

IMPORTANT: Always follow safety guidelines when operating real drones.

Tech Stack

smolagents for agent functionality
Hugging Face's Qwen2.5-Coder model for natural language understanding
DroneKit-Python for real drone control
Streamlit for the user interface
Pandas, Matplotlib and Seaborn for data analysis and visualization

Spaces:

evangelosmeklis
/

deepdrone

Running