DroneKit-Python Documentation

DroneKit-Python is a library used to create applications that communicate with flight controllers (like ArduPilot) via the MAVLink protocol. It is designed for two primary use cases:

1. Onboard Companion Computers: Running on hardware physically connected to the vehicle via a low-latency link to perform computationally intensive tasks like computer vision, path planning, or 3D modelling.
2. Ground Station Apps: Communicating with vehicles over higher-latency RF links.

The API provides programmatic access to telemetry, state, and parameters, allowing for mission management and direct vehicle control.

DroneKit-Python is designed for two primary deployment scenarios:

1. Companion Computers: Running on Linux-based hardware attached to a vehicle, communicating with the autopilot via a serial port.
2. Ground Stations: Running on Linux, Windows, or macOS, communicating with the vehicle via WiFi or a telemetry radio.

Development Workflow: During development, the recommended pattern is to run DroneKit-Python on a development computer and connect it to a Simulated Vehicle (SITL) running on the same machine via a UDP connection.

DroneKit-Python is a Python implementation of DroneKit that allows developers to create applications for Unmanned Aerial Vehicles (UAVs). It communicates with vehicles using the MAVLink protocol.

Key capabilities include:

• Programmatic access to vehicle telemetry, state, and parameter information.
• Mission management.
• Direct control over vehicle movement and operations.

It is designed for use on onboard companion computers (supporting computer vision, path planning, etc.) or for ground station applications communicating via RF-links.

DroneKit-Python is compatible with vehicles using the MAVLink protocol.

• Autopilots: While compatible with various MAVLink-based systems, it is most highly validated against the ArduPilot UAV Platform. Other autopilots may exhibit compatibility issues due to different MAVLink specification interpretations.
• Operating Systems: Runs on Linux, Mac OS X, and Windows.

If you build SITL from source (on Linux, Windows, or Mac) instead of using the dronekit-sitl package, note the following differences:

• MAVProxy: Included and started by default. You can use the MAVProxy terminal to control the autopilot.
• Connection: You connect via UDP on 127.0.0.1:14550 instead of TCP on 5760.
• Configuration: You may need to manually disable arming checks or load autotest parameters.
• Features: It is easier to add virtual rangefinders or virtual gimbals for testing.

The primary mechanism for being notified of changes in vehicle state in DroneKit is through attribute observers. You can register a callback function that is triggered whenever a specific attribute (e.g., location) changes.

Alternatively, you can use the Vehicle.on_attribute decorator for a more elegant implementation.

# Registering a listener for the 'location' attribute
self.vehicle.add_attribute_listener('location', self.location_callback)

# The callback function signature
def location_callback(self, vehicle, name, location):
    if location.global_relative_frame.alt is not None:
        self.altitude = location.global_relative_frame.alt

    self.current_location = location.global_relative_frame

GUIDED mode is the recommended mode for controlling a Copter autonomously without a predefined mission. It allows a Ground Control Station (GCS) or a Companion Computer to control the vehicle "on the fly," enabling real-time reactions to new events or situations.

Note: This functionality is Copter-specific. While Plane supports GUIDED mode, it is less feature-complete than on Copter, where AUTO mode is typically preferred for autonomous flight.

In DroneKit, vehicle information is split into two categories:

1. State Information: Real-time data such as position, speed, and orientation are exposed as Python attributes on the Vehicle class.
2. Parameters (Settings): Configuration settings and flight modes are accessed through the Vehicle.parameters object using named elements.

You can interact with these values by reading them directly as attributes or by accessing specific keys within the parameters collection.

A companion computer is an onboard device that communicates with the autopilot over a low-latency link to perform computationally intensive or time-sensitive tasks.

DroneKit is compatible with onboard computers running Linux variants that support:

1. Python
2. Installation of Python packages from the Internet.

Commonly supported platforms include Raspberry Pi, Intel Edison, BeagleBoneBlack, and Odroid.