Connect a Raspberry Pi to ArduPilot With MAVLink

This guide contains Amazon affiliate links. If you buy through them, DroneCoreGuide earns a small commission at no extra cost to you.

MAVLink is the language ArduPilot and PX4 speak. Once your Raspberry Pi can talk MAVLink to your flight controller, it can read telemetry, change modes and send movement commands: the foundation of every autonomous project.

Do this with props removed until the final step.

What you need

• A flight controller running ArduPilot (this guide) or PX4
• A Raspberry Pi 4 or 5 with Raspberry Pi OS
• Three jumper wires
• A separate 5 V regulator to power the Pi (shop regulators)

Step 1: Wire the serial connection

Connect a TELEM port on the flight controller to the Pi’s UART pins:

Don’t connect the FC’s 5 V to the Pi. Power the Pi from its own regulator and share only ground.

Step 2: Enable the Pi’s serial port

1. Run sudo raspi-config.
2. Under Interface Options, Serial Port: disable the login shell over serial, enable the serial hardware.
3. Reboot. The port is usually /dev/serial0.

Step 3: Configure ArduPilot

In Mission Planner or QGroundControl, set parameters for TELEM2 (SERIAL2 in ArduPilot):

• SERIAL2_PROTOCOL = 2 (MAVLink 2)
• SERIAL2_BAUD = 921 (921600 baud)

Reboot the flight controller.

Step 4: Install a MAVLink library

python3 -m venv ~/drone && source ~/drone/bin/activate
pip install pymavlink

Step 5: Read telemetry

Save this as telemetry.py:

from pymavlink import mavutil

link = mavutil.mavlink_connection('/dev/serial0', baud=921600)
link.wait_heartbeat()
print("Connected to system", link.target_system)

while True:
    msg = link.recv_match(type=['ATTITUDE', 'GLOBAL_POSITION_INT', 'SYS_STATUS'], blocking=True)
    print(msg.get_type(), msg.to_dict())

Run it. You should see attitude, position and battery messages streaming. If nothing arrives, double-check TX/RX crossing, baud rate and the serial port settings.

Step 6: Change modes and send commands

With the connection working, you can switch modes, arm and send position or velocity targets in GUIDED mode. A simple example to change mode:

link.set_mode(link.mode_mapping()['GUIDED'])

For full missions and velocity control, many builders prefer MAVSDK (a higher-level API) or ROS 2 with MAVROS for bigger robotics projects.

Step 7: Test in simulation before real flight

Run the same script against ArduPilot SITL by changing the connection string to the simulator’s UDP address. Confirm every command does what you expect. See SITL and Gazebo.

Understanding MAVLink messages

A few messages you’ll use constantly:

Requesting data at the right rate

Autopilots send some messages at default rates. You can request specific messages at higher rates using a SET_MESSAGE_INTERVAL command or by setting the stream rate parameters on the serial port. Don’t request everything at maximum rate: it wastes bandwidth and CPU on both sides.

Running your script automatically

Use a systemd service so your script starts at boot:

1. Create a service file that runs your script inside its virtual environment.
2. Enable it with systemctl enable.
3. Check logs with journalctl when debugging.

Debugging connection problems

• No heartbeat: check TX/RX crossing, ground, baud rate, SERIAL2_PROTOCOL and that the Pi’s serial login shell is disabled.
• Garbled data: baud rate mismatch.
• Works on the bench, fails in flight: check power to the Pi; voltage dips reboot it.
• Commands ignored: the drone must be in GUIDED mode, and often armed, for movement commands to take effect.

Safety rules for companion-computer control

• Keep a mode switch on your radio that overrides GUIDED mode, such as Loiter or Stabilize.
• Set failsafes in ArduPilot for radio loss and companion-computer loss.
• Limit speed and enable a geofence.
• First real tests: low altitude, open field, props on only when you’re ready.

Next: add object tracking.