Building an Autonomous Drone with Raspberry Pi & Pixhawk

Bridging Autonomy and Flight Control

Autonomous drone systems need to separate two fundamentally different types of computation:

1. Hard Real-Time Stabilization: 400Hz IMU attitude corrections, PID loops, and electronic speed controller signals. Handled by the Pixhawk Flight Controller running ArduPilot.
2. High-Level Autonomy: Computer vision, waypoint trajectory generation, and wireless mission commands. Handled by a Raspberry Pi 4B companion computer.

---

Hardware UART Communication Pipeline

Connecting the Raspberry Pi 4B to the Pixhawk over USB introduces latency and loose connector risk during flight vibrations. Instead, I connected directly via GPIO 14/15 (UART) @ 57600 baud to the Pixhawk's TELEM2 port:

[Raspberry Pi 4B (GPIO 14 TX, GPIO 15 RX)] 
                    |
              (UART Serial @ 57600 baud)
                    |
[Pixhawk TELEM2 Port (RX / TX / GND)]

---

Custom Autonomy with DroneKit-Python

Using DroneKit and PyMAVLink, the Raspberry Pi can upload waypoints dynamically and monitor real-time vehicle telemetry:

from dronekit import connect, Command, VehicleMode
from pymavlink import mavutil
import time

# Connect to Pixhawk via hardware UART
vehicle = connect('/dev/ttyAMA0', baud=57600, wait_ready=True)

print(f"Connected! Mode: {vehicle.mode.name}, GPS: {vehicle.gps_0}")

def upload_waypoint_mission(waypoints):
    cmds = vehicle.commands
    cmds.clear()
    for wp in waypoints:
        cmd = Command(
            0, 0, 0,
            mavutil.mavlink.MAV_FRAME_GLOBAL_RELATIVE_ALT,
            mavutil.mavlink.MAV_CMD_NAV_WAYPOINT,
            0, 0, 0, 0, 0, 0,
            wp['lat'], wp['lon'], wp['alt']
        )
        cmds.add(cmd)
    cmds.upload()
    print("Mission uploaded successfully.")

---

Custom 3D-Printed Vibration-Damping Mount

Motor vibrations transmitted into the Raspberry Pi frame can shake the onboard camera sensor and cause rolling shutter artifacts in computer vision frames.

I designed and 3D printed a custom PLA + TPU vibration-isolated mount that cradles the Raspberry Pi 4B, securing the UART wiring harness and eliminating jitter in aerial frames.

---

Performance & Load Testing Results

The drone was subjected to rigorous payload and waypoint precision testing:

• Waypoint Precision: Executed automated grid missions with ±2 meters GPS waypoint accuracy.
• Payload Capacity: Successfully carried 500g additional payload with zero degradation in attitude stability.
• Formed the aerial testing platform for the Autonomous Litter Detection and Recovery System research project.