Integrating Machine Learning Compute Mechanisms for Unmanned Aerial Vehicle Control

Abstract

We present a novel integration of Pixhawk autopilot technology with a Jetson Nano for real-time mission control using Computer Vision (CV). The Pixhawk (PX4) autopilot system provides a robust platform for autonomous missions in unmanned aerial vehicles (UAVs), offering precise control and navigation capabilities. The system gains precise and efficient onboard processing capabilities by incorporating the Jetson Nano, a powerful AI computing device, alongside the PX4. Leveraging CV algorithms, the integrated system can autonomously analyze visual data from multiple cameras in real time, allowing for dynamic changes during flight missions. This enables the UAV to respond quickly to obstacles and changing environmental conditions, changing the mission as necessary. We aim to highlight the synergy between the Jetson and the PX4, demonstrating their combined potential to enhance UAV autonomy through intelligent CV-based mechanisms.