Autonomous Drones II

You have mastered the fundamentals (Volume 1). Now it is time to take it to the next level: intelligent autonomy.

This book teaches you how to integrate robotics, computer vision, and artificial intelligence into real drones. Perfect for engineers looking to specialise in autonomous perception.

Chapter 1 - ROS2 and Robotic Architecture

- The de facto operating system in professional robotics.

- Nodes, topics, messages - decentralised architecture.

- Ardupilot + ROS2 integration (MAVLink bridge).

- Nav2: 3D autonomous navigation stack.

- Simulation with Gazebo + SITL.

Chapter 2 - Computer Vision and Object Detection

- OpenCV: real-time image processing.

- YOLOv8: ultra-fast detection (45 FPS on GPU).

- Classical methods vs. Deep Learning.

- Integration with ROS2 (image publishers/subscribers).

- Real-world use cases: detection of people, vehicles, points of interest.

Chapter 3 - AI in Drones

- Edge Computing: processing on the drone, not in the cloud.

- Jetson line (Nano â' Orin): selection based on latency and budget.

- Latency < 100ms: mandatory for autonomous flight.

- TensorRT: 2-3x acceleration of NN models.

- Complete architecture: Jetson + ROS2 + Ardupilot + vision.

Key Features:

- â" 246 pages of applied content.

- â" Ready-to-use Python/C++ code.

- â" 20+ graphs and flow diagrams.

- â" Compatible with hardware: Jetson Nano, Orin NX, RTX.

- â" Preparation for research/commercial drones.

Prerequisites:

- Familiarity with Python (Appendix A2).

- Drone concepts (Volume 1).

- Ubuntu 22.04 recommended.

Who is it for?

- Engineers specialising in robotic autonomy.

- Researchers in computer vision.

- AI drone startups.

- Makers who want "intelligent" drones.

From object detection to autonomous decision-making, you will learn the complete stack of intelligent drones.

All examples, practices, exercises and exams are solved and available on GitHub:

https://github.com/DroneBooks/AutonomousDrones