This course is designed for students who have some fundamental knowledge and experience in building and flying of unmanned aerial vehicles (UAVs) and are keen on exploring multi-UAV capabilities. In the course, students will learn how to implement sensors and algorithms for autonomous navigation in unknown environments, integrate artificial intelligence for perception, and coordinate multi-UAV missions. The course culminates in a team-based project where students will use multiple UAVs to execute a collaborative mission.

At the end of this course, students should be able to:

• Implement autonomous navigation in unknown environments using SLAM and obstacle avoidance.
• Integrate advanced sensors and companion computers for real-time processing.
• Apply AI techniques for perception.
• Design and execute multi-UAV missions, including formation flying and collaborative tasks.

Workload: 4 units (letter-graded)

Note: Students who wish to enrol in this course should have some basic knowledge in UAV design, assembly and programming from CDE1302 Introduction to UAVs: From Basics to Autonomous Missions or other courses.

Course syllabus

Review of UAV fundamentals and introduction to advanced UAV architecture:

• Review of UAV fundamentals
• Companion computers
• High-level system integration for autonomy

Sensors for advanced navigation:

• LiDAR, depth cameras, and stereo vision
• Sensor fusion for localisation

SLAM fundamentals:

• Simultaneous Localisation and Mapping (SLAM) concepts
• Implementing SLAM with ROS2

Autonomous exploration and obstacle avoidance:

• Real-time path planning algorithms
• Dynamic waypoint generation
• Collision avoidance strategies

AI for UAV perception:

• Object detection and classification using deep learning
• Lightweight models for onboard inference

Simulation and testing:

• Gazebo and PX4 SITL for advanced mission simulation
• Testing SLAM simulation virtually

Multi-UAV communication:

• MAVLink for swarm coordination
• ROS2 multi-agent frameworks

Swarm algorithms:

• Leader-follower and consensus-based control
• Formation flying