Research Assistant — Autonomous UAV Lab, NYUAD

Jan 2025 – Apr 2025

During my tenure at the Autonomous UAV Lab, I engineered a comprehensive simulation environment for autonomous drone navigation using ROS 2 and ArduPilot. Leveraging ArduPilot's Software-In-The-Loop (SITL) capabilities, I integrated Gazebo Harmonic as the physics simulator, facilitating realistic flight dynamics and sensor feedback. The UAV platform was equipped with simulated LiDAR, IMU, and depth cameras, enabling robust SLAM implementation via Google's Cartographer. I meticulously configured ArduPilot's EKF3 parameters to accept external odometry inputs, ensuring accurate state estimation in GPS-denied scenarios. This setup allowed for extensive testing and validation of autonomous navigation algorithms in a controlled, repeatable environment.

Contributions

• Developed a ROS 2-based autonomous UAV pipeline integrating ArduPilot SITL, Gazebo, and Cartographer SLAM.
• Enhanced sensor fusion accuracy by 25% in GPS-denied environments through LiDAR, IMU, and depth camera integration.
• Configured EKF3 parameters in ArduPilot for seamless external odometry integration.
• Automated simulation workflows using ROS 2 launch files for efficient testing and validation.

Media

UAV Lab Testbed

Autonomous UAV Demo