Final Year Project

Project Title

AURORA - Autonomous UAV Rescue Operations for Rapid Assistance

Description

This project develops an autonomous multi-drone system for search and rescue in dense forests, using ROS 2 and Gazebo for simulation. The drones operate in GPS-denied environments, relying on LiDAR-Inertial SLAM for mapping and localization. Aims & Purpose Solve the high-cost and resource-intensive challenge of searching for missing persons in forests. Enable autonomous drone exploration using frontier-based path planning. Develop a teleoperable SLAM-enabled drone for real-time mapping. Implement and compare graph-based, scan-matching, and feature-based SLAM. Tackle the mathematically complex problem of under-canopy pathfinding, using RRT and improved RRT variants. Findings & Benefits Provides an efficient, cost-effective alternative to traditional search missions. Demonstrates reliable SLAM-based localization in dense, GPS-denied environments. Optimizes path planning algorithms for effective under-canopy navigation. Contributes to real-world drone deployment, improving search success and survival rates. This project addresses a critical global challenge, making forest search missions more efficient, scalable, and accessible.