Autonomous Landing Perception System for UAVs in Maritime Contexts - A Modular Sensor Fusion and AI Vision System for Real-Time Estimation of Landing Platform Position and Orientation on Maritime Vessels

Abstract

Precision landing on a dynamically moving platform presents significant challenges for autonomous unmanned aerial vehicles (UAVs), particularly when the target moves in all six degrees of freedom. This thesis proposes a sensor fusion approach to autonomous landing in maritime contexts by integrating data from multiple onboard sensors including light detection and ranging (LiDAR), cameras, inertial measurement units (IMUs), and depth sensors. The system combines AI-driven algorithms with an extended Kalman filter (EKF) to continuously estimate the position and orientation of the landing platform in real time. Analysis of results obtained in a simulated environment show the potential of the proposed method under challenging environmental conditions. These findings indicate the ability of sensor fusion to assist in autonomous landing in a maritime context.