Control based on GLRT algorithm for Unmanned Aerial Vehicle

Abstract

This paper proposes the study of a vision-based control scheme for an Unmanned Aerial Vehicle (UAV) for tracking objects floating on the sea surface. The applied vision scheme is based on the generalized likelihood ratio test (GLRT) algorithm. Once the target is detected, its coordinates are computed and used by the UAV control to track the target. The quadrotor mathematical model is developed using the Newton-Euler approach and a PID controller is implemented to track the vector containing the coordinates obtained through the vision scheme. To verify the effectiveness of the proposal, simulation tests are developed in MATLAB/Simulink based on a real video of an objective floating in the sea surface. The obtained results show an appropriate detection and tracking of the objective.