Landing a UAV in a dynamical target using Fuzzy Control and Computer Vision

João Pedro Carvalho de Souza¹; Marco Aurélio Jucá¹; Alexandre Menezes Teixeira¹; André Luis Marques Marcato¹; Alexandre Bessa dos Santos¹; Leonardo Rocha Olivi¹

¹ Universidade Federal de Juiz de Fora - Faculdade de Engenharia

Baixar PDF

Resumo

Unmanned Aerial Vehicles (UAVs) are a recent brand of technological research field where multiple applications can be designed, such as surveillance, deliveries and others. The localization of the UAVs are done with Global Positioning System (GPS), which does not offer sufficient precision for landing, especially at dynamical landing points. To solve this imperative issue, many applications with computational vision and control theory have been developed. This paper presents a novel controller based on Mamdani-type fuzzy logic and augmented reality tag identification for landing of a UAV in a dynamical target point. Results are presented with simulation and a real drone, the IRIS+ with Odroid and ROS.

Palavras-chave: Unmanned Aerial Vehicles; Fuzzy Control; Computer Vision; Augmented Reality; ROS; Odroid