A CNN Pilot for Autonomous Drone Racing

Convolutional neural networks (CNN) and deep learning (DL) have become a popular tool for addressing all kinds of artificial intelligence challenges. The Autonomous Drone Race is a challenge that consists of developing a drone capable of defeating a human in a drone race. DL is a tool that has been...

Descripción completa

Detalles Bibliográficos
Autor: Leticia Oyuki Rojas-Perez
Tipo de recurso: tesis de maestría
Estado:Versión aceptada para publicación
Fecha de publicación:2020
País:México
Institución:Instituto Nacional de Astrofísica, Óptica y Electrónica
Repositorio:Repositorio Institucional del INAOE
Idioma:inglés
OAI Identifier:oai:inaoe.repositorioinstitucional.mx:1009/2162
Acceso en línea:http://inaoe.repositorioinstitucional.mx/jspui/handle/1009/2162
Access Level:acceso abierto
Palabra clave:info:eu-repo/classification/Inspec/Autonomous drone racing
info:eu-repo/classification/Inspec/CNN y Deep learning
info:eu-repo/classification/cti/1
info:eu-repo/classification/cti/12
info:eu-repo/classification/cti/1203
info:eu-repo/classification/cti/330412
Descripción
Sumario:Convolutional neural networks (CNN) and deep learning (DL) have become a popular tool for addressing all kinds of artificial intelligence challenges. The Autonomous Drone Race is a challenge that consists of developing a drone capable of defeating a human in a drone race. DL is a tool that has been included in state-of-the-art solutions to address this problem. Current work has proposed using CNN and DL to detect the gates, while other work has proposed using a CNN to obtain the flight speed and a three-dimensional reference point, these data are used by the controller to generate the corresponding control signals. It should be noted that all these approaches use a single frame as input. Motivated by the above, this work aims to develop a CNN to obtain the control signals directly for a drone to navigate autonomously in a drone racing circuit. This implies two levels of difficulty: 1) navigating through a predefined sequence of gates; 2) navigating in an environment where the location of the gates is not known a priory. The performance tests were carried out in the Gazebo simulator, using the AR drone vehicle.