Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/148502
|Navegación autónoma de vehículos aéreos no tripulados mediante técnicas de aprendizaje por refuerzo profundo
|Núñez Sánchez-Agustino, Francisco José
|This work is a study on the practical application of Deep Reinforcement Learning (DRL) techniques in the development of an autonomous navigation system for Unmanned Aerial Vehicles (UAVs), also known as ”drones”. These types of aircraft are used for a wide variety of tasks such as surveillance, search and rescue, topography, or scientific research, among others. In all of these tasks, autonomous navigation is a desirable feature due to the safety and efficiency it can provide to the operation of these devices. However, eliminating human intervention presents significant challenges, such as detecting and avoiding obstacles to ensure the integrity of the aircraft and the safety of people, animals, plans or objects in the environment. In this sense, deep reinforcement learning has become a very promising alternative to carry out this complex task, allowing the drone to learn through its experience and interactions with different environments, both simulated and real. This document aims to demonstrate the viability of a navigation system that combines deep reinforcement learning algorithms with computer vision techniques, to enable a low-cost, simple drone equipped only with a front-facing video camera to navigate autonomously for as long as possible.
|deep reinforcement learning
|Appears in Collections:
|Bachelor thesis, research projects, etc.
Files in This Item:
|Memoria del TFM