Control de un sistema aéreo no tripulado mediante acelerómetro y aprendizaje automático

Abstract

The use of the technology dron is a top topic, it is a sector that is already moving more than 10.000 million of euros in the whole world and it is expected that its use will be tripled for 2020. Infinity of different applications are appearing every day both in the civil and in the military ambience. On the other hand, with the advance of the computers, a more ancient technology is gaining strength, and takes years fighting for its own place in the industrial, commercial sectors and investigation sector most important of the world. The \textit{Machine Learning} or automatic learning, inside the branch of the technologies of artificial intelligence. This project is framed inside the continuation of another final grade project, which intention was the movement's detection by means of thresholds for the control of a dron thanks to a sport watch with an acelerómeter. The intentionally of this project of is combine the drone technology with the use of Machine Learning algorithims to replace the ancient method of movement detection and to provide in this way a more efficient and flexible method. For it, the study and the validation of the algorithims of automatic learning, the management optimization of the information provided by the clock, and the performance of experimental tests have been fundamental. To achieve it, the programming language Python has been used to develop a code capable of providing to the user the collection of samples, the control of the dron or the performance of tests of the algorithm with a mode without flight.