Estimación de posición basada en Deep Learning con sensores ambientales UWB y dispositivos móviles
Fecha
2023-02-14
Autores
Título de la revista
ISSN de la revista
Título del volumen
Editor
Jaén: Universidad de Jaén
Resumen
El propósito de este trabajo es la creación de un sistema de localización de usuarios en espacios interiores mediante la
integración de balizas ambientales con tecnología UWB y una herramienta de recolección de señales UWB y envío de
datos a un nodo Fog que recibirá y procesará dicha información. Usando la técnica de fingerprinting un modelo de
Deep Learning estima la posición del usuario en base a la señal UWB y un etiquetado realizado previamente por el
usuario mediante la herramienta desarrollada para dispositivos móviles. Este sistema ofrece la suficiente precisión y
robustez como para localizar al usuario dentro de entornos interiores y se orienta al uso en aplicaciones de
seguimiento y control de personas mayores o dependientes para prevenir accidentes y monitorizar su actividad diaria
mediante un sistema de bajo coste, larga autonomía, fácil despliegue y alta escalabilidad.
The purpose of this work is the creation of an indoor user location system by integrating environmental beacons with UWB technology and a tool for collecting UWB signals and sending data to a Fog node that will receive and process this information. Using the fingerprinting technique, a Deep Learning model will estimate the user's position based on the UWB signal and a tagging previously performed by the user through the tool developed for mobile devices. This system offers sufficient accuracy and robustness to locate the user at all times in indoor environments and is intended for use in applications for tracking and control of elderly or dependent people to prevent accidents and monitor their daily activity through a low-cost system, long autonomy, easy deployment and high scalability.
The purpose of this work is the creation of an indoor user location system by integrating environmental beacons with UWB technology and a tool for collecting UWB signals and sending data to a Fog node that will receive and process this information. Using the fingerprinting technique, a Deep Learning model will estimate the user's position based on the UWB signal and a tagging previously performed by the user through the tool developed for mobile devices. This system offers sufficient accuracy and robustness to locate the user at all times in indoor environments and is intended for use in applications for tracking and control of elderly or dependent people to prevent accidents and monitor their daily activity through a low-cost system, long autonomy, easy deployment and high scalability.