Use this link to cite:
https://hdl.handle.net/2183/46378 Optimización de recursos en sistemas multiusuario para las nuevas generaciones de datos móviles
Loading...
Identifiers
Publication date
Authors
Martínez Rey, Aitana
Other responsabilities
Universidade da Coruña. Facultade de Informática
Journal Title
Bibliographic citation
Type of academic work
Academic degree
Abstract
[Resumen]: Con el creciente número de usuarios de redes móviles y la demanda de datos en constante aumento, las redes tradicionales basadas en el uso de estaciones base con pocas antenas y una asignación estática de recursos ya no pueden ofrecer la capacidad y eficiencia requeridas. En respuesta a estos desafíos, la tecnología massive Multiple-Input Multiple-Output (MIMO) se está consolidando como una solución clave, a la vez que se necesitan estrategias más avanzadas para distribuir los recursos disponibles en el sistema. Massive MIMO emplea un elevado número de antenas en cada estación base, lo que permite aumentar de manera significativa tanto la capacidad de la red como la calidad del servicio. Un aspecto fundamental de esta tecnología es la asignación de portadoras piloto, que son señales que se utilizan para medir y estimar las condiciones del enlace entre una Base Station (BS) y diversos User Equipments (UEs). La asignación eficiente de estas señales piloto resulta crucial para asegurar una comunicación precisa y de alta calidad, especialmente en entornos con múltiples usuarios y condiciones cambiantes. En este tipo de escenarios, los bloques de tiempo disponibles para estimar la respuesta del enlace son muy pequeños, lo que implica que el número de portadoras piloto va a ser mucho menor que el número de usuarios conectados a la BS. Por lo tanto, la misma portadora debe ser necesariamente reusada para varios usuarios, causando una potencial interferencia entre ellos si esta asignación no se hace de forma correcta y provocando que la estimación de los enlaces sea deficiente. En estos escenarios, la asignación inadecuada entre portadoras y usuarios (UEs) puede incrementar el Normalized Mean Squared Error (NMSE) asociado al proceso de estimación, lo que degrada el rendimiento general del sistema. Este proyecto se centra en desarrollar una solución basada en Redes Neuronales Artificiales (RNAs) para optimizar esta asignación de portadoras y otros recursos en sistemas massive MIMO. Utilizando técnicas de Machine Learning (ML), el modelo de RNA será entrenado con datos representativos para ajustar dinámicamente la asignación de recursos, mejorando así la eficiencia y minimizando el NMSE. La solución desarrollada se comparará con métodos más tradicionales para evaluar su eficacia en términos de calidad de servicio y complejidad computacional, con el objetivo de establecer una base sólida para las redes móviles del futuro.
[Abstract]: With the growing number of mobile network users and the increasing demand for data, traditional networks based on base stations with few antennas and static resource allocation can no longer provide the required capacity and efficiency. In response to these challenges, massive Multiple-Input Multiple-Output (MIMO) technology is emerging as a key solution, while more advanced strategies are needed to distribute the available resources in the system. Massive MIMO employs a large number of antennas at each base station, significantly increasing both the network capacity and service quality. A key aspect of this technology is the allocation of pilot carriers, which are signals used to measure and estimate the link conditions between a BS and various user equipment UEs. Efficient allocation of these pilot signals is crucial to ensure accurate and high-quality communication, especially in environments with multiple users and changing conditions. In such scenarios, the available time slots for estimating the link response are very small, which means the number of pilot carriers will be much smaller than the number of users connected to the Base Station (BS). As a result, the same carrier must be reused by multiple users, potentially causing interference if this allocation is not done correctly, leading to poor link estimation. In these situations, inadequate allocation between carriers and User Equipments (UEs) can increase the Normalized Mean Squared Error (NMSE) associated with the estimation process, degrading the overall system performance. This project focuses on developing a solution based on Artificial Neural Networks (ANNs) to optimize the allocation of pilot carriers and other resources in massive MIMO systems. Using Machine Learning (ML) techniques, the ANN model will be trained with representative data to dynamically adjust resource allocation, thereby improving efficiency and minimizing NMSE. The developed solution will be compared with more traditional methods to evaluate its effectiveness in terms of quality of service and computational complexity, aiming to establish a solid foundation for future mobile networks.
[Abstract]: With the growing number of mobile network users and the increasing demand for data, traditional networks based on base stations with few antennas and static resource allocation can no longer provide the required capacity and efficiency. In response to these challenges, massive Multiple-Input Multiple-Output (MIMO) technology is emerging as a key solution, while more advanced strategies are needed to distribute the available resources in the system. Massive MIMO employs a large number of antennas at each base station, significantly increasing both the network capacity and service quality. A key aspect of this technology is the allocation of pilot carriers, which are signals used to measure and estimate the link conditions between a BS and various user equipment UEs. Efficient allocation of these pilot signals is crucial to ensure accurate and high-quality communication, especially in environments with multiple users and changing conditions. In such scenarios, the available time slots for estimating the link response are very small, which means the number of pilot carriers will be much smaller than the number of users connected to the Base Station (BS). As a result, the same carrier must be reused by multiple users, potentially causing interference if this allocation is not done correctly, leading to poor link estimation. In these situations, inadequate allocation between carriers and User Equipments (UEs) can increase the Normalized Mean Squared Error (NMSE) associated with the estimation process, degrading the overall system performance. This project focuses on developing a solution based on Artificial Neural Networks (ANNs) to optimize the allocation of pilot carriers and other resources in massive MIMO systems. Using Machine Learning (ML) techniques, the ANN model will be trained with representative data to dynamically adjust resource allocation, thereby improving efficiency and minimizing NMSE. The developed solution will be compared with more traditional methods to evaluate its effectiveness in terms of quality of service and computational complexity, aiming to establish a solid foundation for future mobile networks.
Description
Keywords
RNA Aprendizaje automático Redes de Datos Móviles Massive MIMO Asignación de Portadoras Piloto Sistemas Multiusuario Estación Base Redes Celulares Interferencia Eficiencia Espectral Posición Optimización ECMN ANN Machine Learning Mobile Networks Pilot Carrier Assignment Multiuser Systems Base Station Cellular Networks Interference Spectral Efficiency Position Optimization NMSE
Editor version
Rights
Attribution-NonCommercial-NoDerivatives 4.0 International








