MmWave communication systems are exposed to fundamental problems when high interference is present, due to the high number of devices, strong attenuation and transmission frequencies above 28GHz. The main questions are how this interference can be described statistically using the correlation matrix, how this interference correlation can be estimated and how the estimated interference correlation can be used to improve the capacity of a mmWave transmission. Furthermore, changes in correlation matrices should be detected, searching for methods which can be used in mmWave environments. In this work a simple mmWave environment was reproduced. A focus was set to the angles of arrival (AoA), which is a parameter that defines the interference. Traditional and novel methods for AoAs estimation, like MUSIC and methods based on a new gridless multidimensional AoAs estimation method, are used in the mmWave environment to estimate the interference correlation. They are compared with the least square estimation method in terms of mean square error (MSE) and in terms of capacity in a mmWave uplink transmission.The characteristics of the mmWave interference was analyzed and used for the search of a possible algorithm to detect changes in the interference correlation. This thesis revealed the importance of knowing the AoAs for the interference correlation estimation and the differences in terms of performance between the various estimation methods, showing an estimation improvement in some novel methods taking into account the MSE. Considering the capacity of a mmWave transmission, all methods, which use AoAs for the estimation of the interference correlation, accelerate the reaching of the maximal capacity significantly. Also a slight improve in terms of capacity of the novel gridless AoA estimation methods, with respect to the MUSIC algorithm, was detected. Moreover a method for change detection, called ODIT, was found and tested. It was asserted that it satisfies the main properties to be used as mmWave interference correlation change detection algorithm.
I sistemi di comunicazione che fanno uso di onde millimetriche sono soggetti a problematiche di interferenza, causate dalla presenza di un alto numero di dispositivi, di una forte attenuazione e di frequenze di trasmissione superiori ai 28GHz. Il presente elaborato si pone l’obiettivo di indagare come queste interferenze possano essere descritte in termini statistici usando la matrice di correlazione, in che modo la correlazione dell’interferenza possa essere stimata e quanto tale stima possa migliorare la capacità nelle comunicazioni con onde millimetriche, mitigandone l’interferenza. Inoltre si è cercato di individuare metodi in grado di rilevare cambiamenti della correlazione d’interferenza anche in contesti che utilizzano onde millimetriche. A questo scopo è stato riprodotto un contesto semplificato in cui vengono simulate comunicazioni con onde millimetriche.Una particolare attenzione è stata data agli angoli di arrivo, parametri che definiscono l’interferenza. Per stimare la matrice di correlazione dell’interferenza nel contesto delle onde millimetriche sono stati utilizzati metodi tradizionali e innovativi per la stima degli angoli di arrivo, come per esempio il “MUSIC” o il “gridless multidimensional angle of arrival estimation method”. Questi metodi sono stati messi a confronto con il metodo least square per la stima della matrice di correlazione in termini di errore quadratico medio (MSE) e in termini di capacità in una comunicazione uplink con onde millimetriche. Al fine di individuare un algoritmo in grado di determinare cambiamenti nella correlazione di interferenza, sono state analizzate le proprietà di contesti in cui si svolgono comunicazioni con onde millimetriche. Il lavoro svolto conferma l’importanza degli angoli di arrivo nella stima della correlazione d’interferenza ed evidenzia una differenza tra i metodi di stima, riportando un miglioramento in termini di MSE da parte di alcuni metodi innovativi. Tutti i metodi che ricorrono agli angoli di arrivo per la stima della correlazione d’interferenza si sono dimostrati in grado di raggiungere in tempi più brevi la massima capacità possibile. Inoltre, i nuovi metodi riportano un lieve miglioramento in termini di capacità rispetto al “MUSIC”. Infine è stato testato il metodo “ODIT”, in grado di rilevare cambiamenti della matrice di correlazione d’interferenza.Dalle simulazioni effettuate emerge che tale metodo soddisfa tutti i principali requisiti necessari per rilevare cambiamenti della correlazione d’interferenza in contesti con comunicazioni che fanno uso di onde millimetriche.
Statistics estimation of interference in mmWave Channels
HASLER, RAPHAEL
2021/2022
Abstract
MmWave communication systems are exposed to fundamental problems when high interference is present, due to the high number of devices, strong attenuation and transmission frequencies above 28GHz. The main questions are how this interference can be described statistically using the correlation matrix, how this interference correlation can be estimated and how the estimated interference correlation can be used to improve the capacity of a mmWave transmission. Furthermore, changes in correlation matrices should be detected, searching for methods which can be used in mmWave environments. In this work a simple mmWave environment was reproduced. A focus was set to the angles of arrival (AoA), which is a parameter that defines the interference. Traditional and novel methods for AoAs estimation, like MUSIC and methods based on a new gridless multidimensional AoAs estimation method, are used in the mmWave environment to estimate the interference correlation. They are compared with the least square estimation method in terms of mean square error (MSE) and in terms of capacity in a mmWave uplink transmission.The characteristics of the mmWave interference was analyzed and used for the search of a possible algorithm to detect changes in the interference correlation. This thesis revealed the importance of knowing the AoAs for the interference correlation estimation and the differences in terms of performance between the various estimation methods, showing an estimation improvement in some novel methods taking into account the MSE. Considering the capacity of a mmWave transmission, all methods, which use AoAs for the estimation of the interference correlation, accelerate the reaching of the maximal capacity significantly. Also a slight improve in terms of capacity of the novel gridless AoA estimation methods, with respect to the MUSIC algorithm, was detected. Moreover a method for change detection, called ODIT, was found and tested. It was asserted that it satisfies the main properties to be used as mmWave interference correlation change detection algorithm.File | Dimensione | Formato | |
---|---|---|---|
Hasler_Raphael.pdf
accesso aperto
Dimensione
1.13 MB
Formato
Adobe PDF
|
1.13 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/29054