Dissertaçõespara 2017/2018 1º Semestre

Dissertações para 2017/2018 1º Semestre

Grupo de telecomunicaçõesDept. Engenharia ElectrotécnicaFaculdade de Ciências e TecnologiaUniversidade Nova de Lisboa

junho 22, 2017FCT-UNL, Portugal

2

Resumo1. Introdução

2. Apresentaçãodas dissertações

3. Perguntas

ResumoØ Introdução às áreas de investigação no grupo

Ø Apresentação das propostas de dissertação

Ø Perguntas

3

Áreas de investigaçãoØ Redes de TelecomunicaçõesDesenho e arquitectura de redes; Protocolos de comunicação; Software de rede e aplicações; Operação e gestão de redes; etc.

Ø ComunicaçãoComunicações sem fios; Comunicações com espalhamentode espectro; Sincronização, equalização, detecção e estimação de canal; Modulação e desenho de sinal; Teoriada codificação e aplicações; etc.

Ø Antenas e Propagação



3. Perguntas

4

Áreas de investigaçãoØ Redes de TelecomunicaçõesPaulo Pinto, Luis Bernardo, Rodolfo Oliveira, Pedro Amaral

ComunicaçãoRui Dinis, Paulo Montezuma, Marko Beko

Ø Propagação e RadiaçãoIsabel Ventim Neves



3. Perguntas

DISSERTAÇÕES PROPOSTAS

5



3. Perguntas

6

#1. Dynamic scheduling of slices in 5G Networks



3. Perguntas

Requirements: approval in Configuração e Gestão de Redes

Apply the concept of slices to provide network services complying with a SLA and are suitable for different use cases, subscriber types, and applications.

The purpose is to offer different conditions of latency, throughput and availability.

The scope of the dynamic nature will be defined in the thesis. It can mean on-the-fly deployment, bulk transfer of functions and traffic, etc. Scalability issues will be addressed. Supervisor: Paulo Pinto

#2 Near Optimal TE and NFV placement using SDN tools



3. Perguntas

SDN open way to approaches to both problems based on LinearOptimization:• Complex, require a high number of rules in the switches (critical

resource) and the control loop can lead to transient congestion

Classical Networks use distributed routing and forwarding:• Simple and scalable but using local information and ECMP forwarding

leads to poor utilization

Network Function Virtualization, virtualizes in software several traditionalHardware funtionality (load balancing, NAT, Firewalls etc..). NFVs mustbe placed in the Network and traffic must be steered to them.• NFV placement problem

#2 Near Optimal TE and NFV placement using SDN tools



3. Perguntas

Approach:• Build on top of ongoing work (SDN architecture with distributed

control)• Compare the scalability and performance of the alternative schemes to

linear optimization, in terms of performance and scalability.• Schemes based in Social Network theory (Eigen Vector Centralities

etc..) adapting the topological measures to capacity/demand basedmesures.

• Schemes based in ML algorithms.• Apply the same principles to the NFV placement problem.

Supervisor: Pedro Amaral

#3 Machine Learning for network control



3. Perguntas

In modern networks (with SDN solutions and NFVs running in software)richer network information is obtainable (flow durations ; packet counts;byte counts etc...) and control is via software.

Approach:• Build on top of ongoing work, SDN based architecture to collect traffic

statistics from network vantage points.• Explore data mining algorithms for user profiling and network

profiling.• Use learned information for traffic Managment and learning based

network optimization.

What can we learn from collected data using data analytics:• Pattern recognition, User profiling, Etc..

Reinforcement learning for traffic Managment


#4 Hybrid SDN network implementation



3. Perguntas

Interconnecting SDN networks with legacy networks poses challenges.

Dealing with Broadcast L2 Traffic. Interworking with distributed controlplane protocols, Address translation, interactions with the hosts etc..

Solving these problems with current SDN architectures might not scale.

Objectives:• Implement a 2.0 version of the currently deployed SDN network at the

DEE.• Phase out the current legacy network.• Study the interconnection problems with the rest of the non-SDN

network of the campus• Interconnect with the OpenStack cloud at the Datacenter.


11



3. Perguntas

#5 Protocol’s Behavior Analysis

q Main goal:

- identify and assess efficient techiques to attest theprotocols’ behavior (data science techniques +support of BigData technology);

- practical setup with large amount of data;

- protocols (GSM signaling, GTP, SIP, SDP, RTP)

12



3. Perguntas


q Opten-BTS (LTE, UMTS, GSM)

q Asterisk-SIP Server

q 4G IMS core

q Supervisor: Rodolfo Oliveira

GSM

Asterisk-SIP Server

IMS 4G Core

13



3. Perguntas

q Tentative work plan:0 - State-of-the-art (initial readings);

1 – Testbed Cognizance Stage

2 - Offline Data Acquisition;

3 - Online Data Analysis;

4 - Performance Evaluation;

5 - Design of Online Analysis;

6 - Reporting and thesis writing;


14



3. Perguntas

#6 IEEE 802.15.4 MAC performance

q Main goal:- contribute with a theoretical model for theperformance of IEEE 802.15.4 and similar protocols

- focus: beacon mode

- validation of the model through simulation andreal data (using SunSpots)


15



3. Perguntas

#6 IEEE 802.15.4 MAC performance

q Tentative work plan:0 - state-of-the-art (initial readings);

1 - definition of the simulations;

2 - theoretical model development;

3 - operational scenarios (definition);

4 – simulation, validation and evaluation(comparison);

5 - reporting and thesis writing;

16



3. Perguntas

#7 How much energy can be harvested?

q Main goal:- quantify the amount of energy transmitted bydifferent devices (cellular, wifi, bluetooth, DECT,15.4) in an experimental way;

minorgoals:- RF exposure and SAR


17



3. Perguntas

#7 How much energy can be harvested?q Tentative work plan:

0 - state-of-the-art (initial readings);

1 - identification of accurate methodologies tomeasure RF energy (SAR standards);

2 - theoretical model development;

3 - operational scenarios (definition);4 – simulation, validation and evaluation(comparison);

5 - reporting and thesis writing;

18



3. Perguntas

#8 Cognitive Radio

q Software Defined Radio + Sensing + MAC


19



3. Perguntas

#8 Cognitive Radio

q Software Defined Radio + Sensing + MAC

20



3. Perguntas

#9 MAC for Positioning Systems

q Main goal:

- Positioning systems are focused on obtainingaccurate location results

- How MAC can take advantage of the nodes’position while integrating the operationalfeatures of the Positioning System?

q Supervisor: Rodolfo Oliveira + Marko Beko

21



3. Perguntas

#10 Distributed Access in 5G Systems

q Main goal:

- MPR + MAC for distributed access;

- Random Access


22

Research area: Transmission techniques and efficient amplificationtarget and applications?Resumo

1. Introdução


3. Perguntas

• Next generation of 5G systems;• Power efficient transmission structures for 4G and 5G;• Power efficient wireless transmission systems with physical

layer security -> higher battery life on mobile devices;• Physical layer security -> increased security against

interceptions;• Power amplification of single carrier and multi carrier systems.

23

#11 Impairments in quantized linear amplification with nonlinear amplifiers



3. Perguntas

A linear amplification method based on nonlinear amplifiers for single carrier and multi-carrier communication systems:1. Achieves a linear amplification with nonlinear amplifiers. 2. Performs a linear amplification of the quantized values of the

samples of a single carrier (SC) signal or the time domain samples resulting from an inverse fast Fourier transform (IFFT) of a multi-carrier signal (MC).

3. Suitable for further communication systems and the existing ones with envelope fluctuations.

Objectives: Analyse the impairments that may affect the performance of such structure. This includes the effect of the quantizer, the mismatches between amplification branches and spectral impact.Supervisors: Paulo Montezuma & Rui Dinis

24

#12 On-off structures for linear amplification with multiple nonlinear amplifiers



3. Perguntas

A linear amplification method based on nonlinear amplifiers for single carrier and multi-carrier communication systems:1. Achieves a linear amplification with multiple nonlinear amplifiers. 2. Performs a linear amplification of the quantized values of the

samples of a single carrier (SC) signal or the time domain samples resulting from an inverse fast Fourier transform (IFFT) of a multi-carrier signal (MC).

3. Suitable for further communication systems and the existing ones with envelope fluctuations.

Objectives: design and optimization of the combiners and pulse shape design for optimization of the spectral masks in these on-off amplification structures.Supervisors: Paulo Montezuma & Rui Dinis

25

#13 Implementation of OFDM-Based Communications in Software Defined Radio



3. Perguntas

The goal of this thesis is to use SDR (Software-Defined Radio) tools to implement and demonstrate wireless communications based on OFDM schemes.This work will have the following steps:1) Implementation in MatLab of the OFDM-based transmission scheme2) Simulink implementation3) SDR implementation4) Analysis of the transmission characteristics (spectral densities, error

rates, etc.)5) Transmission of specific data (files, pictures, video)Supervisors: Rui Dinis, João Guerreiro, Paulo Montezuma

- Massive MIMO Systems- Huge number of antennas- High capacity gains- Channel estimation difficulties

- Several thesis- Capacity gains under realisticchannel conditions

- Hybrid analy/digital massiveMIMO schemes

- Impact of hardwareimpairments

- Supervisors:- Rui Dinis, Paulo Montezuma,Marko Beko 26

#14 Massive MIMO techniques for 5G Systems



3. Perguntas

27

#15 Protocol for Extreme Low Latency M2M Communication Networks



3. Perguntas

Requirements: approval in Redes Móveis

Protocol for Ultra dense Reliable Low Latency (URLL) communication in 5G machine-type communication (MTC). Design a MAC protocol and dynamic spectrum management algorithm, that adapts access to the expected load / terminal densityPerformance evaluation using simulation Optimize the protocol’s performance

Supervisor: Luis Bernardo

28

#16 C-RAN CoMP Methods for MPR Receivers



3. Perguntas

Requirements: approval in Redes Móveis;recommended Comunicação Digital

Fronthaul/Backhaul configuration in a C-RAN for 5G Ultra Reliable Low Latency (URLL) communicationConfigure C-RAN fronthaul SDN flows and processing units to support CoMP and provide URLL using a MPR receiverUse data analytics and learning techniques to implement C-RAN CoMP self-configuration in real-timePerformance evaluation using simulation

Supervisors: Luis Bernardo & Rui Dinis

29

# Cooperation with NokiaResumo1. Introdução


3. Perguntas

Additional thesis will be proposed in co-supervision with Nokia

Details will be provided later!

Subjects will possibly cover:- Automatic network configuration tool- Cognitive Radio- SDN - C-RAN optimization

30



3. Perguntas

Documents

Dissertaçõespara 2017/2018 1º Semestre