Deliverable D4.1 TERRANOVA’s MAC layer definition ... · D4.1 – TERRANOVA’s MA layer...

D4.1 – TERRANOVA’s MAC layer definition & resource management formulation

TERRANOVA Project Page 1 of 85

This project has received funding from Horizon 2020, European Union’s

Framework Programme for Research and Innovation, under grant agreement

No. 761794

Deliverable D4.1 TERRANOVA’s MAC layer definition &

resource management formulation Work Package 4 - THz Wireless Access and Resource Management

TERRANOVA Project

Grant Agreement No. 761794

Call: H2020-ICT-2016-2

Topic: ICT-09-2017 - Networking research beyond 5G

Start date of the project: 1 July 2017

Duration of the project: 30 months



Disclaimer This document contains material, which is the copyright of certain TERRANOVA contractors, and

may not be reproduced or copied without permission. All TERRANOVA consortium partners have

agreed to the full publication of this document. The commercial use of any information contained

in this document may require a license from the proprietor of that information. The reproduction

of this document or of parts of it requires an agreement with the proprietor of that information.

The document must be referenced if used in a publication.

The TERRANOVA consortium consists of the following partners.

No. Name Short Name Country

1

(Coordinator)

University of Piraeus Research Center UPRC Greece

2 Fraunhofer Gesellschaft (FhG-HHI & FhG-IAF) FhG Germany

3 Intracom Telecom ICOM Greece

4 University of Oulu UOULU Finland

5 JCP-Connect JCP-C France

6 Altice Labs ALB Portugal

7 PICAdvanced PIC Portugal



Document Information

Project short name and number TERRANOVA (653355)

Work package WP4

Number D4.1

Title TERRANOVA’s MAC layer definition &

resource management formulation

Version V1.0

Responsible unit UPRC

Involved units JCP-C, FhG, ICOM, UOULU, UPRC, ALB, PIC

Type1 R

Dissemination level2 PU

Contractual date of delivery 30.06.2018

Last update 30.06.2018

1 Types. R: Document, report (excluding the periodic and final reports); DEM: Demonstrator, pilot, prototype, plan designs; DEC: Websites, patents filing, press & media actions, videos, etc.; OTHER: Software, technical diagram, etc. 2 Dissemination levels. PU: Public, fully open, e.g. web; CO: Confidential, restricted under conditions set out in Model Grant Agreement; CI: Classified, information as referred to in Commission Decision 2001/844/EC.



Document History

Version Date Status Authors, Reviewers Description

v0.1 29.03.2018 Draft Alexandros-Apostolos A.

Boulogeorgos (UPRC)

Initial version, definition of a

structure

v0.2 11.05.2018 Draft Ahmed Mokhtar (JCP) Creation and participation of

section 3.7


Boulogeorgos (UPRC)

Contribution on Sections 2 to 4.

v0.4 31.05.2018 Draft Joonas Kokkoniemi

(OULU)

Janne Lethtomaki

(OULU)

Marku Juntti (OULU)


v0.5 10.05.2018 Draft Ahmed Mokhtar (JCP) Revise of Figure 10.


(OULU)

Janne Lethtomaki

(OULU)

Marku Juntti (OULU)

Contribution on Section 4.

v0.7 12.05.2018 Draft Jean-Charles Point (JCP) Contribution on Section 4.


Boulogeorgos (UPRC)



(OULU)

Janne Lethtomaki

(OULU)

Marku Juntti (OULU)



Boulogeorgos (UPRC)






v0.13 20.06.2018 Draft Janne Lethtomaki

(OULU)

Review Sections 3.1, 3.2, 3.4,

4.2, and 4.4.


Boulogeorgos (UPRC)

Address comments on Sections

3.1, 3.2, 3.4, 4.2 and 4.4.

Review Sections 3.5, 3.7, and

4.5.

v0.15 25.06.2018 Draft Ahmed Mokhtar (JCP) Contribution on Section 3.4.

Review of Sections 3.1, 3.3, 3.6,

and 4.1.

v0.16 25.06.2018 Draft Georgios Stratidakis

(UPRC)

Editing in Sections 1 – 5


(OULU)

Editing in Sections 1 – 5


Boulogeorgos (UPRC)

Proofreading


(OULU)

Proofreading

v0.20 29.06.2018 Draft Angeliki Alexiou (UPRC) Review of all sections

V1.0 30.06.2018 Final Angeliki Alexiou (UPRC) Revision of all sections



Acronyms and Abbreviations

Acronym/Abbreviation Description

2G Second Generation

3G Third Generation

3GPP Third Generation Partnership Project

5G Fifth Generation

A-BFT Associate BeamForming Training

A-MSDU Aggregated Medium access control Service Data Unit

A-MSPU Aggregated Medium access control Service Protocol Unit

ACK Acknowledgement

ACO Analog Coherent Optics

ADC Analog-to-Digital Converter

AFC Automatic Frequency Correction

AFE Analogue FrontEnd

AGC Automatic Gain Control

AiP Antenna-in-Package

AM Amplitude Modulation

AMC Adaptive Modulation and Coding

AP Access Point

ASIC Application-Specific Integrated Circuit

ATDE Adaptive Time Domain Equalizer

ATI Announcement Transmission Interval

AWG Arrayed Waveguide Gratings

AWGN Additive White Gaussian Noise

AWV Antenna Weight Vector

BB BaseBand

BC Beam Combining

BER Bit Error Rate

BF BeamForming



BHI Beacon Header Interval

BI Beacon Interval

BOC Back-Off Counter

BPSK Binary Phase Shift Keying

BRP Beam Refinement Protocol

BS Base Station

BTI Beacon Transmission Interval

BW BandWidth

CA Consortium Agreement

CAP Contention Access Period

CAUI 100 gigabit Attachment Unit Interface

CBAP Contention-Based Access Period

CapEx Capital Expenditure

CC Central Cloud

CCH Control CHannel

CDR Clock and Data Recovery

CFP C-Form Factor Pluggable

CMOS Complementary Metal–Oxide–Semiconductor

CoMP Coordination Multi-Point

COTS Commercial Off-The-Shelf

CPR Carrier Phase Recovery

CRC Cyclic Redundancy Code

CS Compressive Sensing

CSI Channel State Information

CSMA/CA Carrier Sense Multiple Access with Collision Avoidance

CTA Channel Time Allocation

CTAP Channel Time Allocation Period

CTS Clear-To-Send

CTS-NI Clear-To-Send-Node-Information

CW Continuous Wave

D2D Device-to-Device



DAC Digital to Analog Converter

DC Direct Current

DCH Data CHannel

DDC Digital Down Conversion

DEMUX DE-MUltipleXer

DL DownLink

DMG Directional Multi-Gigabit

DMT Discrete Multi-Tone

DO Directional-Omni

DoA Direction of Arrival

DoF Degree of Freedom

DP Detection Probability

DP-IQ Dual Polarization In-phase and Quadrature

DPD Digital PreDistortion

DSP Digital Signal Processing

DTI Data Transfer Interval

DUC Digital Up Conversion

DWDM Dense Wavelength Division Multiplexing

EC European Commission

EDCA Enhanced Distributed Channel Access

EDMG Enhanced Directional Multi-Gigabit

E/O Electrical-Optical

ESE Extended Schedule Element

ETSI European Telecommunications Standards Institute

eWLB embedded Wafer Level Ball grid array

FAP False-Alarm Probability

FEC Forward Error Correction

FCS Frame Check Sequence

FD Full Duplex

FDD Frequency Division Duplexing

FDMA Frequency Division Multiple Access



FIFO First In First Out

FM Frequency Modulation

FPGA Field-Programmable Gate Array

FSO Free-Space Optics

FSPL Free Space Path Loss

FTTH Fiber To The Home

FWA Fixed Wireless Access

GA Grant Agreement

GaAs Gallium Arsenide

HEMT High Electron Mobility Transistor

HFT High Frequency Trading

HSPA High Speed Packet Access

HSPA+ evolved High Speed Packet Access

I/Q In-phase and Quadrature

I2C Inter-Integrated Circuit

IA Initial Access

ICF Intermediate Carrier Frequency

IEEE Institute of Electrical and Electronics Engineers

IF Intermediate Frequency

IoT Internet of Things

IM/DD Intensity Modulation/Direct Detection

IP Internet protocol layer

ISI InterSymbol Interference

ISM Industrial Scientific and Medical band

ITU International Telecommunication Union

ITU-R Radiocommunication sector of the International

Telecommunication Union

IQ COMP. In-phase and Quadrature impairments COMPensator

IQD Indoor Quasi Directional

KPI Key Performance Indicator

LDPC Low-Density Parity-Check



LO Local Oscillator

LoS Line of Sight

LTE-A Long Term Evolution Advanced

MAC Medium Access Control

MCE MAC Coordination Entity

MID Multiple sector IDentifier

MIMO Multiple Input Multiple Output

MMIC Monolithic Microwave Integrated Circuit

mmWave Millimeter Wave

MUE Mobile User Equipment

MUX MUltipleXer

MZI Mach-Zehnder Interferometer

NAV Network Allocation Vector

NETCONF NETwork CONFiguration

NI Node Information

NGPON2 Next-Generation Passive Optical Network 2

NLOS Non-Line Of Sight

NR New Radio

NRZ Non-Return to Zero

OFDM Orthogonal Frequency Division Modulation

OIF Optical Internetworking Forum

OLT Optical Line Terminal

ONUs Optical Network Units

OOK On-Off Keying

OpEx Operating Expenses

P2MP Point-to-Multi-Point

P2P Point-to-Point

PA Power Amplifier

PAM Pulse Amplitude Modulation

PBSS Personal Basic Service Set

PCB Printed Circuit Board



PCP Personal basic service set control point

PDM Polarization-Division Multiplexing

PDM-QAM Polarization Multiplexed Quadrature Amplitude

Modulation

PER Packet Error Rate

PFIS Point coordination Function Inter-frame Space

PHY PHYsical

PIN Positive-Intrinsic-Negative

PLL Phased Locked Loop

PNC Picocell Network Coordinator

PONs Passive Optical Networks

PSP Pulse Shaping Filter

PSS Primary Synchronization Signal

PtMP Point-to-Multi-Point

QAM Quadrature Amplitude Modulation

QoE Quality of Experience

QoS Quality-of-Service

QSFP Quad Small Form-Factor Pluggable

RA Random Access

RAT Radio Access Technology

RAR Random Access Response

RAU Remote Antenna Unit

RF Radio Frequency

RoF Radio over Fiber

RRM Radio Resource Management

RSRP Reference Signal Received Power

RSSI Received Signal Strength Indicator

RTS Request-To-Send

RTS-NI Request-To-Send-Node Information

RX Receiver

SC Small Cell



SD-FEC Soft-Decision Forward-Error Correction

SDM Space Division Multiplexing

SDMA Space Division Multiple Access

SDN Software Define Network

SFF Small Form Factor

SFP Small Form-Factor Pluggable

SiGe Silicon-Germanium

SISO Single Input Single Output

SLS Sector Level Sweep

SM Spatial Multiplexing

SME Small and Medium-sized Enterprise

SMF Single Mode Fiber

SNR Signal to Noise Ratio

SOTA State Of The Art

SP Service Period

SPI Serial Parallel Interface

SRC Sample Rate Conversion

SSB Single-SideBand

SSW Sector SWeep

SSW-FBCK Sector SWeep FeedBaCK

STA STAtion

STM-1 Synchronous Transport Module, level 1

STS Symbol Timing Synchronization

TAB-MAC Terahertz Assisted Beamforming Medium Access Control

TDD Time Division Duplexing

TDM Time Division Multiplexing

TDMA Time Division Multiple Access

TERRANOVA Terabit/s Wireless Connectivity by Terahertz innovative

technologies to deliver Optical Network Quality of

Experience in Systems beyond 5G

THz Terahertz



TIA TransImpedance Amplifier

TWDM Time and Wavelength Division Multiplexed

Tx Transmitter

TXOP Transmission Opportunity

UL Uplink

UE User Equipment

VCO Voltage Controlled Oscillator

VGA Variable Gain Amplifier

VLC Visible Light Communication

WLAN Wireless Local Area Network

WDM Wavelength Division Multiplexing

WiFi Wireless Fidelity

WiGig Wireless Gigabit alliance

WLBGA Wafer Level Ball Grid Array

WM Wireless Microwave

XG-PON 10 Gbit/s Passive Optical Network

XPIC Cross Polarization Interference Cancellation

YANG Yet Another Next Generation



Contents

1. INTRODUCTION ..................................................................................................................... 19

1.1 Scope ............................................................................................................................. 21

1.2 Structure ........................................................................................................................ 21

2. Fundamental particularities of THz systems & TERRANOVA Challenges .............................. 22

2.1 THz channel & system particularities ............................................................................ 22

2.1.1 THz channel characteristics ................................................................................... 22

2.1.2 Directed THz wireless channel............................................................................... 23

2.1.3 Heterogeneity ........................................................................................................ 26

3. MAC and RRM layer functionalities ....................................................................................... 27

3.1 Channelization ............................................................................................................... 27

3.2 Users association ........................................................................................................... 28

3.3 Initial access ................................................................................................................... 32

3.4 Mobility management and handovers .......................................................................... 35

3.4.1 UE tracking ............................................................................................................. 35

3.4.2 Handovers .............................................................................................................. 36

3.5 Interference management ............................................................................................ 36

3.6 Multiple access schemes ............................................................................................... 38

3.6.1 Scheduled access ................................................................................................... 41

3.6.2 Random access ...................................................................................................... 42

3.6.3 Hybrid .................................................................................................................... 46

3.7 Caching .......................................................................................................................... 46

3.7.1 System Infrastructure Architecture ....................................................................... 47

3.7.2 Caching Systems Main Entities .............................................................................. 49

3.7.3 System operation .................................................................................................. 50

3.7.4 System signalling and call flows ............................................................................ 51

3.7.4.1 Metamac signalling ............................................................................................ 51

3.7.4.2 Cache flows ........................................................................................................ 54

3.7.4.3 Content localization ........................................................................................... 58

3.7.4.4 Handover ........................................................................................................... 62

4. Performance evaluation, initial results and discussion ......................................................... 64



4.1 User association ............................................................................................................ 65

4.2 Initial access ................................................................................................................... 68

4.3 Random access .............................................................................................................. 72

4.4 Random vs Scheduled Access ........................................................................................ 74

4.5 Caching .......................................................................................................................... 75

5. Conclusions ............................................................................................................................ 78

6. References ............................................................................................................................. 80



List of Figures

Figure 1: Uncovered area as a function of the average LoS BSs per square meter. Options 1, 2

and 3 respectively correspond to omni-, semi- and full-directional operation modes. ............... 24

Figure 2: The coverage area as a function of the operating beamwidth, assuming 10-5 BS/m2.

Options 1, 2 and 3 respectively correspond to omni-, semi- and full-directional operation modes.

....................................................................................................................................................... 25

Figure 3: The coverage area as a function of the operating beamwidth, assuming 10-5 BS/m2.

Options 1, 2 and 3 respectively correspond to omni-, semi- and full-directional operation modes.

....................................................................................................................................................... 26

Figure 4: Space-dependent UE grouping. ...................................................................................... 28

Figure 5: Dynamic cell network topology. ..................................................................................... 29

Figure 6: IA procedure. .................................................................................................................. 33

Figure 7: PSS transmission. ............................................................................................................ 34

Figure 8: The components of the BI. ............................................................................................. 40

Figure 9: The structure of the BHI phase....................................................................................... 41

Figure 10: The system architecture based in caching. .................................................................. 48

Figure 11: Metamac signalling without cache. .............................................................................. 52

Figure 12: An example of cache flows between different system’s entities................................. 55

Figure 13: Content localization sequence. .................................................................................... 59

Figure 14: Handover call flows from caching perspectives ........................................................... 63

Figure 15: Indicative BS-UE association example. ......................................................................... 65

Figure 16: Convergence rate plot of the average fitness per iteration obtained by GWO and PSO.

....................................................................................................................................................... 66

Figure 17: CDF of the total data rates found by GWO and PSO. ................................................... 66

Figure 18: Utility function value obtained by both algorithms in different executions. ............... 67

Figure 19: Percentage of UEs served by the network obtained by both algorithms in each

execution. ...................................................................................................................................... 67

Figure 20: Average utility function values and average parentage of UE covered as a function of

the number of UEs, obtained by both algorithms for radius equals 10 m. ................................... 68

Figure 21: The average number of epochs for discovering a UE as a function of the LoS BS density

per unit of area. ............................................................................................................................. 69

Figure 22: Receiver operation curves (ROCs) for Ntx=Nrx=16, K=1, Ts = 10 μs, Ws = 1 MHz, and SNR

= 0 dB. ............................................................................................................................................ 70

Figure 23: False-alarm probability as a function of the detection threshold, for different values of

Ndiv, assuming Ntx=Nrx=16, K=1, Ts=10 μs, Ws = 1 MHz, and SNR=0 dB. ........................................ 71

Figure 24: Detection probability as a function of the detection threshold, for different values of

Ndiv, assuming Ntx=64 and Nrx=16, K=1, Ts=10 μs, Ws = 1 MHz, and SNR=0 dB.............................. 71

Figure 25: Optimal transmit probability for slotted ALOHA [68]. ................................................. 72

Figure 26: Throughput vs distance for scheduled and random access, assuming different number

of sectors. ...................................................................................................................................... 75

Figure 27: Caching vs no-caching. ................................................................................................. 77



List of Tables

Table 1: Technical literature on directional MAC protocols.......................................................... 39 Table 2: Caching Layers vs Technologies. ...................................................................................... 48 Table 3: Entities concerned by each system operation mode. ..................................................... 50 Table 4: Signalling entities. ............................................................................................................ 52 Table 5: MetaMac signalling and call flows. ................................................................................. 53 Table 6: Entities definition and abbreviation. ............................................................................... 54 Table 7: An example of call flows between different system’s entities. ....................................... 56 Table 8: Content localization Sequence. ....................................................................................... 59 Table 9: Handover entities definition. ........................................................................................... 62 Table 10: Handover call flows from caching perspectives ............................................................ 63 Table 11: Simulation parameters for the random access scenarios. ............................................ 73 Table 12: Simulation parameters - Scheduled vs random access scenarios. ................................ 74



Executive Summary

The present deliverable “D4.1 TERRANOVA’s MAC layer definition & resource management

formulation,” focuses on the definition of the TERRANOVA’s medium access layer (MAC) and radio

resource management (RRM) approaches for realizing Terabit/s wireless connectivity by using

Terahertz innovative technologies to deliver optical network quality of experience in systems

beyond 5G. In more detail, Section 2 discusses the fundamental characteristics of the THz channel

and the systems particularities that TERRANOVA MAC and RRM layers need to countermeasure.

Next, Section 3, after defining the radio resource block in the time-frequency-space domain, the

MAC and RRM required functionalities are listed and described. Moreover, possible

implementations of those functionalities are provided. Based on these implementations, in

Section 4, initial results that quantify their performance are given.

The main outcomes of the deliverable are:

An overview of the particularities of the wireless THz system that need to be counter-

measured by the MAC and RRM layers;

The required functionalities in both MAC and RRM layer to support the identified use

cases of the TERRANOVA system;

The description of possible utilization and approaches for the THz MAC/RRM protocols;

The presentation of initial simulation results that evaluate the performance of the

proposed procedures and schemes.



1. INTRODUCTION

Wireless data traffic has drastically increased accompanied by an increasing demand for higher

data rate transmission. In particular and according to the Edholm’s law of bandwidth [1], wireless

data rates have been doubled every 18 months over the last three decades and are quickly

approaching the capacity of wired communication systems [2]. In order to address this

tremendous capacity demands, the wireless word has moved towards the fifth generation (5G)

era, by introducing several novel approaches, such as massive multiple-input multiple-output

(MIMO) systems, full-duplexing, and millimetre wave (mmWave) communications. However,

there is a lack of efficiency and flexibility in handling the huge dynamic range of quality of service

(QoS) and experience (QoE) oriented data services [3].

In view of the fact that the currently used frequency spectrum for 5G has limited capacity, THz

wireless became an attractive complementing technology to the less flexible and more expensive

optical-fibre connections as well as to the lower data-rate systems, such as visible light

communications (VLCs), microwave links, and wireless fidelity (WiFi) [4], [5], [6], [7], [8], [9], [10].

As a consequence, by enabling wireless THz communications, we expect not only to address the

spectrum scarcity and capacity limitations [11] of the current cellular systems, but also to boost a

plethora of life-changing applications. Therefore, they will influence the main technology trends

in wireless networks within the next 10 years and beyond.

Motivated by this, the objective of the project TERRANOVA is to provide unprecedented

performance excellence, not only by targeting data rates in the Tbit/s regime, but also by

inherently supporting novel usage scenarios and applications, such as virtual reality, virtual office,

etc., which combine the extreme data rates with agility, reliability and almost-zero response time.

Additionally, in the near future, users in both rural and remote regions, in which the access is not

easily established (e.g., mountains and islands), should be connected with high data rates up to

10 Gbit/s per user, since it has been proven that access to high-speed internet for all, is crucial in

order to guarantee equal opportunities in the global competition. Nowadays, this is either

infeasible or prohibitively costly, when using solely optical fibre solutions. As a result, the use of

wireless THz transmission as backhaul extension of the optical fibre is an important building block

to bridge the ‘divide’ between rural areas and major cities and guarantee high-speed internet

access everywhere, in the beyond 5G era. Finally, the increasing number of mobile and fixed end

users as well as in the industry and the service sector will require hundreds of Gbit/s in the

communication to or between cell towers (backhaul) or between remote radio heads located at

the cell towers and centralized baseband units (fronthaul).

In this context, the previously submitted deliverables of the WP2, namely D2.1 and D2.2

presented the concept of blending wired-optical and wireless-THz links, its requirement and the

possible applications of TERRANOVA. Moreover, the TERRANOVA system hardware enablers and

physical layer functionalities were identified in D2.2. This deliverable, entitled “D4.1

TERRANOVA’s MAC layer definition & resource management formulation” focuses on the

definition of the TERRANOVA’s medium access layer (MAC) and radio resource management

(RRM) approaches for realizing Terabit/s wireless connectivity by using Terahertz innovative

technologies to deliver optical network quality of experience in systems beyond 5G. In more



detail, after identifying the THz system particularities and characteristics that affect the design of

the MAC and RRM layers, we determine the corresponding functionalities that need to be utilized

and we present an initial design of multi-user/device medium access control (MAC) protocols for

THz communication systems able to cope with distance-dependent bandwidth, the new channel

and interference models and pencil-beam antenna concepts. By building on top of the THz

network, a system of distributed replicated caches in the small cell access points that allows for

decreased congestion in the backhaul network and ensures fast and distributed delivery of

content, is provided. Finally, by exploiting the findings of the work package 3, regarding the

channel and propagation characterisation, the pencil beam-forming and the THz information

theory along with other peculiarities of THz communications, this deliverable introduces a highly

adaptable framework for optimal THz user equipment association.



1.1 Scope

This deliverable focuses on THz wireless access and resource management concepts with a view

to optimise the end-to-end (E2E) TERRANOVA system. In more detail, we explore and design

multi-user/device medium access control (MAC) protocols for THz communication systems able

to cope with distance-dependent bandwidth, new channel and interference models and pencil-

beam antenna concepts, by capitalising on extreme densification and hybrid/flexible

deterministic/random access principles. Moreover, building on top of the sub-THz/THz

communication network, a system of distributed replicated caches is provided -in the small cell

access points- that allows for decreased congestion in the backhaul network and ensures fast and

distributed delivery of content. Finally, by exploiting the outputs of the work package 3 (WP3),

regarding the channel and propagation characterisation, the pencil beam-forming and the THz

information theory along with the other peculiarities of THz communication, this deliverable aims

at introducing a highly adaptable framework for optimal THz resource management.

1.2 Structure

The structure of this deliverable is as follows:

In Section 2, the fundamental characteristics of the THz channel and the system

particularities that TERRANOVA MAC and RRM layers need to counter-measure, are

discussed.

In Section 3, after defining the radio resource block in the time-frequency-space domain,

the MAC and RRM required functionalities are listed and described. Moreover, possible

implementations of these functionalities are provided.

In Section 4, initial results that quantify their performance are given.

In Section 5, the main messages and findings of D4.1 are summarized, conclusions are

drawn and sets of the consortium’s future goals concerning the design of MAC and RRM

layer are defined.



2. FUNDAMENTAL PARTICULARITIES OF THZ SYSTEMS &

TERRANOVA CHALLENGES

2.1 THz channel & system particularities

In this section, we focus on the fundamental characteristics of THz channels and systems that will

affect the design of the MAC mechanisms and schemes. These characteristics can be summarized

as follows:

In the THz region, because of the small wavelength, we are able to design highly

directional transmit antennas and receive antennas with low acceptance angle. These

antennas are employed to countermeasure the high channel attenuation. However, at

the same time, they require an extremely accurate alignment between the

communication nodes.

The large material absorption in the THz band rendersquestionable the use of the non-

line of sight (NLOS) communications. As a consequence, beam tracking schemes as well

as coordinated multi-point (CoMP) needs to be used in order to guarantee uninterrupted

communication.

Molecular absorption in the THz frequencies causes frequency- and distance-dependent

pathloss (PL), which makes specific frequency windows unsuitable for establishing a

communication link. Therefore, despite the high bandwidth availability in the THz region,

windowed transmission with time varying loss and per-window adaptive bandwidth

usage is expected to be employed.

In order to increase the links capacity, suitable multiple-input and multiple-output

(MIMO) techniques in combination with beamforming (BF) need to be employed.

In order to countermeasure the impact of transmitter (TX) and RX misalignment and

support tracking of mobile or moving user equipment (UE), adaptive beamsteering is

expected to be utilized. Adaptive beamsteering enables the low-complexity link

installation and guarantees that the TX and RX antennas are aligned.

In all cases, due to small wavelengths, there are high requirement on intra and inter beam

coherence.

Furthermore, due to small wavelengths, multi-path fading in case of NLOS link will be

quickly changing at small spatial movements leading to highly time-varying non-flat

channel characteristics in nomadic applications.

Finally, adaptive modulation and coding (AMC) schemes will be employed in order to

increase both the range and the throughput of the THz system, while, at the same time,

guarantee a pre-defined degree of reliability.

2.1.1 THz channel characteristics

The THz links suffer from several path loss mechanisms, including the free space path-loss (FSPL)

and the molecular absorption loss (MAL). The latter is a distinguishing feature of the millimetre

and sub-millimetre bands. The main difference between the mmWave and the THz band is the

progressively increasing MAL. At longer link distances, below 300 GHz links are dominantly

attenuated by the FSPL, whereas in the THz regime, the molecular absorption becomes more



important, due to its exponentially increasing impact as a function of distance. At short distances

(tens of meters), the FSPL remains the dominant loss below one THz frequencies. The existence

of the molecular absorption loss highly depends on the level of the water vapor in the

atmosphere, since, in this band, water vapor efficiently absorbs energy. This indicates that there

are global variations in the level of absorption. Therefore, for standardization, it is important to

take into account the regional propagation properties.

Although, the THz frequencies are often considered in full line-of-sight (LOS) conditions, it has

been shown that communications via non-LOS (NLOS) paths is also possible, e.g., by reflections

[12] and penetrations [13]. The THz signals are rather easily blocked by thicker objects, such as

humans. Thus, the possibility of utilizing the NLOS paths as primary communications channels, is

plausible in the case the LOS path is not available. If the THz links are established outdoors, one

will have to take into account that, e.g., rain, fog, and clouds also attenuate THz signals [14] [15].

These are random phenomena, but for which the probability of occurrence is also dependent on

the time of the year, location, and altitude on Earth.

In order to accommodate with the THz channel particularities, several models have been

presented in the open technical literature. For instance, in [16] a simple channel model for 275 –

400 GHz frequency band is proposed. In [17], the authors presented an initial path-loss model for

nano-sensor networks that describes the THz propagation behaviour through plant foliage,

whereas, in [18], a channel model for intra-body nanoscale communications was provided.

Additionally, in [19], a model was proposed, which evaluates the total absorption loss, assuming

that the propagation medium is the air, natural gas and/or water. In [20], the authors presented

a multi-ray THz model for THz communications. All the published results agree on the fact that

the THz communication channel has a strong dependence on both the molecular composition of

the medium and the transmission distance.

2.1.2 Directed THz wireless channel

Due to the fundamental characteristics of the THz systems, it is evident that the propagation

environment suffers from sparse-scattering. This causes to the majority of the channel direction

of arrivals (DoAs) below noise floor figures. As a consequence, in order to achieve a sufficient

coverage, as illustrated in Figure 1, a channel in a wireless THz system can be established in a

specific direction with a range that varies according to the directionality level. However, the

directionality of wireless THz channels results in two consequences, namely:

Blockage, which refers to the high penetration loss, due to obstacles and cannot be solved

by just increasing the transmission power; and

Deafness, which refers to the situation, in which the main beams of the transmitter and

the receiver are not aligned to each other. This prevents the establishment of the

communication link.

In order to overcome blockage, the wireless THz system is required to search for and identify

alternatives directed spatial channels, which are not blocked. However, this search entails a new

BF overhead of significant amount and hence it introduces a new type of latency, in which we will

refer to as BF latency. As a consequence, the medium access control (MAC) design for cellular

networks is more complicated than the one of the conventional wireless local area networks



(WLANs), in which short range communications can be also established through NLOS

components. Additionally, the conventional notion of cell boundary becomes questionable in

these systems, due to the randomly located obstacles. As a result, a redefinition of the notion of

the “traditional cell” into “dynamic cell” is required.

On the other hand, deafness has a detrimental effect on the complexity of establishing the link

and causes a synchronization overhead increase. This indicates the importance of redesigning the

initial access (IA) procedures.

Figure 1: Uncovered area as a function of the average LoS BSs per square meter. Options 1, 2 and 3 respectively correspond to omni-, semi- and full-directional operation modes.

Operation frequency 0.4 THz

Bandwidth of the CC 50 KHz

SNR threshold 0 dB

BS transmission power 30 dBm



Figure 2: The coverage area as a function of the operating beamwidth, assuming 10-5 BS/m2. Options 1, 2 and 3 respectively correspond to omni-, semi- and full-directional operation modes.

Figures 2 and 3 depict the coverage area as a function of the operating beamwidth assuming

10-5 and 10-2 BSs/m2, respectively, for three different deployment options, namely:

Option 1 - Omni-directional operation mode: In this option both the BS and UE employ

omni-directional antennas;

Option 2 – Semi-directional operation mode: In this option the BS uses a directional

antenna, while the UE employs an omni-directional antenna; and

Option 3 – Full-directional operation mode: In this option both the BS and UE employ

directional antennas. Note that in this option, it is assumed that the TX and RX antennas

are aligned.

In both figures, the operation frequency is set to 0.4 THz, while the bandwidth of the CC is 50 kHz.

Moreover, the BS transmission power is 0 dB and the SNR threshold equals 0 dB.



SNR threshold 0 dB


BS density [BS/m2] 10-5



Figure 3: The coverage area as a function of the operating beamwidth, assuming 10-5 BS/m2. Options 1, 2 and 3 respectively correspond to omni-, semi- and full-directional operation modes.

2.1.3 Heterogeneity

In order to countermeasure the physical limitations of wireless THz systems, the MAC mechanisms

may simultaneously exploit both the microwave and THz bands [21]. This was initially presented

in [21], where the authors provided a MAC protocol that employs microwave frequencies for

establishing the control channels (CCHs), and THz frequencies for the data channels (DCH).

Additionally, MAC mechanisms may need to facilitate the co-existence of several communication

technologies with different coverage. As a consequence, two different types of heterogeneity are

observed in wireless THz networks, namely:

Spectrum heterogeneity; and

Deployment heterogeneity.

By spectrum heterogeneity, we refer to the scenarios in which wireless THz UEs use both high

(THz) and lower frequencies (e.g. in the microwave band). On the one hand, THz frequencies

provide a massive amount of bandwidth for high data rate communications. On the other hand,

the microwave frequencies are used for control message exchange, which demands much lower

data rates, but higher reliability than data communications. This facilitates the deployment of

wireless THz networks, due to possible omnidirectional transmission/reception of control

messages, as well as higher link stability, at lower frequencies. However, the use of both

microwave and THz bands in UEs increases the fabrication cost and may result to an important

reduction of the mobile UE (MUE) energy autonomy. Moreover, due to the blockage and deafness

effects, the establishment of a microwave CCH might not (necessarily also) result in establishing

the corresponding THz data transmission channel.



SNR threshold 0 dB


BS density [BS/m2] 10-2



The deployment heterogeneity introduces two scenarios for THz cellular network, namely:

stand-alone networks; and

integrated networks.

In the stand-alone scenario, a complete THz network (from macro to pico levels) will be deployed,

whereas the integrated network solution is an amendment to existing microwave networks for

performance enhancement, and includes wireless THz small cells (SCs) and/or THz hotspots [22].

3. MAC AND RRM LAYER FUNCTIONALITIES

The organization of this section is as follows:

Section 3.1 defines and clarifies the structure of the resource block (RB).

Section 3.2 presents the optimization problem that need to be solved in order to associate

UEs with BS in a wireless THz network. Moreover, a machine learning approach, which

provide the solution to this problem is provided.

Section 3.3 discusses the initial access procedure.

Section 3.4 provides mobility management and handover mechanisms.

Section 3.5 clarifies the types of interference that we need to deal with in wireless THz

communications.

Section 3.6 presents random and scheduled access multiple access schemes and

mechanisms.

Finally, Section 3.7 provides the caching architecture and its functionalities.

3.1 Channelization

An important decision in the design of the MAC and RRM layers is the definition of the resource

blocks (RBs), i.e., the smallest unit of the physical resources that can be allocated. In LTE, a RB is

defined as a portion of the time-frequency domain with certain frequency range and time

duration. On the other hand, in wireless THz systems, directional transmissions enable to also use

the spatial dimension. As a result, RBs can be defined in the time-frequency-space domain,

providing possibility of much improved capacity for network due to resource reuse. As illustrated

in Figure 4, in order to properly utilize this type of RB, the BS needs to group a set of UEs together,

which are non-distinguishable in the transmitted beam, and serve each group with one analogue

BF vector [23]. This indicates that an analogue BF will be used to serve only one group at a time,

while hybrid BF should be employed in order to simultaneously serve multiple groups. The

analogue BF partially utilizes the spatial dimension of the RB, whereas the digital BF increases the

multiplexing gain within each group.



Figure 4: Space-dependent UE grouping.

3.2 Users association

Most of the current standards define a cell by the set of UEs that are associated using a minimum-

distance rule, which leads to non-overlapping Voronoi tessellation of the serving area of every BS,

exemplified by the well-known hexagonal cells [24], [25]. The minimum-distance rule leads to a

simple association metric based on the reference signal received power (RSRP) and RSSI.

However, the traditional RSRP/RSSI-based association may become significantly inefficient in the

presence of non-uniform UE spatial distribution, heterogeneous BSs with a different number of

antenna elements, and different transmission powers [26]. This association entails an unbalanced

number of UEs per cell, which limits the available resources per UE in highly populated cells,

irrespective of individual signal strengths, while wasting them in sparse ones. The main

disadvantage of the current static definition of a cell, as a predetermined geographical area

covered by a BS, is that the static cell formation is independent of the cell load as well as of the

UEs’ capabilities. Therefore, the parameters that should affect the cell formation are:

UE traffic demand;

channel between UE and BSs; and

BSs’ loads. By taking into account, the massive number of degrees of freedom that fully-directional

communication offers and possible MAC layer analogue BF, we can define a dynamic cell as a set

of not necessarily collocated UEs that are served by the same beam, independent from the fact

that they are close or far away from the BS (as long as they are within the same beam of certain

beamwidth) and dynamically selected to improve some objective function. Upon any significant

fluctuations of the above three parameters, dynamic cell redefinition may be required. In this

context, a full database in the macro-cell BS is needed, which records the following attributes:



Dynamic cell formations;

UEs’ traffic demands;

UEs’ QoS levels; and

UEs’ connectivity to the neighbouring BSs. Depending on the UEs’ demands, microcell BSs dynamically group UEs together and form new

cells so that:

individual UE’s demands are met (QoS provisioning);

the trade-off between macro-level fairness and spectral efficiency is improved - e.g., through proportional fair resource allocation (network utility maximization); and

every UE is associated with at least two groups, to guarantee robustness to blockage (connection robustness).

Figure 5: Dynamic cell network topology.

Based on the above criteria, we formulate the corresponding optimization problem for the BS-UE

association, assuming a scenario, in which Nb BS can serve Nu UE (see Figure 5) in a fully-directional

mode. In more detail and as reported in [27], for a fixed network topology, which is assumed to

be prior known to every BS i and UE j, the optimal cell formation should satisfy the following

optimization problem:

max𝜳,𝑪

∑ 𝑟𝑖𝑗𝑐𝑖𝑗𝑗∈𝑈,𝑖∈𝐵

,

s. t. C1: ∑ 𝑐𝑖𝑗 ≤ 1,𝑗∈𝑈 for each 𝑖 ∈ 𝐵,

C2: ∑ 𝜓𝑖𝑗 = 1𝑖∈𝐵 , for each 𝑗 ∈ 𝑈,

C3: 0 ≤ 𝑐𝑖𝑗 ≤ 𝜓𝑖𝑗, 𝜓𝑖𝑗 ∈ {0, 1}, for each 𝑖 ∈ 𝐵,

C4: 𝜓𝑖𝑗 = 0, if 𝑅𝑖𝑗 ≤ 𝑅𝑗,𝑚𝑖𝑛,

C5: 0 ≤ 𝜑𝑖𝑏 ≤ 2𝜋, for each 𝑖 ∈ 𝐵,



C6: 0 ≤ 𝜑𝑗𝑢 ≤ 2𝜋, for each 𝑗 ∈ 𝑈,

where 𝑐𝑖𝑗 is the fraction of resources that the BS i employs to serve the UE j. Likewise, 𝜑𝑖𝑏and 𝜑𝑗

𝑢

respectively stand for the boresight angles of the BS i and UE j, whereas 𝑅𝑗,𝑚𝑖𝑛 is the minimum

required rate for the UE j. Finally, Ψ and C are matrices that collect all the user association

variables, ψij, and fraction of resources, cij, used by the BS i to serve the UE j, respectively.

In the above optimization problem, the constraint C2 guarantees the association of the UE j to

exactly one BS, whereas the constraints C3 and C4 ensure that the minimum acceptable QoS for

every UE. Additionally, C3 ensures that hat every BS i will provide a positive resource share only

to its associated UEs. Note that the solution of the optimization problem provides a long-term

association policy along with proper orientation and operating beamwidths for fully-directional

wireless THz communications. This solution guarantees the optimal UE-BS association as long as

the inputs of the optimization problem, namely network topology and UEs’ demands, are

unchanged. If a UE requires more resources or loses its connection, the optimization needs to be

performed again.

In order to find the optimal solution for the association problem, we present Algorithm 1, which

is based on the swarm intelligence meta-heuristic algorithm called grey wolf optimizer (GWO) [28]

and returns the optimal user association matrix. According to the GWO, the solutions are grouped

into four groups, namely α, β, δ, and ω, based on their optimality. In more detail, xα vector

represents the best solution, while xβ and xδ are respectively the second and third optimal solution

vectors. Finally, ω is the set containing all other feasible solutions.

Algorithm 1: GWO algorithm for UE association

Input: Population size: Np Maximum iterations: Gmax

Output: X Initialize: xα the best solution

f(xα) the objective function corresponding value. 1: Calculate the objective function of each vector.

Use Algorithm 2. 2: Sort vectors according to the objective function value. 3: Select the best solution xα, the second best xβ, and the third best xδ. 4: set G=1. 5: while 𝐺 ≤ 𝐺𝑚𝑎𝑥 do 6: for k=1 to Np do 7: Update vector positions according to equation (1) 8: end for 9: Update vectors α, A, and K.

Calculate objective function value of each vector k according to Algorithm 2. 10: Select the new xα, xβ, and xδ vectors. 11: G=G+1 12: end while 13: return the best objective function vector xα and best objective function value f(xα).



A brief description of the GWO algorithm for the user association problem is given (above) in

Algorithm 1. Therein, NP denotes the population size, and Gmax is the maximum number of

generations. We notice that these are the only inputs for a given problem, since all the other

parameters are randomly generated. Thus, GWO does not require additional control parameter

setting like other algorithms. The population of NP vectors is initialized randomly from a uniform

distribution. After the objective function value of each vector is calculated the algorithm sorts the

vectors descending according to objective function value. The best solution is selected as the xα

vector, while the second best is the xβ, and the third best is the xδ vector respectively. Then the

algorithm main loop starts.

In the G+1 generation, the position of the k-th vector in the n-th dimension, is evaluated as

𝐱𝐧𝐤(𝐆 + 𝟏) = 𝐱𝐧𝐤

𝐚𝟏 (𝐆)+𝐱𝐧𝐤𝛃𝟐(𝐆)+𝐱𝐧𝐤

𝛅𝟑(𝐆)

𝟑 ( 1 )

where

𝑥𝑛𝑘𝑖𝑗

= 𝑥𝑛𝑘𝑖 (𝐺) − 𝐴𝑛𝑘

𝑗 (𝐺) 𝐷𝑛𝑘𝑖 (𝐺), for (𝑖, 𝑗) ∈ {(𝑎, 1), (𝛽, 2), (𝛿, 3)}

and

𝐷𝑛𝑘𝑖 (𝐺) = |𝐾𝑛𝑘

𝑗 (𝐺) 𝑥𝑛𝑘𝑖 (𝐺) − 𝑥𝑛𝑘(𝐺)|

with

𝐴𝑛𝑘𝑗 (𝐺) = 𝑎 (2 𝑓𝑛𝑘

1 (𝐺) − 1)

and

𝐾𝑛𝑘𝑗 (𝐺) = 2 𝑓𝑛𝑘

2 (𝐺)

Please note that 𝑓𝑛𝑘1 (𝐺) and 𝑓𝑛𝑘

2 (𝐺) are uniformly distributed randomly selected numbers in the

range of [0, 1].

The algorithm updates the α, A and K vectors according to the above relations and evaluates the

objective function value of the new vectors. Then again, a new set of xα, xβ, xδ vectors is selected,

if it improves the objective function value, provided by algorithm 2. Finally, note that the stopping

criterion is the maximum number of generations.

Algorithm 2: Evaluation of the objective function.

1: Input: A possible solution y 2: for i=1 to Nu do 3: Calculate rate rij for the i-th UE connecting to the j-th BS 4: if 𝑅𝑖𝑗 ≥ 𝑅𝑗,𝑚𝑖𝑛 then

5: Calculate cij 6: end if



7: end for 8: Calculate objective function

∑ 𝑟𝑖𝑗𝑐𝑖𝑗𝑗∈𝑈,𝑖∈𝐵

9: Return objective function value

3.3 Initial access

At THz frequencies, high antenna gains are needed to overcome the large path loss and other

losses. In order to support this scenario, IA, where UE discovers THz access point and forms link-

layer connection, needs to support BF/directionality at least in one end of the link (unless

communication distances are very small or very long preambles/high coding gain are used to

make it easier to discover users even with omnidirectional antenna modes). However, the use of

low-complexity and low-power THz devices along with the massive number of antennas, make

traditional digital BF based on instantaneous channel state information (CSI) very expensive and,

in several cases, infeasible. On the other hand, the use of analogue BF, based on predefined beam

steering vectors (beam training codebook), each covering a certain direction with a certain

beamwidth (BW), is considered as a feasible and effective alternative solution. However, one of

the main drawbacks of analogue BF is the lack of multiplexing gain, which is addressed by the

hybrid digital/analogue BF architecture.

Based on the assumed BF (i.e., fully digital, fully digital but low-resolution converters, hybrid,

analogue, etc.) multiple beams could be transmitted/received at the same time, helping to reduce

IA time. Otherwise, options include sequential scanning of all possible BF directions at the THz

access point and UE (exhaustive search), using hierarchical search, etc. The reliance on the

directional transmission in the IA in the mmWave communications renders the initial cell search

procedures different from the existing microwave systems. In the existing microwave systems,

sophisticated BF is used only after the IA procedures have been completed. In order to reduce the

time overhead, a hybrid approach has been proposed in several papers, where microwave bands

(omnidirectional search) are used for the BS discovery and then reverting to mmWave bands for

subsequent beam alignment and data transmission. As a result, the required functions for UE

detection depend on the chosen approach for IA. For example, for hierarchical search, it is

required to be able to support successively narrower beams.



Figure 6: IA procedure.

In general, all the design options considered for IA follow the same basic steps shown in Figure 6.

The main steps of the IA procedure are identical to the ones used in the third generation

partnership project (3GPP) LTE-A, which are described in [29]. However, the depicted procedure

needs major modifications for THz to enable both the UE and the BS to determine the initial BF

directions in addition to detecting the presence of the BS and the access request from the UE. In

more detail, the IA steps are:

1. Synchronization and signal detection: During this step, each BS periodically transmits

synchronization signals that the UE can scan to detect the presence of the BS and obtain the

downlink frame timing. In LTE-A, the first synchronization signal to detect is the Primary

Synchronization Signal (PSS). On the other hand, for THz systems, the synchronization signal

will additionally be used to identify the UE’s BF direction, which (for analogue BF) is related

to the angles of arrival of the signal paths from the BS.

2. Random access (RA) preamble transmission: Motivated by the LTE-A uplink frame structure,

we assume that the uplink contained dedicated slots exclusively for the purpose of RA

messages. After the synchronization and signal detection is completed, the location of these

slots is known to the UE. The UE transmits a RA preamble in one of the RA slots. Since the UE

BF direction is known after step 1, the RA preamble can be transmitted directionally. The BS

scans for the presence of the RA preamble and learns the BF direction at the BS side.

3. RA response (RAR): After detecting a RA preamble, the BS transmits a RAR to the UE, which

indicates the index of the detected response. At this point, both the BS and the UE know the



BF directions so all transmissions can obtain the full BF gain. The UE receiving the RAR knows

its preamble was detected.

4. Connection request: After receiving RAR, the UE desiring initial access will send some sort of

connection request message on the resources scheduled in the uplink grant in the RAR.

5. Scheduled communication: All subsequent communication can occur on scheduled channels

with the full BF gain on both sides.

Note that this IA procedure was initially presented in [29].

Figure 7: PSS transmission.

Similar, to long term evolution (LTE), in order to achieve spatial synchronization, as illustrated in

Figure 7, we assume that the primary synchronization signal (PSS) is transmitted periodically once

every Tp seconds in a brief interval of length of length Ts. In what follows, we call the short interval

of length Ts the PSS time slot, and the period of length Tp between two concurrent PSS slots, the

PSS period. In the THz regime, due to the severe pathloss attenuation, we consider that the BS

randomly transmits the signal in a different direction in each PSS time slot; in other words, it

randomly scans the angular space. Of note, deterministic search patterns, such as the ones

employed in IEEE 802.11ad [30], could also been employed.

Moreover, in order to exploit the higher BW available in the THz range, we assume that, in each

PSS time slot, the PSS waveform is transmitted over Nd PSS sub-signals with each sub-signal being

transmitted over a small BW, Ws. It is worth noting that the use of multiple sub-signals can provide

frequency diversity, while narrowband signalling enables the use of very low power RXs with high

SNR capabilities. Additionally, the PSS signal should fit in the minimum possible BW. A UE

searching for the BS has no knowledge of the actual BW during the cell search; therefore, all the

BSs transmit the PSS over the same BW. For THz systems, we expect that the minimum BW will

be much larger than the current fifth generation (5G) systems. Hence, a reasonable number of

Fre

qu

en

cy

… … …

…Wt

Ws

m=1 m=2 m=Ntot

d=1

d=2

d=Nd

Tp Ts



narrowband sub-signals can be accommodated, even in the minimum bandwidth. Finally, at the

UE, we assume that a generalized likelihood ratio test (GLRT) is employed to determine whether

a PSS signal is present and in which direction.

3.4 Mobility management and handovers

This section is dedicated to mobility management and handover processing. In order to address

mobility management, UE tracking with low latency and high accuracy is required. Moreover, in

order to countermeasure multiple handovers and to reduce the latency, physical layer caching is

proposed. The following subsections discuss UE tracking and handovers.

3.4.1 UE tracking

Once the THz access point (AP) and the associated UE have discovered the proper BF directions,

the UE should be tracked in order to support nomadic mobility of UE or small movements at the

transceivers of a fixed P2P link. If we adopt the conventional real-time channel estimation

schemes, the pilot overhead will be unaffordable [31], [32]. As a consequence, the MAC and RRM

layers should support beam-tracking method that consumes fewer resources compared with the

full exhaustive beam discovery. For example, we can search just the beam direction adjacent to

the current direction; this assumes that changes are small, which may not always be the case.

Searching the technical literature, there exist several papers that provide different UE tracking

approaches (see e.g., [33], [34], [35], [36], [37]). In particular, in [33], the authors proposed a

beam tracking method, which is based on detecting the signal strength via training sequences

appended to data packets. This approach has been experimentally validated by monitoring

adjacent beams (in the angular domain) in [35]. The drawback of this approach is that it requires

multiple beam pairs training. Moreover, in [36], the authors exploited the scarcity of the channel,

and reduced the problem of estimating the AoAs/AoDs, in order to estimate the virtual channel

matrix. This tracking solution relies on the assumption that the previous estimate has small

difference to the current AoA/AoD and does not exploit any dynamics of the channels. Likewise,

in [34], the authors presented a beam tracking approach that is based on fully scanning all possible

beam combinations in order to create a measurement matrix to which the extended Kalman filter

(EKF) is applied. The main drawback in [34] is the full scan requirement, which makes it difficult

to track fast-changing environments, due to the long measurement time. In contrast to [34], in

[37], the beam tracking approach relies on a single measurement; therefore, it is considered to

be more reliable for tracking fast-changing channels.

In practice, more efficient UE tracking schemes exploiting the temporal correlation of the time-

varying channels are preferred. In general, excising UE tracking schemes can be classified into two

categories. The key idea behind the schemes that belong in the first category is to model the time-

varying channels in adjacent time-slots by a one-order Markov chain and apply classical Kalman

filter in order to track the time-varying channels with low pilot overhead [38], [39]. However, the

special sparse structure of the THz beamspace channels cannot be modelled as a one-order

Markov process. As a consequence, these approaches cannot be directly extended to wireless THz

systems with multiple antennas [40].



3.4.2 Handovers

As discussed in TERRANOVA Deliverable D.2.2, the suppression of interference in THz systems

with pencil-beam operation comes at the expense of more complicated mobility management

and handover strategies. In contrast with the long term evolution advanced (LTE-A), frequent

handovers even for fixed UEs, is a potential drawback of THz systems, due to their vulnerability to

random obstacles (such as humans moving and blocking the THz signal due to the very large

penetration loss).

As a consequence, the MAC/RRM should be able to recover from user mobility outside the

coverage of the current AP/BS or from sudden blockage such as human blockage preventing to

use current AP, preferably without full initial access after current connection has been lost. Due

to cost reasons, it is not expected that THz networks will cover entire areas. Therefore, after losing

current THz connection it may be necessary to move to lower frequency communication networks

such as WLAN/LTE. Multi-connectivity may be required.

For handover preparation, the caching system should be aware of the next possible AP/BS in order

to prefetch the contents in the correct next hub. The prefetching of contents in next hub in

advance will reduce content delivery latency from the user point of view and will save backhaul

bandwidth consumption from the resource utilization point of view. These two advantages are

discussed in details in section 3.7. In this case, the cache controller should synchronize with the

resource manager. After the resource manager performs the necessary calculations for the path

prediction, this information is passed to the cache controller. This information includes the

candidate AP/BS for handover. Based on the forwarded information, the cache controller will

trigger prefetching commands to the concerned caches. When the relevant contents are retrieved

in the concerned caches, the cache controller will be informed and trigger an update to the

resource manager. Based on such an update, besides the updated network conditions, the

handover decision is made.

The efficiency of the cache controller during this process can be seen as the efficiency of caching

system to prefetch the needed data in an optimized manner (no unnecessary prefetching is

performed). The QoE of a user is dependent on the communication performance between the

cache controller and the resource manager. The precision of the path prediction from the

resource manager affects both the user QoE and the caching system performance.

3.5 Interference management

Due to the pencil beam nature of THz transmissions, interference from sidelobes of the antenna

can be ignored in many cases; similarly, interference from non-LOS transmitters can be negligible.

Therefore, typical interference could be modelled with on/off behaviour with quite low

probability of interference. In some cases, this could lead to noise-limited behaviour instead of

interference-limited behaviour [41] . In other cases, scheduling and proper beamforming design

can be used. This also leads to challenges for MAC/RRM, as to how to extract maximum

performance from systems taking into account spatial reuse.

Let us next discuss how directionality can help with different interference types. In wireless

networks, there are three main types of interference that should be managed [42]:



Intra-cell interference: This is the interference among the UEs in a cell. It can be avoided

by orthogonally allocating resources to users (in time/frequency/space). Since orthogonal

allocation would limit spectral efficiency, approaches enabling spectrum re-use are more

beneficial. Device to device (D2D) users (that communicate with each other directly) will

potentially cause and receive additional interference. However, highly directional pencil

beams will avoid most of this kind of interference. For random access, even very simple

approaches such as slotted Aloha can be efficient.

Inter-cell interference: This means interference among different cells. Especially at higher

THz frequencies and longer distances this is expected to be very low due to exponentially

increasing molecular absorption. In effect, molecular absorption protects from inter-cell

interference. Even at lower THz frequencies, highly directional beams make interference

unlikely. If link density is very high, we may need to take approaches to manage inter-cell

interference. Well known from LTE networks are for example fractional frequency reuse

(FFR) and soft frequency reuse (SFR). Additionally, if information exchange is enabled

between cells (such as by X2 interface), more advanced interference management

methods can be used. Due to highly available bandwidth at THz, it may even be feasibly

to employ reuse factor greater than one for simplicity. In some scenarios, such as THz

networks inside office rooms (attocells), there is substantially reduced need for any kind

of interference management since the walls/doors highly attenuate the signals between

different rooms.

Inter-layer interference: This means interference among different levels of network (such

as macro, micro, pico, femto, atto). It is expected that THz frequencies will be mainly used

or small cells (such as femto/atto) and not for levels requiring longer communication

distances. This separation will avoid this kind of interference. Of course, with large dish

antennas (or massive antenna arrays) THz communication can be used even over longer

distances but this will be typically limited to fixed point-to-point links. In case that THz is

used at several layers simultaneously, techniques from LTE such Enhanced-ICIC (e-ICIC)

can be used to mitigate interference among different levels [43]. Naturally, the highly

directional nature of the THz pencil beams will help also in this case.

Additionally, in case of multipath inter-symbol interference (ISI) may occur. To mitigate this, well-

known approaches such as OFDM or equalizer can be used.

To sum up, directional communications with pencil-beam operation drastically reduce multiuser

interference in THz networks. An interesting question is whether in this case a THz network is

noise-limited, as opposed to interference-limited networks. This fundamental question affects the

design principles of almost all MAC layer functions. For instance, as the system moves to the noise-

limited regime, the required complexity for proper resource allocation and interference

avoidance functions at the MAC layer is substantially reduced. On the other hand, pencil-beam

operation complicates negotiation among different devices in a network, as control message

exchange may require a time consuming beam training procedure between transmitter and

receiver.



3.6 Multiple access schemes

Novel MAC protocols are required for THz band communication networks, since classical solutions

are unable to accommodate the following characteristics:

The THz band provides devices with a very large bandwidth. As a consequence, the THz

devices do not need to aggressively contend for the channel. In addition, this very large

bandwidth results in extremely high bit-rates; hence, very low transmission durations are

achieved. This causes a very low collision probability.

The use of very large arrays and very narrow directional beams can also clearly reduce the

multi-user interference. This comes with the cost that high-bit-rates and pencil-beams

significantly increase the synchronization requirements [41].

As a result, the new MAC protocol need to be designed and developed that accommodates the

following objectives:

To support both random and scheduled access;

To employ pencil-beamforming to countermeasure the high channel attenuation;

To countermeasure deafness, by guaranteeing alignment between the TX and the RX,

through the development of RX-initiated transmission schemes.

To exploit fast-steerable pencil beams for network-wide objectives. The fast steering and

pattern control capabilities of very large antenna arrays enable new functionalities that

can be exploited to control interference. For example, in the DL, a TX can simultaneously

send messages to multiple receivers, by transmitting time-interleaved pulses in different

directions so that the pulses intended for each user add coherently at the desired RX,

with no interference. There exists a trade-off between user acquisition complexity and

communication robustness, in terms of BW choice. A possible scheme to fully harness

the huge antenna array is to adapt wider BW during scanning phase for fast user

acquisition, and intelligently steer focused thin beam for the subsequent data

communication.

Table 1 reports the technical literature on directional MAC protocols. In particular, the first

proposed MAC protocols [42], [43] and [44], employed omni-directional RTS/CTS (ORTS/OCTS)

packets in order to determine the best beam direction. The disadvantage of this approach is that

the communication range of the devices is significantly constrained. Next, directional-omni (DO)

MAC protocols were presented, which allowed omni-directional reception of directional

transmitted signals (see for example [45], [46], [47], and [48]). The main challenge of using

directional RTS/CTS (DRTS/DCTS) is to determine the best beam direction. As a consequence,

these approaches assume that the location information of the nodes are acquired through the

global positioning system (GPS) [45], upper layer [47], or the determination of the DoA of

incoming transmissions by more complicated RXs, which employ digital BF. The disadvantage of

these protocols is that they usually suffer from hidden terminals and deafness problems, since

they do not inform their neighbours about the ongoing communication. Likewise, the protocols

presented in [49], [50], [51] and [52], utilized multi-directional circular RTS/CTS (CRTS/CCTS)

control packets in order to suppress the hidden terminal and deafness problems, while supporting

transmission range to DO neighbours. In particular, the RTS and CTS packets are emitted in



directional mode sequentially from each of the sectors in order to inform all the neighboring

nodes about the ongoing transmission and hence minimize the deafness problem. This approach

causes the pollution of the medium with an enormous amount of control packet overhead, before

each transmission. As a result, the overall throughput of the system is significantly reduced. In

order to overcome the throughput reduction, the transmission of multi-directional concurrent

RTS/CTS packets in [52], [53], and [54], as well as the use of an additional channel for control

packet transmission [55], [56], [57], [58], [59], [60], [61], [62] were proposed. Despite the fact

that these protocols solve the overhead problem, they are not compatible with IEEE 802.11ay

transceiver, due to the requirement of sophisticated transceivers for concurrent transmission.

Additionally, all these protocols assume either the availability of the location information or

continuous tracking of the location of neighboring nodes by either demanding extra packet

overhead, such as periodic hello packet transmission, or more complex hardware for DoA

detection.

Table 1: Technical literature on directional MAC protocols.

Ref. RTS CTS Range Beamforming Information

Antenna Chan-nel(s)

MAC Challenge Addressed

Deafness Hidden terminal

[42] Omni Omni OO3 DoA Switched Single No No

[43] Omni Omni OO Exchange Antenna Weights

Adaptive array

Multi No Yes

[44] Dir Omni OO Exchange Antenna Weights

Adaptive array

Multi No Yes

[45] Dir Omni OO GPS Switched Single No No

[46] Dir Dir DO4 DoA Adaptive array

Single No No

[47] Dir Dir DO Upper layer Adaptive array

Single No No

[48] Dir Dir DO Assumed available

Switched Single Yes No

[63] Multi-Dir

sequen-cial

Dir DO DoA Switched Single Yes Yes

[49] Multi-Dir

sequen-cial

Multi-Dir

seque-ncial

DO DoA Switched Single Yes Yes

[50] Multi-Dir

Multi-Dir

DO Upper layer Switched Single Yes Yes

3 OO: Omni-Omni. 4 DO: Directional-Omni.



sequen-cial

seque-ncial

[51] Multi-Dir

sequen-cial

Multi-Dir

seque-ncial

DO Assumed available

Switched Single Yes Yes

[52] Multi-Dir

sequen-cial

Multi-Dir

seque-ncial

OO Periodic updates

Adaptive array

Single Yes Yes

[53] Multi-Dir

sequen-cial

Multi-Dir

seque-ncial

DO Assumed available

Adaptive array

Single Yes Yes

[54] Dir Dir DO Hello Packet Adaptive array

Single Yes Yes


Switched Multi Yes No

[56] Dir Dir DO DoA Switched 2 channels

Yes No


Yes Yes

[58] Dir Dir DO DoA Switched Multi Yes Yes


Switched Multi Yes No


Yes Yes

[61] Omni Omni OO GPS Switched Multi Yes Yes

[62] Omni Omni OO Assumed available

Adaptive array

Multi Yes Yes

Motivated by the above requirement and based on the previous published technical literature,

the medium access can be organized in beacon intervals (BIs), similar to IEEE 802.11ay [64]. Each

BI should support BF association, management frame exchange between the access point (AP)

and the beam-trained stations, and data transmission. As a consequence, as illustrated in Figure

8, the BI consists of three phases, namely:

Beacon header interval (BHI);

Announcement transmission interval (ATI); and

Data transmission interval.

BHI ATI DTI

Figure 8: The components of the BI.

The BHI is responsible for the spatial synchronization, during the IA. Therefore, as demonstrated

in Figure 9, it consists of two periods, namely beacon transmission interval (BTI) and associate BF



training (A-BFT). During the BTI, the AP performs transmit sector sweep (TXSS), by sending

synchronization signal (beacons), i.e., a sequence of packets, with each of them using a different

antenna weight vector (AWV), in such a way that the transmissions include all directions, i.e., the

sectors, of coverage of the AP. In the meanwhile, the STA continuously tries to detect a packet,

by performing receive sector sweep (RXSS) and when it is able to lock to a beacon frame, it

decodes its header and data field. The information contained in the payload enables the UE to

associate to the AP. After the UE has been associated with the AP, it randomly selects one of the

A-BFT slots in order to perform TXSS. Note that since A-BFT is slotted, collisions may occur, during

this phase, when two UEs select the same A-BFT slot. Once collision occurs, the A-BFT slot may be

unavailable for BF training, which will result in great waste of BF training opportunities. In order

to further alleviate the problem of high collision in ultra-dense user scenarios, a secondary back-

off mechanism (SBOM) is utilized5.

BHI

BTI A-BFT

Figure 9: The structure of the BHI phase.

During ATI, the AP exchanges management information with associated and beam-trained STAs,

and only the AP can initiate a frame transmission. The frames transmitted in ATI are limited to the

request and response frames such as management frame, acknowledgment (ACK) frame, etc. The

details of specific request and response frames can be found in [65]. In particular, ATI is

Date post:	03-Feb-2021
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

Deliverable D4.1 TERRANOVA’s MAC layer definition ... · D4.1 – TERRANOVA’s MA layer...

Documents