Study of macro and small cell deployment options in 5G

mmWave bands using ATOLL radio planning tool

A Master's Thesis
Submitted to the Faculty of the
Technical School of Telecommunications Engineering of
Barcelona
Polytechnic University of Catalonia
by
Layla Maite Guerrero Condo

In partial fulfilment
of the requirements for the degree of
MASTER IN TELECOMMUNICATIONS ENGINEERING

Advisors: Anna Umbert, PhD; Ramon Ferrús, PhD

Barcelona, February 2020

Title of the thesis: Study of macro and small cell deployment options in 5G mmWave
bands using ATOLL radio planning tool.

Author: Layla Maite Guerrero Condo.

Advisors: Anna Umbert, PhD; Ramon Ferrús, PhD.

Abstract

The new radio (NR) air interface specified for 5G systems is able to operate in frequency
bands in the range of 24.25–52.6 GHz, known as millimeter wave (mmWave) frequencies,
where important amounts of spectrum can be used to support ever-growing mobile traffic and
deliver bit data rates in the gigabit per second range.

In this context, this thesis studies the performance in terms of capacity and coverage of
different macro and small cell deployment options for the exploitation of the 26 GHz band. The
study is conducted using the ATOLL radio planning tool. Considering a geographic area in the
city of Barcelona for which a detailed characterization including the modelling of buildings at
1 meter resolution, a set of representative cell sites at different locations (e.g. macro cell sites
in open areas, small cell sites at street level, etc.) have been defined and evaluated. Results
for 26 GHz are compared with configurations using mid-frequency bands. The impact of the
channel bandwidth and the beamforming configuration is also assessed. In addition, a usage
scenario representing a crowded event is simulated under different deployment configurations
in order to evaluate the experience data rates for different service loads.

1
Resumen

La nueva interfaz aérea de radio (NR) especificada para sistemas 5G puede operar en
bandas de frecuencia en el rango de 24.25–52.6 GHz, conocidas como frecuencias de onda
milimétrica (mmWave), donde se pueden usar cantidades importantes de espectro para
soportar el crecimiento constante de tráfico móvil y entregar tasas de bits de datos en el rango
de gigabits por segundo.

En este contexto, esta tesis estudia el rendimiento en términos de capacidad y cobertura

de diferentes opciones de despliegue de células macro y small para la explotación de la banda
de 26 GHz. El estudio se realiza utilizando la herramienta de planificación de radio ATOLL.
Considerando un área geográfica en la ciudad de Barcelona para la cual una caracterización
detallada que incluye el modelado de edificios con una resolución de 1 metro, un conjunto de
sitios de células representativas en diferentes ubicaciones (por ejemplo, sitios de célula macro
en áreas abiertas, sitios de célula small a nivel de calle, etc.) han sido definidos y evaluados.
Los resultados para 26 GHz se comparan con configuraciones que utilizan bandas de
frecuencia media. También se evalúa el impacto del ancho de banda del canal y la
configuración de formación de haz. Además, se simula un escenario de uso que representa
un evento lleno de gente bajo diferentes configuraciones de implementación para evaluar las
tasas de datos experimentadas para diferentes cargas de servicio.

2
Resum

La nova interfície aèria de ràdio (NR) especificada per als sistemes 5G és capaç d’operar
en bandes de freqüència en el rang de 24,25-52,6 GHz, conegudes com a freqüències d’ona
mil·limètrica (mmWave), on es poden utilitzar quantitats importants d’espectre per suportar
un creixement constant de tràfic mòbil i proporcionar taxes de dades en el rang de gigabits
per segon.

En aquest context, aquesta tesi estudia el rendiment en termes de capacitat i cobertura

de diferents opcions de desplegament de cel·les macro i small per a l’explotació de la banda
de 26 GHz. L’estudi es realitza mitjançant l’eina de planificació de la ràdio ATOLL. Penseu en
una zona geogràfica de la ciutat de Barcelona per a la qual es pugui caracteritzar
detalladament incloent el modelat d’edificis a una resolució d’un metre, un conjunt de llocs de
cel·les representatius a diferents llocs (per exemple, llocs de cel·les macro en zones obertes,
llocs de cel·les small a nivell de carrer, etc.) s'han definit i avaluat. Els resultats per a 26 GHz
es comparen amb configuracions que utilitzen bandes de freqüència mitjana. També s’avalua
l’impacte de l’amplada de banda del canal i la configuració de la formació de feix. A més, es
simula un escenari d’ús que representa un esdeveniment multitudinari en diferents
configuracions de desplegament per avaluar les taxes de dades experimentades per diferents
càrregues de servei.

3
Dedication

I devote this work to my lovely family from whom I have obtained the support and strength
for every step during this process and all their love during my entire life. To them, I owe the
confidence and the opportunity to open my wings to discover other world and achieve new
goals.

4
Acknowledgements

I would like to thank my thesis advisors Anna Umbert and Ramon Ferrús for their guidance
and constant support during this time, as well as for providing me with necessary information
and tools for the current study. I would also like to express my gratitude to my country and the
outstanding scholarships program led by the National Secretariat of Higher Education,
Science, Technology and Innovation (SENESCYT) of Ecuador, which gave me the opportunity
to study this master program. Finally, I must thank to my family for providing me with
continuous encouragement and support. This accomplishment would not have been possible
without them.

Thank you,

Layla Guerrero

5
Revision history and approval record
Revision Date Purpose
0 25/10/19 Document creation
1 30/10/19 Document revision
2 15/11/19 Document correction
3 20/11/19 Document revision
4 04/12/19 Document correction
5 09/12/19 Document revision
6 08/01/20 Document correction
7 10/01/20 Document revision
8 23/01/20 Document correction
9 27/01/20 Document revision

Written by: Reviewed and approved by:

Date 27/01/20 Date 28/01/20

Name Layla Guerrero Name Anna Umbert / Ramon Ferrús

Position Project Author Position Project Supervisor

6
Table of contents

Abstract ........................................................................................................................... 1
Resumen......................................................................................................................... 2
Resum............................................................................................................................. 3
Acknowledgements ......................................................................................................... 5
Revision history and approval record .............................................................................. 6
Table of contents............................................................................................................. 7
List of Figures.................................................................................................................. 9
List of Tables ................................................................................................................. 11
1. Introduction ........................................................................................................... 12
1.1. Context and motivation .................................................................................. 12
1.2. Objectives ...................................................................................................... 12
1.3. Document structure........................................................................................ 13
2. 5G fundamentals for mmWave planning ............................................................... 14
2.1. Justification of the necessity of 5G in mmWave band .................................... 14
2.2. Technical requirements and frequency range of 5G NR in mmWave band .... 15
2.3. NR physical layer ........................................................................................... 16
2.4. Multi-antenna transmission ............................................................................ 17
2.5. Throughput .................................................................................................... 18
3. ATOLL radio planning tool .................................................................................... 19
3.1. Configuration of network equipment ............................................................... 19
3.1.1. Beamforming features............................................................................. 22
3.1.2. MIMO features ........................................................................................ 23
3.2. Geographic data imported to Atoll .................................................................. 23
3.2.1. Digital Terrain Model (DTM) .................................................................... 23
3.2.2. Clutter Classes ....................................................................................... 23
3.2.3. Clutter Heights ........................................................................................ 24
3.2.4. Vector Layers ......................................................................................... 25
3.3. Propagation models for mmWave planning .................................................... 25
3.3.1. 3GPP 38.900 propagation model ............................................................ 25
3.3.2. mmWave Aster propagation model ......................................................... 28
3.3.3. ATOLL model for clutter-based indoor losses ......................................... 33
3.4. Study 5G network coverage ........................................................................... 33
3.4.1. Downlink Coverage Prediction ................................................................ 33
3.4.2. Downlink Quality Prediction .................................................................... 34
3.5. Studying 5G Network Capacity ...................................................................... 34
3.5.1. Modelling Services .................................................................................. 34
3.5.2. Modelling User Profiles ........................................................................... 35
3.5.3. Modelling Environments.......................................................................... 35
3.5.4. Simulations ............................................................................................. 35
4. Analysis of mmWave deployment options ............................................................. 37
4.1. Geographical area of study and data ............................................................. 37
4.2. Configuration, validation and comparison of the propagation models ............. 38
4.3. Initial Parameters in mmWave propagation models ....................................... 41
4.4. 5G macro cell deployments............................................................................ 42
4.4.1. Determination of representative cell site locations and configuration ...... 42
4.4.2. Coverage analysis .................................................................................. 46
4.4.3. Capacity analysis .................................................................................... 51

7
 4.5. 5G small cell deployments ............................................................................. 54
4.5.1. Determination of representative cell site locations and configuration ...... 54
4.5.2. Coverage analysis .................................................................................. 57
4.5.3. Capacity analysis .................................................................................... 59
4.6. Impact of increasing channel bandwidth to 400 MHz ..................................... 61
4.7. Impact of beamforming configuration ............................................................. 62
5. Network Deployment............................................................................................. 68
5.1. Determination and modelling of network services and users .......................... 68
5.2. Creation of the traffic map. ............................................................................. 69
5.3. Justification for network deployments ............................................................ 70
5.3.1. Site locations for Macro cell deployments ............................................... 70
5.3.2. Site locations for Small cell deployments ................................................ 72
5.4. Simulation of network deployments................................................................ 74
5.4.1. Analysis with variations of parameters .................................................... 78
6. Conclusions and future development .................................................................... 81
Bibliography .................................................................................................................. 85
Abbreviations and acronyms ......................................................................................... 86

8
List of Figures

Figure 1. Current 2G, 3G, and 4G spectrum and bandwidth allocations [3]. ................. 14
Figure 2. Bearer selection threshold according to the C/(I+N)....................................... 21
Figure 3. Height and distance parameters considered in the 38.900 model [10]. .......... 27
Figure 4. Vertical and Horizontal Components of Global Attenuation [15]. .................... 28
Figure 5. Enabling Indoor calculation without and with Indoor Losses applied. ............. 32
Figure 6. Geographic data: (a) Digital Terrain Model data, (b) Clutter classes, (c) Clutter
height and (d) Vector layers ................................................................................................ 37
Figure 7. Display of Point Analysis tool. ........................................................................ 38
Figure 8. Geographical Profile of the mmWave Aster Propagation Model available in Point
Analysis tool. ....................................................................................................................... 39
Figure 9. The mmWave Aster Propagation Model configured without indoor losses applied.
........................................................................................................................................... 39
Figure 10. Link Budget with mmWave Aster Propagation Model: (a) Apply indoor losses
set to “NO”, (b) Apply indoor losses set to “NO” and Clutter-based Indoor Loss from Atoll
applied, (c) Apply indoor losses set to “YES”. ..................................................................... 39
Figure 11. Configuration in Point Analysis to apply Clutter-based Indoor Loss from Atoll.
........................................................................................................................................... 40
Figure 12. The mmWave Aster Propagation Model configured with indoor losses applied.
........................................................................................................................................... 40
Figure 13. Geographical Profile of the 3GPP 38.900 Propagation Model available in Point
Analysis tool. ....................................................................................................................... 41
Figure 14. Link Budget with 3GPP 38.900 Propagation Model: (a) without indoor losses,
(b) with indoor losses. ......................................................................................................... 41
Figure 15. Penetration loss and linear loss values in mmWave Aster model................. 42
Figure 16. Locations of real telecommunication base stations ...................................... 43
Figure 17. Selected sites for 5G macro cell characterization. ....................................... 44
Figure 18. Transmitter parameters for sites A, B and C (except the azimuth value). ..... 45
Figure 19. Representation of location of the sites and transmitters in Atoll. .................. 46
Figure 20. Downlink Coverage SS-RSRP predictions per each site and its computation
zone. ................................................................................................................................... 47
Figure 21. Sites using mmWave Aster propagation model with 26 GHz and using Aster
propagation model with 3.5 GHz. ........................................................................................ 47
Figure 22. Configuration to Indoor calculation only needed to perform Multi-storey
prediction ............................................................................................................................ 48
Figure 23. Downlink coverage using the mmWave Aster Propagation Model with 20m
receiver height. ................................................................................................................... 49
Figure 24. Downlink coverage using the mmWave Aster Propagation Model and receiver
height at 1.5m, 5m, 10m, 20m and 30m. ............................................................................. 49
Figure 25. Downlink coverage using the Aster Propagation Model and receiver height at
1.5m, 5m, 10m, 20m and 30m. ........................................................................................... 50
Figure 26. Downlink Capacity predictions per each site and its computation zone. ....... 52
Figure 27. Distribution of peak cell capacity for the different sites and frequency bands.
........................................................................................................................................... 53
Figure 28. Selected sites for 5G small cell characterization. ......................................... 55
Figure 29. Transmitter parameters for sites X, Y and Z. ............................................... 56
Figure 30. Representation of location of the sites and transmitters in Atoll. .................. 56

9
 Figure 31. Downlink Coverage SS-RSRP predictions per each site and its computation
zone. ................................................................................................................................... 57
Figure 32. Sites using Aster propagation model with 3.5 GHz and using mmWave Aster
propagation model with 26 GHz. ......................................................................................... 58
Figure 33. Downlink Capacity predictions per each site and its computation zone. ....... 60
Figure 34. Downlink capacity for the different sites and frequency bands. .................... 60
Figure 35. Distribution of peak cell capacity for the different sites in 26 GHz. ............... 61
Figure 36. 3GPP Antenna Properties. .......................................................................... 63
Figure 37. Downlink Capacity predictions using different beamforming configurations in
Small Cell............................................................................................................................ 64
Figure 38. Downlink Capacity using different beamforming configurations in Small Cell.
........................................................................................................................................... 65
Figure 39. Downlink Capacity predictions using different beamforming configurations in
Macro Cell........................................................................................................................... 66
Figure 40. Downlink Capacity using different beamforming configurations in Macro Cell.
........................................................................................................................................... 67
Figure 41. Properties of the new Required Service. ...................................................... 68
Figure 42. Properties of the new user. .......................................................................... 69
Figure 43. Properties of the new Environment. ............................................................. 69
Figure 44. Traffic zone.................................................................................................. 70
Figure 45. Quality predictions and interference results. ................................................ 72
Figure 46. Quality predictions and interference results with two small cells. ................. 73
Figure 47. Selection of traffic map in a new Simulation. ................................................ 74
Figure 48. Traffic simulation in Macro Cell environment. .............................................. 76
Figure 49. Traffic simulation in Small Cell network........................................................ 77
Figure 50. Experienced Uplink bit rate from simulations results. ................................... 78
Figure 51. Experienced Downlink bit rate from simulations results. .............................. 79
Figure 52. Active users from simulations results. .......................................................... 79
Figure 53. Rejected users from simulations results....................................................... 80

10
List of Tables

Table 1. NR Operating Bands for Frequency Range 2 (FR2) [4]. .................................. 15

Table 2. Minimum Technical Performance Requirements for IMT-2020 [5]. .................. 15
Table 3. Physical Signals. ............................................................................................. 17
Table 4. Upper limits for Physical Layer Throughput per MIMO layer (Mbps) ................ 18
Table 5. Transmitter Total Losses ................................................................................. 19
Table 6. Radio Bearer for 5G NR Radio Equipment ...................................................... 20
Table 7. DTM parameters of accuracy [12] ................................................................... 23
Table 8. Clutter Classes parameters of accuracy [12]. .................................................. 23
Table 9. Clutter Classes [12]. ........................................................................................ 24
Table 10. Clutter Heights parameters of accuracy [12].................................................. 25
Table 11. Vector Layers [12]. ........................................................................................ 25
Table 12. Propagation scenarios and distances. ........................................................... 26
Table 13. Path loss for each propagation scenario. ...................................................... 27
Table 14. Calculation of vertical and horizontal components. ........................................ 29
Table 15. Deterministic propagation classes available in Aster [14]. ............................. 30
Table 16. Statistical propagation classes available in Aster [14].................................... 30
Table 17. Mapping Clutter Classes in Propagation Classes - Low Resolution case. ..... 31
Table 18. Mapping Clutter Classes in Propagation Classes - High Resolution case...... 31
Table 19. Coverage Prediction in Atoll. ......................................................................... 33
Table 20. Material penetration losses in 3GPP TR 38.900 model. ................................ 41
Table 21. 5G NR Macro Cell Parameters ...................................................................... 44
Table 22. Beamforming Antenna Parameters in Macro Cell .......................................... 45
Table 23. Downlink coverage SS-RSRP ....................................................................... 48
Table 24. Outdoor and Indoor Coverage ....................................................................... 51
Table 25. Downlink Capacity per covered area ............................................................. 53
Table 26. 5G NR Small Cell Parameters ....................................................................... 55
Table 27. Beamforming Antenna Parameters in Small Cell ........................................... 56
Table 28. Downlink coverage SS-RSRP ....................................................................... 58
Table 29. Outdoor and Indoor Coverage ....................................................................... 59
Table 30. Downlink Capacity per covered area with carrier of 50MHz. .......................... 61
Table 31. Downlink Capacity per covered area using 26 Ghz band with carriers of 400MHz
and 50MHz. ........................................................................................................................ 62
Table 32. Beamforming Antenna Parameters ............................................................... 63
Table 33. Downlink Capacity per covered area using different beamforming configurations
in Small Cell. ....................................................................................................................... 65
Table 34. Downlink Capacity per covered area using different beamforming configurations
in Macro Cell. ...................................................................................................................... 67
Table 35. Quality Prediction Results in Macro Cell. ....................................................... 72
Table 36. Quality Prediction Results in Small Cell with 2 sites. ..................................... 74
Table 37. Quality Prediction Results in Small Cell with 3 sites. ..................................... 74
Table 38. Simulations results in Macro and Small Cell deployment............................... 75

11
1. Introduction

1.1. Context and motivation

Nowadays, having a wireless network connection is one of the most common and
necessary things in our daily life. During decades, we have experimented different generations
on mobile communications, from analog voice to the current 4G LTE Generation in which there
has been an improvement in the mobile broadband due to the technology development, the
benefits for the users can be translated as a fast wireless internet access experience.

As the technology in telecommunications are always evolving, the International

Telecommunication Union Radiocommunications Sector (ITU-R) started to develop “IMT for
2020 and beyond” in early 2012, and also, due to the dramatic growth of mobile data services
driven by wireless Internet and smart devices. International Telecommunication Union (ITU)
and its partners are working in the investigation of 5G for the next generation of terrestrial
mobile telecommunications. They have recognized the relationship between International
Mobile Telecommunications (IMT) and 5G.

In 2015, ITU-R has finalized its “Vision” of the 5G mobile broadband connected society.
This view of the horizon for the future of mobile technology was instrumental in setting the
agenda for the last World Radiocommunication Conference 2019 (Oct 28th-Nov 22nd), where
deliberations on additional spectrum were taking place in support of the future growth of IMT.
Future mobile networks will face great challenges, including higher capacity, higher
performance, lower power consumption, higher spectrum efficiency, more spectrum resource,
and lower cost [1].

As the demand for capacity in mobile broadband communications increases dramatically

every year, wireless carriers must be prepared to support up to a thousand-fold increase in
total mobile traffic, requiring researchers to seek greater capacity and to find new wireless
spectrum beyond the current bands used in existing standards and telecommunication
technologies.

To face the increasing of demand for capacity in mobile broadband communications,

researchers are required to find new wireless spectrum, with higher bandwidth available,
beyond the current bands used in existing standards and telecommunication technologies. In
that sense 5G new radio (NR) access, will include new bands in the range of 24.25 – 52.6
GHz, known as millimeter wave (mmWave) frequencies.

1.2. Objectives

This thesis aims to test the behaviour of different 5G scenarios, working in mmWave bands,
with different configurations using a radio planning tool, ATOLL. The following objectives were
defined:
• To analyse and get familiar with the 5G model supported in the ATOLL tool for the
planning and performance evaluation of 5G networks.
• To analyse and get familiar with the use of detailed geographic data including digital
terrain heights, clutter classes and buildings. This includes also the determination of
the suitable configuration to support the planning studies, including the selection of the
propagation model, for mmWave bands.

12
 • To identify and characterize a number of representative configurations for both macro
and small cell deployments, detailing site, transmitter and cell characteristics.
• To study the performance of the defined representative configurations, including
capacity and coverage predictions.
• Based on the analysis of the representative configurations, to carry out the planning of
an illustrative 5G scenario considering different deployment options and 5G service
requirements.

1.3. Document structure

The reminder of the document is organized in five chapters.

Chapter 2 describes the 5G fundamentals for mmWave planning, it let to the readers to
know about the key theoretical concepts needed to work in this topic.

Chapter 3 is the compressed explanation of some tools and characteristics of ATOLL that
are applied for the study and analysis in the current work.

Chapter 4 describes the analysis of mmWave deployment options, from the geographical
site location and characteristic to the final sites configurations that will be used for the
deployments.

Chapter 5 is about the Network Deployment, from the determination and modelled of the
network services and users to the analysis of the different deployments options in Macro and
Small Cell.

Chapter 6 presents the conclusions for all the analysis and studies achieved in the current
work. Also, it includes some suggestions for future work.

13
2. 5G fundamentals for mmWave planning

This chapter explains concepts of some key features that are necessary to develop the 5G
mmWave network planning.

First, it is important to remark that 5G wireless access is envisioned to enable a networked

society, there are three major challenges:
• A massive growth in the number of connected devices.
• A massive growth in traffic volume.
• A wide range of applications with diverse requirements and characteristics.

The current cellular systems operate below 6 GHz. Third generation partnership project
(3GPP) is currently developing a global standard for new radio access technology, 5G new
radio (NR), which will operate in frequencies from below 1 GHz up to 100 GHz. 5G NR shall
unleash new frequencies and new functionalities to support ever-growing human-centric and
machine-centric applications [2].

2.1. Justification of the necessity of 5G in mmWave band

Nowadays, mobile broadband networks are facing the growing consumer data rate
demands and the exponential increasing of traffic volumes. Also, wireless service providers
are limited to carrier frequency spectrum ranging between 700 MHz and 2.6 GHz. In Figure
1, the global spectrum bandwidth allocation for all cellular technologies does not exceed 780
MHz, where each major wireless provider has approximately 200 MHz across all of the
different cellular bands of spectrum available to them.

Figure 1. Current 2G, 3G, and 4G spectrum and bandwidth allocations [3].

14
 Thus, an efficient radio access technology combined with more spectrum availability is
essential to achieve the ongoing demands faced by wireless carriers.

In the development and implementation of the new fifth generation (5G), there will be main
differences compared to 4G. It is required the use of much greater spectrum allocations at
untapped mmWave frequency bands, highly directional beamforming antennas, longer battery
life, higher bit rates in larger portions of the coverage area, lower infrastructure costs, and
higher aggregate capacity for many simultaneous users in both licensed and unlicensed
spectrum (e.g. the convergence of Wi-Fi and cellular). The networks of 5G mmWave wireless
connections should allow fast deployment and connectivity with cooperation between base
stations.

2.2. Technical requirements and frequency range of 5G NR in mmWave band

5G NR can be deployed in different frequency bands, which were divided into two
frequency ranges by [4], because of different RF requirements (e.g. maximum transmission
power):
• Frequency range 1 (FR1) that includes all existing and new bands below 6 GHz.
• Frequency range 2 (FR2) that includes new bands in the range of 24.25 – 52.6 GHz.

Also, 3GPP defined the operating bands, corresponding to different frequency ranges for
downlink and uplink. Table 1 shows the details of the FR2, also known as millimeter wave
(mmWave) frequencies, which is of interest for the present work.

Table 1. NR Operating Bands for Frequency Range 2 (FR2) [4].

NR Operating Band Uplink Range (MHz) Downlink Range (MHz) Duplex Mode
n257 26500 – 29500 26500 – 29500 TDD
n258 24250 – 27500 24250 – 27500 TDD
n260 37000 – 40000 37000 – 40000 TDD
n261 27500 – 28350 27500 – 28350 TDD

Table 2 summarizes the minimum requirements related to technical performance for IMT-
2020 radio interfaces, which are based on a set of capabilities needed to support the 5G use
cases and usages scenarios.

Table 2. Minimum Technical Performance Requirements for IMT-2020 [5].

Parameter Minimum Technical Performance Requirement

Peak data rate Downlink: 20 Gbps
Uplink: 10 Gbps
Peak spectral efficiency Downlink: 30 bit/s/Hz
Uplink: 15 bit/s/Hz
User-experienced data rate Downlink: 100 Mbps
Uplink: 50 Mbps
Area traffic capacity 10 Mbps/m2 (indoor hotspot for eMBB)
User plane latency 4 ms for eMBB
1 ms for URLLC
Control plane latency 20 ms

15
 Connection density 1,000,000 devices per km2
Energy efficiency Efficient data transmission in a loaded case
Low energy consumption when there is no data
High sleep ratio
Long sleep duration
Reliability 1-10-5 success probability of transmitting a layer 2 PDU
of 32 bytes within 1 ms, at coverage
edge in Urban Macro for URLLC
Mobility 1.5 bit/s/Hz at 10 km/h for indoor hotspot eMBB
1.12 bit/s/Hz at 30 km/h for dense urban eMBB
0.8 bit/s/Hz at 120 km/h for rural eMBB
0.45 bit/s/Hz at 500 km/h for rural eMBB
Mobility interruption time 0 ms
Bandwidth At least 100 MHz and up to 1 GHz in higher frequency
bands. Scalable bandwidth shall be supported

2.3. NR physical layer

5G NR protocol architecture is separated into control plane and data plane. The physical
layer (PHY) is one of the layers in the user plane protocol stack. PHY layer handles
coding/decoding, modulation/demodulation, multi-antenna processing and mapping of signals
to physical time-frequency resources. The physical channels, time-frequency resources
carrying information from higher layers (above PHY), defined for NR are the following [6]:

• Physical Downlink Shared Channel (PDSCH): main physical channel used for unicast
data transmission. It also used for transmission of paging information and delivery of part
of the system information.
• Physical Broadcast Channel (PBCH): carries part of the system information, required
by the device to access the network.
• Physical Downlink Control Channel (PDCCH): used for downlink control information,
mainly scheduling decisions, required for reception of PDSCH and for scheduling grants
enabling transmission on the PUSCH.
• Physical Uplink Shared Channel (PUSCH): is the uplink counterpart to the PDSCH.
• Physical Uplink Control Channel (PUCCH): used by the device to send hybrid-ARQ
acknowledgements, to send channel-state reports and for requesting resources to transmit
uplink data upon.
• Physical Random-Access Channel (PRACH): used by the UE to request connection
setup referred to as random access.

The downlink and uplink transmission between gNB (the radio access network node) and
the UE work as follows [2]:

In the downlink, the UE monitors the PDCCH, typically once per slot. After detecting a
valid PDCCH, the UE receives one unit of data (transport block) on the PDSCH following the
scheduling decision of the gNB. Afterwards, the UE responds with a hybrid ARQ
acknowledgment indicating if data was successfully decoded or not.

In the uplink, the UE requests the gNB for physical time-frequency resources to transmit
data (scheduling request) and this request is sent over the PUCCH. In response, the gNB

16
 sends a scheduling grant over the PDCCH, which gives permission to a UE to use certain
resources for transmission. Following the scheduling grant, the UE schedules its data
transmission over the PUSCH, data that is received by the gNB and sends hybrid ARQ
acknowledgement indicating if the data was decoded correctly or not.

On the other hand, the physical signals, Table 3, correspond to a set of time-frequency
resources used by the PHY layer but not to carry information. They are reference signals used
for purposes such as demodulation, channel estimation, synchronization and channel-state
information. In particular, the PSS and the SSS are signals used in the cell search process,
used by UE to find a new cell.

Table 3. Physical Signals.

Downlink Physical Signals Uplink Physical Signals
• Demodulation Reference Signal (DM-RS) • Demodulation Reference Signal (DM-RS)
• Phase Tracking Reference Signal (PT-RS) • Phase Tracking Reference Signal (PT-RS)
• Channel State Information Reference Signal• Sounding Reference Signal (SRS)
(CSI-RS)
• Primary Synchronization Signal (PSS)
• Secondary Synchronization Signal (SSS)

Also, 3GPP define some physical layer measurements. One of these measurements is
the SS reference signal received power (SS-RSRP), which is defined as the linear average
over the power contributions (in Watts) of the resource elements that carry secondary
synchronization signals (SSS) [7].

2.4. Multi-antenna transmission

Multi-antenna transmission is a feature of NR, it is achieved using massive Multiple Input

Multiple Output (MIMO) and beamforming, that improves the mobile communication system
performance.

Beamforming consists in the use of multiple antennas on the transmitter to provide

directivity, that is, to focus the transmitted power in a certain direction. Beamforming is just a
special implementation of MIMO. This directivity can increase the data rate and range due to
higher power reaching a specific location in space, and also reduces interference, which
improves the spectrum efficiency.

On the other hand, depending on the number of users, different multi-antenna schemes
can be considered: single-user MIMO (SU-MIMO) or spatial multiplexing and multi-user MIMO
(MU-MIMO).

In SU-MIMO, a single user is scheduled within a given time/frequency resource with

transmission of a large number of layers in parallel (up to 8 layers). This scheme is mainly
targeted at cell centered users with sufficient carrier to interference and noise C(I+N)
conditions, in order to improve the data rate.

In MU-MIMO, multiple users are scheduled simultaneously within the same time/frequency
resource but with a limited number of spatial layers per scheduled device, in this case, a
maximum of 4 layers. In contrast with SU-MIMO, where the spatial multiplexing gain is

17
confined to a single user, MU-MIMO allows multiple co-scheduled users to exploit this gain
among two or more UE [8].

In 5G NR, all these benefits of the MIMO schemes can be taken to a larger scale with the
use of hundreds to thousands of antennas at the base station. This is referred as Massive
MIMO, a larger number of antennas provides benefits in gain, signal to noise ratio, coverage,
capacity, data rates and decreased latency.

2.5. Throughput

The throughput in Frequency Range 2 (24.25 GHz to 52.60 GHz), for the subcarrier
spacing values allowed in it, is presented in Table 4. These results are assuming a single
MIMO layer, the normal cyclic prefix, a coding rate of 1 and zero overhead from the control
channels and reference signals, all Resources Elements transfer data [9].

Table 4. Upper limits for Physical Layer Throughput per MIMO layer (Mbps)
Base Station Channel Bandwidth (MHz)
Subcarrier
Metric
Spacing 50 100 200 400

QPSK 88.7 177.4 354.8

16QAM 177.4 354.8 709.6
60 kHz -
64QAM 266.1 532.2 1064.4
256QAM 354.8 709.6 1419.2
QPSK 86 177.4 354.8 709.6
16QAM 172 354.8 709.6 1419.2
120 kHz
64QAM 258 532.2 1064.4 2128.8
256QAM 344.1 709.6 1419.2 2838.4

18
3. ATOLL radio planning tool
ATOLL is a multi-technology wireless network design tool that offers unique capabilities of
using both predictions and live network data throughout the network planning and optimization
process. ATOLL supports the latest technology advances such as massive MIMO, 3D
beamforming, and mmWave propagation for the design of 5G networks [10].

In this chapter, some of the most relevant parameters related to the radio planning using
ATOLL are described. They will be used to develop the coverage and capacity analysis for
5G NR networks. The values in the configurations will be mentioned in the upcoming chapters
according to the necessities of each study.

3.1. Configuration of network equipment

The site is the geographical point and contains the base station and its transmitters.
Transmitters consist in all the equipment used to generate the radio waves in order to transmit
or receive data with the use of an antenna. Finally, a cell is the RF channel configured on a
transmitter [10].

The base station subsystem consists of three elements considered to calculate the
downlink and uplink losses:
• Tower-mounted amplifier (TMA): are used to reduce the composite noise figure of the
base station.
• Feeder cables: connect the TMA with the antenna.
• Transmitter equipment.
However, their values are set to zero for 5G network planning, because of its own concept.

The predefined equipment in the transmitter has values for total transmission and
reception losses, the default values are shown in Table 5.

Table 5. Transmitter Total Losses

Transmission Reception
Total losses (dB) 1.51 -2.09

Atoll has predefined station templates in which it is possible to modify the following cell,
however, the numerical values also will be showed when it is necessary:
• Carrier: The carrier of the cell in the frequency band. Indicated the carrier bandwidth.
• Max power (dBm): The cell’s maximum transmission power.
• SSS EPRE (dBm): The SSS energy per resource element. This value is used to calculate
the transmission power corresponding to the secondary synchronization signal (SSS).
• PSS EPRE offset / SSS (dB): The difference in the energy of a resource element
belonging to the PSS with respect to the energy of an SSS resource element. This value
is used to calculate the transmission power corresponding to the primary synchronization
signal (PSS).
• Layer: The network layer to which the cell belongs. It can be either macro layer (for macro
cells) or small layer (for small/micro cells).
• Cell type: This indicates whether the cell is configured as primary PCell, a secondary
SCell (UL), or a secondary SCell (DL).

19
• Min SS-RSRP (dBm): The minimum SS-RSRP required for a user to be connected to the
cell. The SS-RSRP is compared with this threshold to determine whether or not a user is
within the cell’s coverage or not.
• SS/PBCH numerology: The numerology used by the cell for SS/PBCH.
• SS/PBCH periodicity: The periodicity of the SS/PBCH bursts.
• SS/PBCH OFDM symbols: The OFDM symbols at which the SS/PBCH blocks start and
the number of SS/PBCH blocks per SS/PBCH burst.
• PDCCH overhead (OFDM symbols): The number of OFDM symbols per subframe
assigned to the physical downlink control channel (PDCCH).
• Traffic numerology: The numerology used by the cell for traffic channels (PDCCH,
PDSCH and PUSCH).
• TDD DL OFDM symbols (%): The percentage of downlink OFDM symbols out of the total
number of OFDM symbols per frame.
• Radio equipment: cell’s radio equipment.
• Scheduler: The scheduler used by the cell for bearer selection and resource allocation.
• Diversity support (DL): The type of antenna diversity technique supported by the cell in
downlink.
• Diversity support (UL): The type of antenna diversity technique supported by the cell in
uplink.
• Number of MU-MIMO users (DL): The average number of MU-MIMO users that share
the same resources on the downlink.
• Number of MU-MIMO users (UL): The average number of MU-MIMO users that share
the same resources on the uplink.
• Traffic load (DL) (%): The downlink traffic load percentage.
• Traffic load (UL) (%): The uplink traffic load percentage.
• Max traffic load (DL) (%): The downlink traffic load not to be exceeded.
• Max traffic load (UL) (%): The uplink traffic load not to be exceeded.
• UL noise rise (dB): The uplink noise rise in dB.
• Fractional power control factor: This factor is used for path loss compensation when
performing fractional power control on the uplink. For example, if this factor is set to 0.8,
only 80% of the actual path loss will be considered when estimating the received power.
• Number of required PRACH RSI: The number of required PRACH RSIs for this cell. The
number or PRACH RSIs needed for any cell depends on the used PRACH preamble
format and the cell size.
• Max number of 4G/5G neighbors: The maximum number of 5G NR neighbors that the
cell can have.
• Max number of inter-technology neighbors: The maximum number of other technology
neighbors that the cell can have.

The 5G NR Radio Equipment defines the radio bearers, Table 6, and bearer selection
threshold, Figure 2, for the transmitter and the user equipment [11].

Table 6. Radio Bearer for 5G NR Radio Equipment

Radio Bearer Channel Bearer Efficiency

Modulation
Index Coding Rate (bits/symbol)
19 QPSK 0.117188 0.2344
20 QPSK 0.188477 0.377
21 QPSK 0.300781 0.6016

20
 22 QPSK 0.438477 0.877
23 QPSK 0.587891 1.1758
24 16QAM 0.369141 1.4766
25 16QAM 0.423828 1.6953
26 16QAM 0.478516 1.9141
27 16QAM 0.540039 2.1602
28 16QAM 0.601563 2.4063
29 16QAM 0.642578 2.5703
30 64QAM 0.455078 2.7305
31 64QAM 0.504883 3.0293
32 64QAM 0.553711 3.3223
33 64QAM 0.601563 3.6094
34 64QAM 0.650391 3.9023
35 64QAM 0.702148 4.2129
36 64QAM 0.753906 4.5234
37 64QAM 0.802734 4.8164
38 64QAM 0.852539 5.1152
39 256QAM 0.666504 5.332
40 256QAM 0.694336 5.5547
41 256QAM 0.736328 5.8906
42 256QAM 0.77832 6.2266
43 256QAM 0.821289 6.5703
44 256QAM 0.864258 6.9141
45 256QAM 0.89502 7.1602
46 256QAM 0.925781 7.4063

Figure 2. Bearer selection threshold according to the C/(I+N)

21
 The antenna properties can be also selected. A physical antenna may have one or more
antenna models (patterns), corresponding to different electrical tilts. Some of the antenna
parameters are the mechanical azimuth, mechanical downtilt and the number of antennas for
transmission and reception. The values for the antenna configuration will be detailed in each
study in Chapter 4.

3.1.1. Beamforming features

In Atoll, 3D beamforming represents uniform planar array antennas with antenna elements
aligned horizontally and vertically across a two-dimensional plane. Atoll models 3D
beamforming by distinguishing between the antenna model and the beam pattern [10]:
• The 3D beamforming models represent the physical beamforming antenna equipment,
which produces multiple antenna patterns.
• The 3D beamforming patterns represent the beam patterns that are produced by the
beamforming model.

The parameters that can be configured that describe the physical characteristics of the 3D
beamforming antenna panel are:
• Frequency range: the range of frequencies in MHz within which the 3D beamforming
antenna is designed to operate.
• Vertical spacing: the distance between two vertical antenna elements in multiples of the
wavelength.
• Horizontal spacing: the distance between two horizontal antenna elements in multiples
of the wavelength.
• Columns (N): the number of columns of the antenna elements within the panel.
• Rows (M): the number of rows of the antenna elements within the panel.
• Transmission ports: the number of ports used in transmission.
• Reception ports: the number of ports used in reception.

On the other hand, each 3D beamforming pattern is represented for the following
parameters:
• Beam type: specifies whether the beam is a control channel beam, a traffic channel beam
or both. Control channel beams are used for the SS/PBCH block in 5G NR, while traffic
channel beams are used for the PDCCH and PDSCH.
• Beam index: specifies the unique index of the beam pattern within the 3D beamforming
model.
• Electrical azimuth (º): the azimuth towards which the beam pattern points.
• Electrical tilt (º): the tilt towards which the beam pattern points.
• Horizontal elements (m): specifies the number of horizontal antenna elements used to
form the beam pattern.
• Vertical elements (n): specifies the number of vertical antenna elements used to form the
beam pattern.
• Boresight gain (dB): specifies the gain of the beam in the direction defined by its azimuth
and tilt.
• Half-power beamwidth (º): displays the aperture of its horizontal pattern corresponding
to the pattern attenuation of 3 dB.

22
3.1.2. MIMO features

Atoll supports different MIMO configurations: transmit and receive diversity, single-user
MIMO (SU-MIMO) or spatial multiplexing and multi-user MIMO (MU-MIMO) or collaborative
MIMO.

During calculations, a user terminal with MIMO capabilities, connected to a cell that
supports SU-MIMO or MU-MIMO, will benefit from the MIMO gain or not depending if the
PDSCH or PUSCH C/(I+N) on the point where the user is located is higher o equal than the
threshold defined in the 5G NR radio equipment. This threshold is defined by default by Atoll
to be 14 dB. It means that only above this value, SU-MIMO and MU-MIMO can be used.

3.2. Geographic data imported to Atoll

The geographic data for this study has been taken from the webpage of Visicom, a data
provider of digital mapping products [12]. The coordinate system used is WGS 84 / UTM zone
31 N. In order to carry out the different calculations (path loss, coverage, signal level, etc),
four geographic data types are imported to Atoll:
o Digital Terrain Model (DTM).
o Clutter Classes.
o Clutter Heights.
o Vector Layers.

3.2.1. Digital Terrain Model (DTM)

The DTM contains the value of elevation above sea level. It is automatically taken into
account by the propagation model during computations. Table 7 shows the accuracy of the
DTM file used in the current study. Also, it is important to know that each image pixel stores
the value of the terrain elevation, with a range from 0 to 27 meters.

Table 7. DTM parameters of accuracy [12]

Parameters of accuracy Value

Resolution (cell size) 1m
Absolute planimetric accuracy (x,y) 2-3 m
Absolute altimetric accuracy (z) 2-3 m

3.2.2. Clutter Classes

The clutter class data file describes land cover or land use. Clutter classes are taken into
account by the propagation model during computations. Each pixel in the clutter class file
contains a code (from a maximum of 256 possible classes) which corresponds to a clutter
class or to a certain type of ground used or covered [10].
The height of the terrain is defined as an average height for each clutter class. Table 8
shows the accuracy of the Clutter Classes file from Visicom.

Table 8. Clutter Classes parameters of accuracy [12].

Parameters of accuracy Value
Resolution (cell size) 1m
Absolute planimetric accuracy (x,y) 2-3 m

23
 Table 9 shows the correspondences between codes and clutter class names (in the
Visicom Clutter Classes file).

Table 9. Clutter Classes [12].

Code Class Name Class Description

1 Open Open space outside the town without vegetation.
Forested lands with closed tree canopy. No distinction is made
2 Forest
between deciduous and coniferous.
3 Sea Sea and ocean.
4 Inland water Rivers, canals of more than 10 m width, lakes, reservoirs.
Houses in suburban environment. Suburban dwellings. Lots
may be as small as 30m by 30m, but are usually larger and
5 Residential
include vegetation cover. Individual houses are frequently
visible. Average height is below 15m.
Areas within urban perimeter. The mean urban should have
mean street density with no pattern, the major streets are
6 Mean urban
visible, the built-up features appear distinct from each other.
Some small vegetation could be included. Height is 15-30m.
Areas within urban perimeter. This includes dense urban areas
with dense development where built-up features do not appear
7 Dense urban
distinct from each other. It also includes built-up features of the
downtown district with heights 15-30m .
Groups of buildings, either parallel or not, that may be
8 Block buildings
separated by large green space. Height is above 30 m..
Areas including buildings with large footprints separated by
9 Industrial
streets (factories, shopping malls, storehouses etc.).
10 Villages Small built-up area in rural surrounding.
Open spaces inside the town: streets, avenues, vacant lots,
11 Open in urban
squares.
12 Parks in urban Park of less than 20m height trees.
13 Airport Territory of airport without buildings and runways.
14 Wetland Swampland.
Groups of houses or collective residential buildings in suburban
environment. Suburban density typically involves laid out street
15 Dense residential
patterns in which streets are visible. There is no open space
between constructions. Average height is below 15m.
16 Dense urban high Financial District. Heights are over 40 m.
17 Buildings Isolated cluster of high towers or skyscrapers higher than 40 m.
Forested lands with a low density of trees and without canopy
18 Sparse forest
tree.
19 Grass The territory covered by low grass.
20 Agricultural Lands used for agriculture.
21 Barren Areas covered by sparse, stunted vegetation.

3.2.3. Clutter Heights

Clutter height maps describe the altitude of a clutter over the DTM, in the file, a height is
given for each point on the map (pixel). Clutter height files allow higher accuracy because they
allow more than one height per clutter class [10].

24
 When the clutter altitude is defined in both clutter class map and clutter height map, the
clutter altitude is taken from the clutter height map. The parameters of accuracy, in the clutter
height file from Visicom, are shown in Table 10. This file includes buildings and vegetated
areas.
Table 10. Clutter Heights parameters of accuracy [12].

Parameters of accuracy Value

Resolution (cell size) 1m
Absolute planimetric accuracy 2-3 m
Accuracy of building heights 2-3 m
Minimal mapping unit for buildings 16 m2
Minimal recognizable height for buildings 3m
Accuracy of other obstacles heights 5m
Minimal mapping unit for vegetation 50 m2
Minimal recognizable height for vegetation 5m

3.2.4. Vector Layers

Atoll supports vector objects such as contours, lines, and points. They are arranged in
vector layers and are used for display purposes, for example polygons that can be used as
filters, computation zones or focus zones. There are defined 8 linear classes in the current
vector file, Table 11.
Table 11. Vector Layers [12].
Vector Name Vector Description
Boundaries Administrative boundaries of populated places.
Buildings Buildings with heights.
Coastlines Coastline of ocean and seas.
Highways International roads.
Coastline of rivers and lakes. Rivers with less
Inland water
than 10 m width.
Major roads Regional roads.
Streets Town street axial lines.
Vegetation Vegetation with heights.

3.3. Propagation models for mmWave planning

In this subsection, the two propagation models supported in ATOLL to cope with mmWave
frequencies are described, namely the 3GPP 38.900 model and the mmWave Aster model.
These models are intended to be used in both outdoor to outdoor (O2O) and outdoor to indoor
(O2I) scenarios. For the last case, the modelling of the indoor penetration losses based on the
characterization of the clutter classes is also described.

3.3.1. 3GPP 38.900 propagation model

The Atoll 3GPP 38.900 is a propagation model for frequencies above 6 GHz up to 100
GHz. According to the technical report 3GPP TR 38.900 [13], the model is applicable to the
following type of scenarios:
o UMi (Street canyon, open area) with O2O and O2I: the BSs are mounted below rooftop
levels of surrounding buildings. UMi open area is intended to capture real-life

25
 scenarios such as a city or station square. The width of the typical open area is in the
order of 50 to 100 m. Example: Tx height: 10m, Rx height: 1.5-2.5 m, Intersite Distance
(ISD): 200m.
o UMa with O2O and O2I: the BSs are mounted above rooftop levels of surrounding
buildings. Example: Tx height: 25m, Rx height: 1.5-2.5 m, ISD: 500m.
o Indoor: This scenario is intended to capture various typical indoor deployment
scenarios, including office environments, and shopping malls. The BSs are mounted
at a height of 2-3 m either on the ceilings or walls. The shopping malls are often 1-5
stories high and may include an open area shared by several floors. The BSs are
mounted at a height of approximately 3 m on the walls or ceilings of the corridors and
shops. Example: Tx height: 2-3m, Rx height: 1.5m, area: 500 square meters.

The 3GPP 38.900 as implemented in ATOLL is a semi-deterministic propagation model in

the sense that it applies empirical formulas to calculate the path loss along transmitter-receiver
profiles for which the LoS/NLoS status is determined based on actual geo data (instead of
LoS/NLoS probabilities as suggested in the 3GPP TR 38.900) [10]. The ATOLL 3GPP 38.900
defines four propagation scenarios: Rural Macro, Urban Macro, Urban Micro and Indoor
Hotspot. Each propagation scenario, called configuration in Atoll, uses defined sets of
empirical formulas for line-of-sight (LoS) and non-line-of-sight (NLoS) propagation. The
coefficients of the path loss terms in these formulas can be user-defined and automatically
calibrated using measurement data. The Atoll 3GPP 38.900 propagation model uses the
following path loss formula [11] common to all supported propagation scenarios:
𝑑
𝐾1 + 𝐾2 × log(𝑑) + 𝐾3 × log(𝑓) + 𝐾4 × log(𝑑𝐵𝑃 2 + (ℎ𝐵𝑆 − ℎ𝑈𝑇 )2 ) + 𝐾5 × log ( )
𝑑𝐵𝑃
+ 𝐾6 min(0.03 × ℎ1.72 , 10) × log(𝑑) + 𝑃𝐿
= 𝐾7 × min(0.044 × ℎ1.72 , 14.77) + 𝐾8 × log(ℎ) × 𝑑 + 𝐾9 × (ℎ𝑈𝑇 − 1. 5) + 𝐾10 × log(𝑊) + 𝐾11 × log(ℎ)
ℎ 2
+ 𝐾12 × log(ℎ𝐵𝑆 ) + 𝐾13 × (log(11.75 × ℎ𝑈𝑇 ))2 + 𝐾14 × log ( ) + 𝐾15 × log(ℎ𝐵𝑆 ) × log(𝑑) + 𝐿𝑜𝑥
ℎ𝐵𝑆

where, d is the distance in meters between the transmitter and receiver (selectable in the
model between the 2D (d2D) or 3D (d3D) distance1, see Figure 3), f is the operating frequency
in GHz, hBS is the transmitter antenna height in meters (between 10 and 150 m), hUT is the
receiver antenna height in meters (between 1 and 10 m), h is the average building height in
meters (height of the clutter) around the receiver (5 m by default), w is the street width (20 m
by default, with possible values between 5 and 50 m), Lox is the additional loss due to oxygen.
And, 𝑑𝐵𝑃 is a breakpoint distance calculated as indicated in Table 12.
Table 12. Propagation scenarios and distances.
Propagation
𝒅𝑩𝑷
Scenario
Rural Macro (RMa) 𝑓
𝑑𝐵𝑃 = 2 × 𝜋 × ℎ𝐵𝑆 × ℎ𝑈𝑇 ×
𝑐

Urban Macro (UMa) 𝑓

𝑑𝐵𝑃 = 4 × (ℎ𝑈𝑇 − 1) × (ℎ𝐵𝑆 − 1) ×
𝑐

Urban Micro (UMi) 𝑓

𝑑𝐵𝑃 = 4 × (ℎ𝑈𝑇 − 1) × (ℎ𝐵𝑆 − 1) ×
𝑐
Indoor Hotspot (InH) N/A
Note: 𝑐 = 3 × 108 𝑚/𝑠

1 In Advanced Properties, the selection of Use 3D distance is available.

26
 Figure 3. Height and distance parameters considered in the 38.900 model [10].

For each propagation scenario, the path loss is calculated using relevant components of
the generic formula mentioned before, as follows in Table 13:

Table 13. Path loss for each propagation scenario.

LoS - near transmitter, where 𝟏𝟎 𝒎 < 𝒅𝟐𝑫 < 𝒅𝑩𝑷
Rural Macro (RMa) 𝑃𝐿𝐿𝑜𝑆−𝑁𝑒𝑎𝑟
= 32.4 + 20 × log(𝑓) + min(0.03 × ℎ1.72 , 10) × log(𝑑) − min(0.044 × ℎ1.72 , 14.77)
+ 0.002 × log(ℎ) × 𝑑

Urban Macro (UMa) 𝑃𝐿𝐿𝑜𝑆−𝑁𝑒𝑎𝑟 = 32.4 + 20 × log(𝑑) + 20 × log(𝑓)

Urban Micro (UMi) 𝑃𝐿𝐿𝑜𝑆−𝑁𝑒𝑎𝑟 = 32.4 + 21 × log(𝑑) + 20 × log(𝑓)

LoS - far from the transmitter, where 𝒅𝑩𝑷 < 𝒅𝟐𝑫 < 𝒅𝑴𝑨𝑿
Rural Macro (RMa) 𝑃𝐿𝐿𝑜𝑆−𝐹𝑎𝑟
𝑑
= 32.4 + 20 × log(𝑑) + 20 × log(𝑓) + 40 × log ( )
𝑑𝐵𝑃
+ min(0.03 × ℎ1.72 , 10) × log(𝑑) − min(0.044 × ℎ1.72 , 14.77) + 0.002 × log(ℎ) × 𝑑

Urban Macro (UMa) 𝑃𝐿𝐿𝑜𝑆−𝐹𝑎𝑟 = 32.4 + 40 × log(𝑑) + 20 × log(𝑓) − 10 × log(𝑑𝐵𝑃 2 + (ℎ𝐵𝑆 − ℎ𝑈𝑇 )2 )

Urban Micro (UMi) 𝑃𝐿𝐿𝑜𝑆−𝐹𝑎𝑟 = 32.4 + 40 × log(𝑑) + 20 × log(𝑓) − 9.5 × log(𝑑𝐵𝑃 2 + (ℎ𝐵𝑆 − ℎ𝑈𝑇 )2 )

NLoS, where 𝟏𝟎 𝒎 < 𝒅𝟐𝑫 < 𝒅𝑴𝑨𝑿

Rural Macro (RMa) 𝑃𝐿𝑁𝐿𝑜𝑆
= 35.75 + 43.42 × log(𝑑) + 20 × log(𝑓) − 7.1 × log(ℎ) − 15.07 × log(ℎ𝐵𝑆 )
ℎ 2
− 10.24 × (log(11.75 × ℎ𝑈𝑇 ))2 + 3.7 × log ( ) × log(ℎ𝐵𝑆 )
ℎ𝐵𝑆
)
− 3.1 × log(ℎ𝐵𝑆 × log(𝑑)

Urban Macro (UMa) 𝑃𝐿𝑁𝐿𝑜𝑆 = 13.54 + 39.08 × log(𝑑) + 20 × log(𝑓) − 0.6 × (ℎ𝑈𝑇 − 1. 5)

Urban Micro (UMi) 𝑃𝐿𝑁𝐿𝑜𝑆 = 22.4 + 35.3 × log(𝑑) + 21.3 × log(𝑓) − 0.3 × (ℎ𝑈𝑇 − 1.5)

LoS and NLos, where 𝟏 𝒎 < 𝒅 < 𝒅𝑴𝑨𝑿

Indoor Hotspot 𝑃𝐿𝐿𝑜𝑆 = 32.4 + 17.3 × log(𝑑) + 20 × log(𝑓)
(InH) 𝑃𝐿𝑁𝐿𝑜𝑆 = 17.3 + 38.3 × log(𝑑) + 24.9 × log(𝑓)

27
 In above formulas, 𝑑𝑀𝐴𝑋 is the user-defined maximum distance beyond which path loss is
not calculated.

Below are some properties of the 3GPP 38.900 propagation model that will be configured
in the predictions:
• In the Parameters Tab, each of the previous scenarios can be selected.
• In the Clutter Tab, there is Clutter taken into account, in which you can select the
following modes of calculation:
▪ Receiver on top of clutter: Whether the receiver is considered always on top of
the clutter or inside clutter where the clutter height is greater that the receiver
height.
▪ Indoor calculations only: Whether path losses are to be calculated on indoor
locations only. This is useful when calculating outdoor predictions are not
convenient or needed, for example inside buildings. Therefore, in 3D coverage
prediction calculations performed by “Multistorey prediction” as it performs
successive 2D layer calculations at different heights.
• In Advanced Tab, there is the option Use 3D distance to select whether the distance has
to be calculated in 3D or 2D, where needed.

For the case of O2I scenarios, indoor penetration losses are not modelled as part of the
Atoll 3GPP 38.900 propagation model. Therefore, to take into account for such losses, the
Atoll modelling of indoor losses based on the characterization of the clutter classes shall be
used, i.e. Clutter-based Indoor Loss from Atoll. This modelling is described below in
subsection 3.3.3.

3.3.2. mmWave Aster propagation model

The mmWave Aster model is part of the Aster propagation model. The mmWave Aster
propagation model is designed for frequencies from 5 GHz up to 60 GHz. It is optimized for
urban areas but it can also be used for suburban and rural areas [14].
The models created by Aster have two major components:
o Vertical diffraction over rooftops: uses optimized radial extraction techniques. The
corresponding attenuation is 𝐿𝑣 , a hybrid diffraction attenuation based on the Deygout
knife-edge method and a modified Walfisch-Ikegami model.
o Horizontal diffraction/reflection below rooftops: uses high-speed ray-launching
techniques based on raster sampling of building angles. The corresponding
attenuation is 𝐿ℎ .

Figure 4. Vertical and Horizontal Components of Global Attenuation [15].

28
 Table 14. Calculation of vertical and horizontal components.
▪ 𝐿𝑐𝑙𝑢𝑡 is the attenuation due to the
clutter of the receiver
▪ 𝐺𝑎𝑛𝑡 is the Antenna Gain
▪ 𝐿𝑓𝑟𝑒𝑞 is the Frequency Correction in
LoS and NLoS
𝐿𝑙𝑜𝑠 + 𝐿𝑓𝑟𝑒𝑞 + 𝐿𝑒𝑎ℎ𝑙𝑜𝑠 ▪ 𝐿𝑙𝑜𝑠 is the LoS main attenuation
𝐿𝑣 = 𝐿𝑐𝑙𝑢𝑡 + 𝐺𝑎𝑛𝑡 + {
𝐿𝑛𝑙𝑜𝑠 + 𝐿𝑓𝑟𝑒𝑞 + 𝐿𝑒𝑎ℎ + 𝐿𝑟𝑚 + 𝐿𝑎𝑟 ▪ 𝐿𝑛𝑙𝑜𝑠 is the NLoS main attenuation
▪ 𝐿𝑒𝑎ℎ is the Effective Antenna Height
attenuation
▪ 𝐿𝑒𝑎ℎ𝑙𝑜𝑠 is the Effective Antenna Height
attenuation in case of LoS
▪ 𝐿𝑟𝑚 is the Roof to Mobile attenuation
▪ 𝐿𝑎𝑟 is the Antenna to Roof attenuation
▪ 𝑛 is the number of different free space
lines of paths causes by a diffraction
or a reflection
▪ 𝐿𝑙𝑜𝑠−𝑘 is the LoS attenuation on path
𝑛 𝑛−1 𝑛−1 number k
𝐿ℎ = 𝐿𝑐𝑙𝑢𝑡 + 𝐺𝑎𝑛𝑡 + ∑ 𝐿𝑙𝑜𝑠−𝑘 + ∑ 𝐷𝑖𝑓𝑓𝑘 + ∑ 𝑅𝑒𝑓𝑙𝑘 ▪ 𝐷𝑖𝑓𝑓𝑘 is the attenuation due to the
𝑘=1 𝑘=1 𝑘=1 diffraction between path number k and
path k+1
▪ 𝑅𝑒𝑓𝑙𝑘 is the attenuation due to the
reflection between path number k and
path k+1

In the Propagation Models folder, it is possible to select a mmWave Aster propagation

model configuration. The following configurations are predefined in Aster [14]:
• Default_Model: A general purpose configuration that works in all environments with
extremely high frequencies. Ray tracing is disabled by default but can be enabled.
• LOS_Model: A configuration optimized for line-of-sight communication with extremely
high frequencies. Ray tracing is disabled.
• Micro_28 Ghz: A general purpose configuration optimized for micro or small cells in
urban or suburban environments. Ray tracing is enabled.
• Macro_28 Ghz: A configuration optimized for macro cells in urban or suburban
environments. Ray tracing is disabled.

Aster requires at least the following geographical data:

• Digital Terrain Model (DTM) that represents the terrain altitude. The recommended
maximum pixel size is 100m.
• Above Surface Object Digital Model (ASODM), which represents the distribution and
composition of objects found above the ground. The ASODM model has two
components:
o ASODM types, which describe the objects or the terrain. These are derived
from the Clutter Classes layers in Atoll, but they can also be defined from
high-resolution clutter heights or vector files.

29
 o ASODM heights, which describe objects’ height above the ground. These are
typically retrieved from the Clutter Heights layers in Atoll, but they can also
be defined from the clutter classes default heights in Atoll or from vector files.

Since the Aster propagation model is specially tailored for urban environments, it is
important that the ASODM type and height are accurate enough to represent buildings and
vegetation that create obstructions, as well as streets and open areas where radio signals can
propagate freely.

Aster applies either of the following propagation classes depending on the precision in the
clutter class description [14]:
• Deterministic propagation classes: Aster uses deterministic propagation when the ASODM
description is precise enough to accurately model a terrain.
• Statistical propagation classes: Aster uses statistical propagation when the ASODM
description is rough (low resolution) and not well geo-referenced.

For Aster to work properly, values from the clutter class layer or from a vector file (terrain
categories) must be mapped to the appropriate Aster propagation classes. The lists of
available deterministic and statistical propagation classes are given in Table 15 and Table 16,
respectively.

Table 15. Deterministic propagation classes available in Aster [14].

Propagation Class Description Calculation
Open All open areas, streets, Open space, radio signals propagate freely.
squares, flat parks, etc.
Water Sea, rivers, lakes, etc.
Elevated Open Elevated areas. Filled obstruction area for radio signals.
Reception done on top.
Trees Trees, forests, etc. Deterministic vegetation. Radio signals can go
through with some loss, and are also diffracted
and
diffused.
Building Type 1 Three types of buildings Deterministic buildings that create an obstruction
Building Type 2 are defined to model on which radio signals reflect and diffract. Radio
Building Type 3 different propagation signal levels are not calculated in a direct way but
effects. interpolated from signal levels in surrounding
open
areas.
Bridge Bridges, highways, etc. Radio propagation is calculated on top of bridges.
Radio signals can propagate below bridges.

Table 16. Statistical propagation classes available in Aster [14].

Propagation Class Description Calculation
Mixed Vegetation Parks or areas with a mixture of Radio signals can go through with
trees and open areas. some loss and diffract vertically.
Sparse Buildup Sparse building areas with a mixture The clearance area is considered at
of buildings and open areas. the receiver location and reception is
Dense Buildup Dense building areas with a mixture done at ground level.
of buildings and open areas.

30
 Deterministic propagation and statistical propagation classes have to be mapped to the
clutter classes used in the ATOLL project. For this purpose, clutter class layers can be
selected on the Clutter tab of Aster properties. It is up to the user to decide on the best mapping
according to the meaning of a clutter class and to the corresponding terrain description
(accurate or rough).

When a low resolution umbrella clutter class layer is used, the propagation classes can be
any of the statistical classes or either the "Open" or "Water" deterministic classes for the clutter
classes which are mostly open areas. For example:

Table 17. Mapping Clutter Classes in Propagation Classes - Low Resolution case.
Clutter Class Propagation Class
Grass Open
Suburban Sparse Buildup
Dense Urban Dense Buildup

When a high resolution clutter class layer is used, choose the corresponding propagation
classes among the deterministic classes. For example:

Table 18. Mapping Clutter Classes in Propagation Classes - High Resolution case.
Clutter Class Propagation Class
Park Open
Mixed Vegetation
Vegetation Trees
Houses Building Type 1
Commercial Building Type 2
Skyscraper Building Type 3
Roads Bridge

It is important to know that when no clutter is defined, only the "No Data" line appears in
the column name of Parameters per clutter class from Clutter Tab; the propagation class
defined under Propagation Class is the one that will be used when no clutter is defined.

Aster enables you to consider indoor areas when you calculate a coverage prediction.
Figure 5 shows the options to be configured:

31
 Figure 5. Enabling Indoor calculation without and with Indoor Losses applied.

• When indoor calculation is disabled, the output calculation grid has undefined values for
all the pixels inside buildings and coverage predictions will not be available for indoor
locations
• When indoor calculation is enabled, two options are possible:
▪ With the option Apply indoor losses set to "No" (default option), Aster actually
interpolates the signal levels from surrounding streets and open areas, wherever
any of the "Building type" statistical propagation classes are defined.
▪ With the option Apply indoor losses set to “Yes”, this method will model indoor
loss from a first wall penetration loss with a variable linear loss based on the signal
propagation depth within the building. The indoor loss is based on the sum of two
losses which are defined for each "Building type" propagation class [14]:
• A penetration loss (for the initial penetration)
• A linear loss (dB/m) applied linearly from the building facade
Indoor loss is defined for each "Building type" propagation class ("Building type
1", "Building type 2", and "Building type 3").

Therefore, Aster has this option that allows us to define the indoor losses, so this model
can use its own modelling for O2I, or (never both options), the option of Atoll modelling for
indoor losses based on the characterization of the clutter classes shall be used, i.e. Clutter-
based Indoor Loss from Atoll, described in the next section.

32
3.3.3. ATOLL model for clutter-based indoor losses

To calculate indoor losses based on clutter data, a clutter class layer is available in Atoll
and clutter based indoor loss can be activated from the propagation model.
o To define clutter-based indoor loss in Atoll: In the Clutter Classes folder, select
Properties from the context menu, then, enter a Frequency (MHz) and its
corresponding Indoor Loss (dB),
o To activate clutter-based indoor loss from Atoll: In Predictions folder, select Properties
from a coverage prediction, then in the Conditions tab, select the Indoor coverage
check box and click OK.

3.4. Study 5G network coverage

5G NR coverage predictions in Atoll allows to analyze the effective signal levels, signal quality
and throughput, where each pixel of the map is considered as a non-interfering user, with a
defined service, mobility type and terminal. Table 19 summarizes the different coverage
predictions in Atoll.
Table 19. Coverage Prediction in Atoll.

Coverage Prediction Description

Network coverage Predicts the effective signal levels of
different type of 5G NR signals for the
downlink and the uplink. This prediction
includes SS-RSRP Level, PDSCH and
PUSCH signal level, best control channel
beam and best traffic channel beam.
Network quality Predicts the interference levels. This
prediction includes PDSCH and PUSCH
C/(I+N) level and SS-SIRN level.
Service areas Displays the 5G NR radio bearers based
on the C/(I+N) level for each pixel for the
downlink and the uplink. Also displays the
modulation schemes used.
Network capacity For 5G NR scenarios, it shows the peak
RLC allocated bandwidth throughput for
downlink and uplink.

In the current work, the network coverage predictions have been applied to study the signal
levels results, the network capacity predictions have been used to show the downlink and
uplink data rates that can be get without considering distribution of 5G users, and the network
quality predictions to select the best locations for the sites according to the interference levels
obtained.

3.4.1. Downlink Coverage Prediction

The downlink coverage predictions are based on the downlink traffic loads of the cells, and
the uplink coverage predictions are based on the uplink noise rise values. These parameters
can either be calculated by Atoll during the Monte Carlo simulations, or set manually by the
user for all the cells. Coverage predictions are generated using a bilinear interpolation method
from multi-resolution path loss matrices. These coverage predictions are all best server

33
coverage predictions, i.e., the coverage area of each cell comprises the pixels where the cell
is the best server [11]. The previous concepts in subsection 3.3 are applied here.

3.4.2. Downlink Quality Prediction

The interference received by any UE, served by a cell TXi (𝑖𝑐) from other cells TXj (𝑗𝑐) can
be defined as the signal levels received from interfering cells TX j (𝑗𝑐) depending on the overlap
that exists between the channels used by the cells TXi (𝑖𝑐) and TXj (𝑗𝑐), on the traffic loads of
the interfering cells TXj (𝑗𝑐), and whether the cells support inter-cell interference coordination
(ICIC) or not.

The carrier signal to interference and noise ratio is calculated in three steps. First Atoll
calculates the received signal level from the studied cell at the UE under study. Next, Atoll
calculates the interference received at the same studied UE from all the interfering.
Interference from each cell is weighted according to the co- and adjacent channel overlap
between the studied and the interfering cells, the traffic loads of the interfering cells, and the
probability of collision in case ICIC is used by the cells. Finally, Atoll takes the ratio of the
signal level and the sum of the total interference from other cells and the downlink noise [11].

3.5. Studying 5G Network Capacity

In Atoll, a simulation is based on a realistic distribution of users at a given point in time.

The distribution of users at a given moment is referred to as a snapshot. Based on this
snapshot, Atoll calculates various network parameters such as the downlink and uplink traffic
loads, the uplink noise rise, the user throughputs, and so on. Simulations are calculated in an
iterative fashion [10].

When several simulations are performed at the same time using the same traffic
information, the distribution of users will be different, according to a Poisson distribution.
Consequently, you can have variations in user distribution from one snapshot to another.

To create snapshots, services and users must be modelled. As well, certain traffic
information in the form of traffic maps must be provided. Once services and users have been
modelled and traffic maps have been created, you can make simulations of the network traffic.

We will describe some of the parameters of interest with the aim that in the next chapters
the reader can understand the chosen values.

3.5.1. Modelling Services

In 5G NR technology, the parameters of interest to configure are in two tabs.

The General Tab:

• Activity factor: The uplink and downlink activity factors are used to determine the
probability of activity for users accessing the service during Monte Carlo simulations.
For packet-switched services (data), this parameter is used when working with sector
traffic maps and user density traffic maps. For circuit-switched services (voice), the
parameter is taken into consideration with any traffic map.

34
 • Average requested throughput: You can enter the average requested throughput for
uplink and downlink. This throughput is the average throughput obtained by a user of
the service. How the average requested throughput is used in Atoll depends on the
type of service.

5G NR Tab:
This tab allows you to configure the quality of service, bearer, and throughput settings for
the service.
• Highest modulation and Lowest modulation: Specify the highest and lowest modulation
that the service can use in the uplink and downlink. These limits are considered for
determining the bearer.
• Highest coding rate and Lowest coding rate: Specify the highest and lowest channel
coding rates that the service can use in the uplink and downlink. These limits are
considered for determining the bearer.
• Min throughput demand and Max throughput demand: Enter the minimum and
maximum throughput that the service can demand in the uplink and downlink.

3.5.2. Modelling User Profiles

The parameters to create different profiles to model variations in user behavior for different
circumstances are the following:
• Service: Select a service from the list
• Calls/Hour: For packet-switched (data) services (max. bit rate), the Calls/Hour value is
defined as the number of sessions per hour. A session is defined by the volume
transferred in the uplink and downlink and not by the time.
• Duration: For circuit-switched services, enter the average duration of a call in seconds.
For packet-switched services, this field is left blank.
• UL Volume: For packet-switched services, enter the average uplink volume per
session in kilobytes.
• DL Volume: For packet-switched services, enter the average downlink volume per
session in kilobytes.

3.5.3. Modelling Environments

An environment class describes its environment using a list of user profiles, each with an
associated mobility type and a given density (i.e., the number of subscribers with the same
profile per km²). During Monte Carlo simulations, indoor losses defined per frequency per
clutter class will be added to the path losses of indoor mobiles.
• User Profile: Select a user profile.
• Mobility: Select a mobility type.
• Density (Subscribers/km²): Enter a density as a number of subscribers per square
kilometer for the combination of user profile and mobility type.

3.5.4. Simulations

Once you have modelled the network services and users and have created traffic maps,
you can create simulations. Depending on the radio technology, the simulation process
consists of several steps:

35
 1. Obtaining a realistic user distribution: Atoll generates a user distribution using a Monte
Carlo algorithm; this user distribution is based on the traffic database and traffic maps
and is weighted by a Poisson distribution between simulations of the same group.
Each user is assigned a service, a mobility type, and an activity status by random trial,
according to a probability law that uses the traffic database. The user activity status is
an important output of the random trial and has direct consequences on the next step
of the simulation and on the network interferences. A user can be either active or
inactive. Both active and inactive users consume radio resources and create
interference.
A shadowing error is randomly assigned to each user using the probability distribution
that describes the shadowing effect. Another random trial determines user positions in
their respective traffic zone (possibly according to the clutter weighting and the indoor
ratio per clutter class).
2. Technology selection: For each mobile generated at the beginning of the simulation,
Atoll searches for its serving cell in each possible technology. For multi-technology
mobiles, an active list of transmitters is generated, possibly using different
technologies. Then, retained transmitters are sorted according to the priorities of
technologies in the services.
3. Modelling network regulation mechanisms: Regulation mechanisms are modelled
according to the technology, or set of technologies, used by the network

Atoll selects the highest priority as defined in the service assigned to the mobile. Once
determined, the serving technology does not change for a given user distribution [10].

The map can be used to display the distribution of traffic that is generated by all simulations
according to different parameters. For example, you can display the traffic according to a
service, activity status or any output of the simulation.

The display of the traffic distribution can be set according to discrete values and the select
the value to be displayed. Or, you can select the display of the traffic distribution according to
value intervals, and then select the parameter and the value intervals that are to be displayed.
You can also define the colours of the icon and the icon itself.

36
4. Analysis of mmWave deployment options
In this chapter, representative sites have been determined with the aim that the cells cover
geographically different areas to analyze the behavior in an open area, in a residential area,
and in a dense residential area. We will study the results obtained in coverage analysis
prediction and in capacity prediction for macro and small cell deployments. The analysis will
be done for different sites and considering two cases for the band frequency: 26 GHz and 3.5
GHz in order to compare their behaviors.

4.1. Geographical area of study and data

The geographic data for this study has been taken from the webpage of Visicom, a data
provider of digital mapping products [12]. In order to carry out the different calculations (path
loss, coverage, signal level, etc), four geographic data types are imported to Atoll, they were
explained in subsection 3.2:
• Digital Terrain Model (DTM).
• Clutter Classes.
• Clutter Heights.
• Vector Layers.

The study takes place in a 4 km2 sample area of Ciutat Vella district, Barcelona (Spain),
the geographic area of study is shown in Figure 6 (b).

Figure 6. Geographic data: (a) Digital Terrain Model data, (b) Clutter classes, (c) Clutter
height and (d) Vector layers

37
4.2. Configuration, validation and comparison of the propagation models

To illustrate how indoor losses are accounted, the Point Analysis tool will be used with one
site located over a building in an urban zone, it is illustrated in Figure 7.

Figure 7. Display of Point Analysis tool.

First, the mmWave Aster Propagation Model, whose geographical profile can be
appreciated in Figure 8, is configured as follows:

• With option Apply indoor losses set to “NO” in the configurations within the model
properties, Figure 9:
o The results are shown in Figure 10 (a), where there are no indoor losses applied.
o Further, without modifying the mentioned configuration, we can apply the indoor losses
directly from the clutter classes values, using Clutter-based Indoor Loss from Atoll:
➢ The selection can be done from the configurations option in the Point Analysis tool,
see Figure 11. The results are shown in Figure 10 (b), where the indoor loss
parameter is equal to the clutter class’ value that was modified in the respective
Clutter Class (dense urban) to 20dB.

• With option Apply indoor losses set to “YES” in the configuration within the model
properties, Figure 12:
o The results are shown in Figure 10 (c), where there are indoor losses applied. This
value is 38 dB of indoor loss corresponding to the propagation classes from Aster
Model. Specifically, 35 dB of penetration loss and 3 dB of path loss in O2I.

38
Figure 8. Geographical Profile of the mmWave Aster Propagation Model available in Point
Analysis tool.

Figure 9. The mmWave Aster Propagation Model configured without indoor losses applied.

Figure 10. Link Budget with mmWave Aster Propagation Model: (a) Apply indoor losses set
to “NO”, (b) Apply indoor losses set to “NO” and Clutter-based Indoor Loss from Atoll
applied, (c) Apply indoor losses set to “YES”.

39
 Figure 11. Configuration in Point Analysis to apply Clutter-based Indoor Loss from Atoll.

Figure 12. The mmWave Aster Propagation Model configured with indoor losses applied.

Second, the 3GPP 38.900 propagation model, whose geographical profile can be
appreciated in Figure 13,
• As this model does not have options inside itself to configure indoor losses, it has been
calculated the values without considering the indoor losses, Figure 14 (a).
• Nevertheless, we can apply the indoor losses directly from the clutter classes values, using
Clutter-based Indoor Loss from Atoll:
▪ The selection can be done from the configurations option in the Point Analysis
tool, see Figure 11. The results are shown in Figure 14 (b) where the indoor
loss parameter is equal to the clutter class’ value that was modified in the
respective Clutter Class (dense urban) to 20 dB.

40
 Figure 13. Geographical Profile of the 3GPP 38.900 Propagation Model available in Point
Analysis tool.

Figure 14. Link Budget with 3GPP 38.900 Propagation Model: (a) without indoor losses, (b)
with indoor losses.

4.3. Initial Parameters in mmWave propagation models

Typical values of material penetration losses in 3GPP TR 38.900 model are given in Table
20. For 26 GHz, if we calculate every penetration loss, the result is that the range for
penetration loss is from 7.2 dB to 106 dB.

Table 20. Material penetration losses in 3GPP TR 38.900 model.

Material Penetration loss [dB]
Standard multi-pane glass Lglass = 2 + 0.2 f
IRR glass LIIRglass = 23 + 0.3 f
Concrete Lconcrete = 5 + 4 f
Wood Lwood = 4.85 + 0.12 f
Note: f is in GHz

In the following subsections, the mmWave Aster propagation model will be used in order
to be able to take into account the O2I losses, the default values given for the propagation
model will be applied. These are reproduced in Figure 15.

41
 Figure 15. Penetration loss and linear loss values in mmWave Aster model.

4.4. 5G macro cell deployments

In this subsection, we will study the results obtained in coverage analysis prediction and
in capacity prediction for macro cell deployments. The analysis will be done for different sites
and considering two cases for the band frequency: 26 GHz and 3.5 GHz in order to compare
their behaviors.

4.4.1. Determination of representative cell site locations and configuration

The site locations have been chosen with the aim that the cells cover geographically
different areas. These sites, in Figure 16, are located in an urban park zone, urban zone, and
dense urban zone where it will be possible to analyze the behavior in an open area, in a
residential area, and in a dense residential area, respectively. In addition, it is important to
remark that they are current locations of telecommunication base stations [16]. Figure 17
provides a description of the characteristics of the selected sites along with a Google Earth
view of the area surrounding the sites.

42
 Figure 16. Locations of real telecommunication base stations

Characteristics Google Earth View of the Area surrounding the site

Tower of 15m
located over a 20m
height building in
an urban park,
Site A –
pointing to open
Open Area
Macro area.

Coordinates taken
from site “Vodafone
España SA 35658”

Tower of 13m
located over a 27m
height building,
Site B – pointing to other
Residential
Area
buildings.
Macro
Coordinates taken
from site “Vodafone
España SA 75799”

43
 Tower of 1m
located over a 47m
height building in a
dense residential
Site C –
Dense zone.
Residential
Area
Macro
Coordinates taken
from site
“Telefónica Móviles
España SAU
0801258”
Figure 17. Selected sites for 5G macro cell characterization.

In Table 21 and Figure 18 the main characteristics of the 5G cells and transmitters are
shown respectively. Also, in Table 22, the Beamforming Antenna Parameters are detailed.

Table 21. 5G NR Macro Cell Parameters

Site A / Site B / Site C
Frequency band n78 (3.5 GHz) / n257 (26 GHz)
Carrier (MHz) 50
Max power (dBm) 50
SSS EPRE (dBm) 15
PSS EPRE offest / SSS (dB) 3
Layer Macro Layer
Cell type PCell
Min SS-RSRP (dBm) -140
SS/PBCH numerology 0 (15 kHz)
SS/PBCH periodicity (ms) 10
SS/PBCH OFDM symbols {4,8,16,20}+28n [Lmax=4]
PDCCH overhead (OFDM symbols) 1
Traffic numerology 2 (60 kHz Normal CP)
TDD DL OFDM symbols (%) 50
Radio equipment 5G NR Radio Equipment
Scheduler Proportional Fair
Diversity support (DL) Transmit Diversity; SU-MIMO; MU-MIMO
Diversity support (UL) Receive Diversity; SU-MIMO; MU-MIMO
Number of MU-MIMO users (DL) 1
Number of MU-MIMO users (UL) 1
Number of users (DL) 1
Number of users (UL) 1
Traffic load (DL) (%) 50
Traffic load (UL) (%) 100
Max traffic load (DL) (%) 100
Max Traffic Load (UL) (%) 100
UL noise rise (dB) 3
Fractional power control factor 1
Number of required PRACH RSI 10
Max number of 4G/5G neighbors 16
Max number of inter-technology neighbors 16

44
 Figure 18. Transmitter parameters for sites A, B and C (except the azimuth value).

The beamforming antennas parameters are shown in the following table:

Table 22. Beamforming Antenna Parameters in Macro Cell

Parameters Default Beamformer
Columns (N) 8
Rows (M) 8
Transmission ports 64
Reception ports 64
Number of beams 60
Azimuth range -45º to 45º
Tilt range -5º to 5º
Boresight gain (dB) 26.0618

The scenario is 5G NR Macro Cell. In Figure 19, we can appreciate the three sites in the
geographic area of study. In order to be able to compare the quantitative results, a
computation zone has been defined according to the possible better shape of the coverage
area for each site. This computation zone has a perimeter of 1.41x1.41 km and a total area of
1.97 km2.

45
 Figure 19. Representation of location of the sites and transmitters in Atoll.

4.4.2. Coverage analysis

The coverage analysis is based on the prediction of the SS-RSRP in the downlink. For
each site, the percentage of the computation zone covered with values of SS-RSRP above a
given threshold has been computed. A representation of the coverage plot for the 3 sites is
represented in Figure 20. In addition to the 26 GHz band, predictions considering the use of
the 3.5 GHz band have been also accounted for the same of comparison. In general terms, it
could be seen that the covered surface is higher in all the values for the case of 3.5 GHz. Also,
see in Figure 21.

26 GHz 3.5 GHz Legend

Site
A

46
Site
B

Site
C

Figure 20. Downlink Coverage SS-RSRP predictions per each site and its computation
zone.

Figure 21. Sites using mmWave Aster propagation model with 26 GHz and using Aster
propagation model with 3.5 GHz.

47
 Focusing now on the numerical values, Table 23 provides a summary of the amount of
covered area (in km2) per each site in both 3.5 GHz and 26 GHz for different SS-RSRP levels.
In Site A that corresponds to the site pointing to an open area, the amount of covered area
with low levels of signal is on average 2.22 times higher with the site in 3.5 GHz than that
corresponding to 26 GHz.
In Site B that corresponds to the site pointing to other buildings (urban area), the amount
of covered area with low levels of signal is on average 3.24 times higher with the site in 3.5
GHz than that corresponding to 26 GHz.
In Site C that corresponds to the site pointing to a dense residential zone, the amount of
covered area with low levels of signal is on average 6.84 times higher with the site in 3.5 GHz
than that corresponding to 26 GHz.
In the three sites, the amount of covered area with better level of signal (SS-RSRP ≥ -
100dBm) is around three times higher with the site in 3.5 GHz than that corresponding to 26
GHz.

In addition, in Table 23, it is possible to compare the results between the sites. For
example, in mmWave case of 26 GHz, it is noticeable that covered area by Site A is higher
than the ones in Site B or C. Therefore, it is important to analyze a case with better level of
signal, taking the case of SS-RSRP ≥ -100dBm, the Site A (open area) covers around the
triple of squared kilometers than Site B (urban area) and around 13 times more area than Site
C (dense residential zone) for both 3.5 GHz and 26 GHz.

Table 23. Downlink coverage SS-RSRP

Site A Site B Site C
26 GHz 3.5 GHz 26 GHz 3.5 GHz 26 GHz 3.5 GHz
SS-RSRP ≥ -140dBm 0.647 1.018 0.618 1.147 0.577 0.972
SS-RSRP ≥ -130dBm 0.191 0.449 0.106 0.378 0.054 0.182
SS-RSRP ≥ -120dBm 0.107 0.256 0.034 0.135 0.004 0.066
(km²)
SS-RSRP ≥ -110dBm 0.081 0.167 0.018 0.073 0.003 0.027
SS-RSRP ≥ -100dBm 0.035 0.096 0.011 0.030 0.002 0.008
SS-RSRP ≥ -90dBm 0.001 0.021 0.003 0.010 0.000 0.002

To analyze indoor coverage in different heights, it is configured with the Indoor calculation
enabled. For example, with the configuration in Figure 22, it will consider a user to be inside
a building if its height is higher than 2m. This means that signals are computed outdoor at
ground level (receiver height < 2m) and only indoor in storeys above 2m.

Figure 22. Configuration to Indoor calculation only needed to perform Multi-storey prediction

48
 Figure 23. Downlink coverage using the mmWave Aster Propagation Model with 20m
receiver height.

Figure 23 shows that the prediction is only performed indoor and not outdoor, since there
are no receivers outdoor at 20m. This also means that SS-RSRP coverage was reduced only
to those users that could actually be inside a building with a height at least 20m.

Figure 24 compares the Downlink coverage predictions for the mmWave Aster
Propagation Model with receivers at 1.5 m, 5 m, 10 m, 20 m and 30 m.

Figure 24. Downlink coverage using the mmWave Aster Propagation Model and receiver
height at 1.5m, 5m, 10m, 20m and 30m.

49
 In addition, the predictions using Aster Propagation Model in 3.5 GHz have been
developed, see Figure 25

Figure 25. Downlink coverage using the Aster Propagation Model and receiver height at
1.5m, 5m, 10m, 20m and 30m.

In Figure 24 and Figure 25, when the receiver height increases, the covered area in
squared kilometers is reduced for those buildings that have at least the same height of the
receiver height. This indicates that both Aster Propagation Models are very sensitive to
variations on the receiver height. Nevertheless, this sensitivity could be modified by changing
the 2m threshold to higher heights to consider indoor losses.

While previous results, in Table 23, estimate the total covered area, Table 24 shows
separately the coverage estimations for indoor and outdoor locations. In particular, for each
site, it provides: the amount of outdoor and indoor areas within the computation zone and the
percentage of those areas covered when using 26 and 3.5 GHz channels. For the indoor case,
five different receiver heights have been used in order to take into account, the different
altitudes where a user can be located.

Table 24 shows the results for the different receiver heights in the mmWave part. First, it
is important to remark that in all the cases, it has been used the total indoor area when the
receiver height is 1.5m because it was not possible to find the total area in other cases. As
conclusion, it is noticeable that the percentage of indoor coverage in mmWave is less than
2%. From the results, it can be noticed that in Site A, Site B and Site C, the percentage of
outdoor area covered in 3.5 GHz is around 1.5 times higher than that corresponding to 26
GHz. Also, Site A as being located in an open area has higher percentage of outdoor area
covered for both frequencies.
In Site A, the percentage of indoor area covered with the site in 3.5 GHz is 7 times higher
than that corresponding to 26 GHz.
In Site B, the percentage of indoor area covered with the site in 3.5 GHz is 12 times higher
than that corresponding to 26 GHz.

50
 In Site C, the percentage of indoor area covered with the site in 3.5 GHz is 5 times higher
than that corresponding to 26 GHz.
As we expected, in general there are huge differences in the indoor case between the
percentages of covered area using 26 GHz and 3.5 GHz, because the high frequencies are
affected with the obstacles.

Table 24. Outdoor and Indoor Coverage

Site A Site B Site C

Total 26 3.5 Total 26 3.5 Total 26 3.5
area GHz GHz area GHz GHz area GHz GHz
(km²) (%) (%) (km²) (%) (%) (km²) (%) (%)
Outdoor 1.018 73.320 93.041 1.139 64.673 91.283 1.037 69.464 91.295
Indoor (1.5 m) 0.957 0.958 6.666 0.836 1.369 16.568 0.938 1.601 7.298
Indoor (5 m) 0.957 0.103 6.972 0.836 0.302 16.267 0.938 0.050 7.635
Indoor (10 m) 0.957 0.212 9.524 0.836 0.819 16.169 0.938 0.064 9.687
Indoor (20 m) 0.957 0.070 12.394 0.836 0.491 10.156 0.938 0.335 12.436
Indoor (30 m) 0.957 0.032 6.561 0.836 0.197 5.635 0.938 0.749 7.065

4.4.3. Capacity analysis

The prediction calculated to evaluate the capacity is Peak RLC channel throughput (DL)
(kbps), which is the maximum RLC channel throughput attainable using the highest bearer
available at the subscriber location in the downlink. The configuration in the cells is set to 50%
for TDD DL OFDM Symbols and 50 % for Traffic load DL. The bandwidth of the carrier is set
to 50 MHz in all the cases, see Table 21.

26 GHz 3.5 GHz Legend

Site
A

51
Site
B

Site
C

Figure 26. Downlink Capacity predictions per each site and its computation zone.

A representation of the capacity plot for the 3 sites is represented in Figure 26. In addition
to the 26 GHz band, predictions considering the use of the 3.5 GHz band have been also
accounted for the same of comparison. It can be observed that the area where the service is
available is smaller than the area previously computed in the coverage analysis of the SS-
RSRP in Figure 20. This is because we are using the Peak RLC channel throughput (DL),
which is the maximum RLC channel throughput attainable using the highest bearer available
at the subscriber location in the downlink and even with beamforming there are some areas
on the map that are not covered or not assigned to a modulation scheme. It means that CINR
in those points of the map is too low that does not achieve the threshold requirements to
assign even the lowest bearer with the lowest modulation scheme.

Figure 27 provides the maximum data rate value, that is 270 Mbps, this value
approximates to the one in Table 4 from Chapter 2, with subcarrier spacing of 60 kHz, 64QAM
as maximum modulation and 50 MHz of bandwidth. In that table the value is 266.1 Mbps, but
it is considering a zero overhead. Also, Figure 27 shows graphically how the sites in 3.5 GHz
have higher covered area with all the capacities.

52
 Figure 27. Distribution of peak cell capacity for the different sites and frequency bands.

Table 25 provides the amount of covered area for different values of peak cell capacity.
From the results it can be observed that in Site A, Site B and Site C, the amount of area that
achieves a capacity higher to 0 Mbps is almost twice as high with the site in 3.5 GHz as with
the site in 26 GHz.
In Site A, the amount of area that achieves a capacity higher to 250 Mbps is 2.76 times
higher with the site in 3.5 GHz than that corresponding to 26 GHz. In Site B, the amount of
area that achieves a capacity higher to 250 Mbps is 3.2 times higher with the site in 3.5 GHz
than that corresponding to 26 GHz. In Site C, the amount of area that achieves a capacity
higher to 250 Mbps is 4.95 times higher with the site in 3.5 GHz than that corresponding to 26
GHz. Further, there are noticeable differences in the capacity achieved per each site into the
same frequency. However, the sites in 3.5 GHz have higher covered area with all the
capacities and the high amount of covered area with high capacity has been achieved with
the Site A located over a building and pointing to an open area.

Table 25. Downlink Capacity per covered area

Site A Site B Site C

26 GHz 3.5 GHz 26 GHz 3.5 GHz 26 GHz 3.5 GHz
Capacity ≥ 0 Mbps 0.315 0.683 0.251 0.699 0.196 0.477
Capacity ≥ 50 Mbps 0.120 0.342 0.049 0.231 0.012 0.122
Capacity ≥ 100 Mbps 0.086 0.197 0.019 0.090 0.003 0.039
(km²)
Capacity ≥ 150 Mbps 0.073 0.155 0.018 0.067 0.003 0.026
Capacity ≥ 200 Mbps 0.060 0.129 0.015 0.051 0.003 0.018
Capacity ≥ 250 Mbps 0.041 0.112 0.012 0.037 0.002 0.012

53
4.5. 5G small cell deployments

In this subsection, we will study the results obtained in coverage analysis prediction and
in capacity prediction for small cell deployments. The analysis will be done for different sites
and considering two cases for the band frequency: 26 GHz and 3.5 GHz in order to compare
their behaviors.

4.5.1. Determination of representative cell site locations and configuration

The site locations have been chosen with the aim to see the behavior of small cells in
different areas, and also taking into account the guidelines for installing small cell equipment
and poles in the City and County of Denver’s rights of way [19]. These sites are located as
follows: the first one in an urban park zone, the second one near to a corner in an urban zone,
and the last one located in the middle of the block street in a residential zone where it will be
possible to analyze the behavior in an open area and in a residential area, respectively. Figure
28 provides a description of the characteristics of the selected sites along with a Google Earth
view of the area surrounding the sites.

Characteristics Google Earth View of the Area surrounding the site

Site located in
Site X – an urban park,
Open Area pointing to
Small
open area.

Site located
near to a
Site Y – corner,
Residential pointing to
Area Small
other
buildings.

54
 Site located in
the middle of
Site Z – the block
Residential street in a
Area Small
residential
zone.

Figure 28. Selected sites for 5G small cell characterization.

In Table 26 and Figure 29 the main characteristics of the 5G small cells and transmitters
are shown respectively. Also, in Table 27, the Beamforming Antenna Parameters are detailed.

Table 26. 5G NR Small Cell Parameters

Site X / Site Y / Site Z

Frequency band n78 (3.5 GHz) / n257 (26 GHz)
Carrier (MHz) 50
Max power (dBm) 34
SSS EPRE (dBm) 5
PSS EPRE offest / SSS (dB) 3
Layer Small Cell Layer
Cell type PCell
Min SS-RSRP (dBm) -140
SS/PBCH numerology 3 (120 kHz)
SS/PBCH periodicity (ms) 10
SS/PBCH OFDM symbols {4,8,16,20}+28n [Lmax=4]
PDCCH overhead (OFDM symbols) 1
Traffic numerology 3 (120 kHz)
TDD DL OFDM symbols (%) 50
Radio equipment 5G NR Radio Equipment
Scheduler Proportional Fair
Diversity support (DL) Transmit Diversity; SU-MIMO; MU-MIMO
Diversity support (UL) Receive Diversity; SU-MIMO; MU-MIMO
Number of MU-MIMO users (DL) 1
Number of MU-MIMO users (UL) 1
Number of users (DL) 1
Number of users (UL) 1
Traffic load (DL) (%) 50
Traffic load (UL) (%) 100
Max traffic load (DL) (%) 100
Max Traffic Load (UL) (%) 100
UL noise rise (dB) 0
Fractional power control factor 1

55
 Number of required PRACH RSI 1
Max number of 4G/5G neighbors 16
Max number of inter-technology neighbors 16

Figure 29. Transmitter parameters for sites X, Y and Z.

The beamforming antennas parameters are shown in the following table:

Table 27. Beamforming Antenna Parameters in Small Cell

Parameters Default Beamformer

Columns (N) 8
Rows (M) 8
Transmission ports 64
Reception ports 64
Number of beams 60
Azimuth range -45º to 45º
Tilt range -5º to 5º
Boresight gain (dB) 26.06

The scenario is 5G NR Small Cell. In Figure 30, we can appreciate the three sites in a part of
the geographic area of study. In order to be able to compare the quantitative results, a
computation zone has been defined according to the possible better shape of the coverage
area for each site. This computation zone has a perimeter of 288 x 244 m and a total area of
0.07272 km2.

Figure 30. Representation of location of the sites and transmitters in Atoll.

56
 4.5.2. Coverage analysis

The coverage analysis is based on the prediction of the SS-RSRP in the downlink. For
each site, the percentage of the computation zone covered with values of SS-RSRP above a
given threshold has been computed. A representation of the coverage plot for the 3 sites is
represented in Figure 31. In addition to the 26 GHz band, predictions considering the use of
the 3.5 GHz band have been also accounted for the same of comparison. In general terms, it
could be seen that the covered surface with better levels of signal is higher in 3.5 GHz case.
Also, see in Figure 32.

26 GHz 3.5 GHz Legend

Site
X

Site
Y

Site
Z

Figure 31. Downlink Coverage SS-RSRP predictions per each site and its computation
zone.

57
 Figure 32. Sites using Aster propagation model with 3.5 GHz and using mmWave Aster
propagation model with 26 GHz.

Table 28 shows that, quantitatively, the behavior in a small cell is similar to frequencies
3.5 GHz and 26 GHz. In reference to the zone, the open area where is located the Site X has
a higher amount of area covered.

Table 28. Downlink coverage SS-RSRP

Site X Site Y Site Z
26 GHz 3.5 GHz 26 GHz 3.5 GHz 26 GHz 3.5 GHz
SS-RSRP ≥ -140dBm 0.0475 0.0526 0.0171 0.0325 0.0123 0.0209
SS-RSRP ≥ -130dBm 0.0343 0.0416 0.0079 0.0226 0.0051 0.0131
SS-RSRP ≥ -120dBm 0.0218 0.0327 0.0053 0.0169 0.0040 0.0092
(km²)
SS-RSRP ≥ -110dBm 0.0184 0.0258 0.0043 0.0106 0.0034 0.0071
SS-RSRP ≥ -100dBm 0.0133 0.0189 0.0032 0.0047 0.0023 0.0049
SS-RSRP ≥ -90dBm 0.0052 0.0076 0.0022 0.0026 0.0013 0.0017

Table 29 shows separately the coverage estimations for indoor and outdoor areas per
each site. In outdoor case, the behaviour in both frequencies is similar, they cover a similar
percentage of squared meters. In indoor case, the area covered with the site in 3.5 GHz is
higher than that corresponding to 26 GHz. Nevertheless, it is important to notice that the site
located near to the corner in the urban area has the higher percentage of indoor covered area
for both frequencies.

58
 Table 29. Outdoor and Indoor Coverage

Site X Site Y Site Z

Total Total Total
area 26 GHz 3.5 GHz area 26 GHz 3.5 GHz area 26 GHz 3.5 GHz
(km²) (%) (%) (km²) (%) (%) (km²) (%) (%)
Outdoor 0.0500 95.3837 97.9416 0.0252 65.3837 77.0179 0.0226 52.3553 60.3977
Indoor 0.0202 1.5718 15.9055 0.0451 1.7552 28.6734 0.0476 1.1251 15.7928

4.5.3. Capacity analysis

The prediction calculated is Peak RLC channel throughput (DL) (kbps), which is the
maximum RLC channel throughput attainable using the highest bearer available at the
subscriber location in the downlink. The configuration in the cells is set to 50% for TDD DL
OFDM Symbols and 50 % for Traffic load DL. The bandwidth of the carrier is set to 50 MHz in
all the cases, see Table 26.

A representation of the capacity plot for the 3 sites is represented in Figure 33. In addition
to the 26 GHz band, predictions considering the use of the 3.5 GHz band have been also
accounted for the same of comparison. It can be observed that the area where the service is
available is not so different than the area previously computed in the coverage analysis of the
SS-RSRP, in Figure 31.

26 GHz 3.5 GHz Legend

Site
X

Site
Y

59
Site
Z

Figure 33. Downlink Capacity predictions per each site and its computation zone.

Figure 34 shows graphically how the sites in 3.5 GHz located in urban area have higher
covered area with all the capacities. Nevertheless, the site in 26 GHz located in open area
covers more area than those located in urban area in both frequencies. Also, the differences
between the covered areas in open zone for both frequencies are small.

Figure 34. Downlink capacity for the different sites and frequency bands.

Table 30 provides the amount of covered area for different values of peak cell capacity.
From the results it can be observed that in Site Y and Site Z, the amount of area that achieves
a capacity higher to 0 Mbps is almost three times as high with the site in 3.5 GHz as with the
site in 26 GHz. For Site X, the differences are not high.
In all the sites, the amount of area that achieves a capacity higher to 200 Mbps is almost
similar with the site in 3.5 GHz than that corresponding to 26 GHz.

60
 Further, there are noticeable differences in the capacity achieved per each site into the
same frequency. However, Site X in 26 GHz have higher covered area with all the capacities
and the high amount of covered area with high capacity.

Table 30. Downlink Capacity per covered area with carrier of 50MHz.

Site X Site Y Site Z

26 GHz 3.5 GHz 26 GHz 3.5 GHz 26 GHz 3.5 GHz
Capacity ≥ 0 Mbps 0.036 0.043 0.009 0.024 0.006 0.014
Capacity ≥ 50 Mbps 0.020 0.030 0.005 0.015 0.004 0.009
km² Capacity ≥ 100 Mbps 0.014 0.020 0.004 0.005 0.003 0.005
Capacity ≥ 150 Mbps 0.014 0.020 0.004 0.005 0.003 0.005
Capacity ≥ 200 Mbps 0.010 0.017 0.003 0.004 0.002 0.004

4.6. Impact of increasing channel bandwidth to 400 MHz

For this section, the bandwidth of the carrier is set to 400 MHz in small cell working at 26
GHz. Figure 35 provides the maximum data rate value, that is 2210 Mbps, this value is
approximated to the one in Table 4 from Chapter 2, subsection 2.5, with subcarrier spacing
of 120 kHz, 64QAM as maximum modulation and 400 MHz of bandwidth. In that table the
value is 2128.8 Mbps, but it is considering a zero overhead. Also, Figure 35 shows graphically
how the site located in the open area has higher covered area with all the capacities.

Figure 35. Distribution of peak cell capacity for the different sites in 26 GHz.

61
 In Table 31 is noticeable that all the sites offer the higher capacity to a higher covered
area. They offer more than 2000 Mbps to nearly the third of the total covered area. Compared
with the results when a 50MHz bandwidth of the carrier is applied in 26 Ghz band, where the
higher capacity was nearly 200 Mbps, the covered areas in that capacity value is nearly three
times when the bandwidth is 400 MHz.

Table 31. Downlink Capacity per covered area using 26 Ghz band with carriers of
400MHz and 50MHz.

Site X Site Y Site Z

Capacity\Carrier 400 MHz 50 MHz 400 MHz 50 MHz 400 MHz 50 MHz
Capacity ≥ 0 Mbps 36100 36200 9100 9100 5600 5700
Capacity ≥ 100 Mbps 33300 14300 7600 3600 4900 2500
Capacity ≥ 150 Mbps 31000 14300 6500 3600 4700 2500
Capacity ≥ 200 Mbps 29900 9800 6200 3000 4600 1900
Capacity ≥ 400 Mbps 22100 5500 4000
Capacity ≥ 600 Mbps 19600 4900 3500
m² Capacity ≥ 800 Mbps 18100 4400 3200
Capacity ≥ 1000 Mbps 14300 3500 2400
Capacity ≥ 1200 Mbps 14300 3500 2400
Capacity ≥ 1400 Mbps 14300 3500 2400
Capacity ≥ 1600 Mbps 14300 3500 2400
Capacity ≥ 1800 Mbps 13400 3500 2400
Capacity ≥ 2000 Mbps 11300 3200 2100

4.7. Impact of beamforming configuration

The analysis of capacity predictions has been done for different beamforming
configurations. We compare the behaviour in small and macro cell scenarios, in which the
transmitter is configured with and without beamforming, working in the 26 GHz band with
400MHz carrier bandwidth. When the transmitter configuration is set to the option ‘Without
Beamforming’, a specific antenna must be chosen. In this study, the 3GPP Antenna has been
selected, its properties are shown in Figure 36.

62
 Figure 36. 3GPP Antenna Properties.

When the transmitter is using beamforming, two configurations of the antenna parameters
have been studied. The number of antenna elements and the number of beams are different
for each beamformer. The beamforming antennas parameters, including their gains are shown
in Table 32 following:

Table 32. Beamforming Antenna Parameters

Parameters Default Beamformer Beamformer 2

Columns (N) 8 4
Rows (M) 8 4
Transmission ports 64 64
Reception ports 64 64
Number of beams 60 20
Azimuth range -45º to 45º -45º to 45º
Tilt range -5º to 5º -5º to 5º
Boresight gain (dBi) 26.0618 20.0412

First, the Small Cell scenario has been studied. The site X has been selected to do the
current capacity analysis. In Figure 37, it is noticeable that the covered area is lower when
an antenna without beamforming is used. Also, Figure 38 shows that the site, when antenna
with Default Beamforming configuration is used, has almost all higher values of covered areas
for each capacity values than the other configurations, and it has the highest values in the
better capacity values.

63
 Site X Legend

Without
Beamforming

Beamforming 2

Default
Beamforming

Figure 37. Downlink Capacity predictions using different beamforming configurations in

Small Cell.

64
 Figure 38. Downlink Capacity using different beamforming configurations in Small Cell.

Table 33 shows the numerical results for the downlink capacity per covered area for the
site working at 26 Ghz band with 400MHz carrier using different beamforming configurations.
When the antenna configuration is Default Beamforming, which is using 64 elements and 60
beams, the results of covered area for the better capacity is the highest.

Table 33. Downlink Capacity per covered area using different beamforming
configurations in Small Cell.

Site X
Without Beamforming Default
Beamforming 2 Beamforming
Capacity ≥ 0 Mbps 17900 25100 36100
Capacity ≥ 200 Mbps 17000 20200 29900
Capacity ≥ 400 Mbps 14000 18800 22100
Capacity ≥ 600 Mbps 8500 17100 19600
Capacity ≥ 800 Mbps 5200 14300 18100
m² Capacity ≥ 1000 Mbps 1900 8500 14300
Capacity ≥ 1200 Mbps 1900 8500 14300
Capacity ≥ 1400 Mbps 1900 8500 14300
Capacity ≥ 1600 Mbps 1900 8500 14300
Capacity ≥ 1800 Mbps 1700 7800 13400
Capacity ≥ 2000 Mbps 1300 6300 11300

Second, the Macro Cell scenario has been studied. The site A has been selected to do
the current capacity analysis. In Figure 39, it is noticeable that the covered area is lower
when an antenna without beamforming is used. Also, Figure 40 shows that the site when

65
antenna with Default Beamforming configuration is used, has almost all higher values of
covered areas for each capacity values than the other configurations.

Site A Legend

Without
Beamforming

Beamforming
2

Default
Beamforming

Figure 39. Downlink Capacity predictions using different beamforming configurations in

Macro Cell.

66
 Figure 40. Downlink Capacity using different beamforming configurations in Macro Cell.

Table 34 shows the numerical results for the downlink capacity per covered area for the
site working at 26 Ghz band with 400MHz carrier using different beamforming configurations.
When the antenna configuration is Default Beamforming, the results of covered area are
higher but in a low capacity value.

Table 34. Downlink Capacity per covered area using different beamforming
configurations in Macro Cell.

Site A
Without Beamforming Default
Beamforming 2 Beamforming
Capacity ≥ 0 Mbps 0.0968 0.1789 0.3803
Capacity ≥ 200 Mbps 0.0897 0.1287 0.2409
Capacity ≥ 400 Mbps 0.0727 0.1071 0.1468
Capacity ≥ 600 Mbps 0.0442 0.0884 0.105
Capacity ≥ 800 Mbps 0.0245 0.0718 0.0954
km² Capacity ≥ 1000 Mbps 0.0167 0.0627 0.0866
Capacity ≥ 1200 Mbps 0.0136 0.0579 0.0827
Capacity ≥ 1400 Mbps 0.0092 0.0504 0.077
Capacity ≥ 1600 Mbps 0.007 0.0427 0.0713
Capacity ≥ 1800 Mbps 0.0053 0.0351 0.0643
Capacity ≥ 2000 Mbps 0.0035 0.0287 0.0558

67
5. Network Deployment
In this chapter, we are going to determine and model all the requirements to analyse the
capacity and connectivity behaviours in 5G macro and small cell deployments working in
mmWave band. Simulations will be used to get the desired results. There are some
requirements before creating simulations: network services and users need to be modelled
and, a traffic map has to be created.

5.1. Determination and modelling of network services and users

We are supposing a hypothetical scenario: a broadband access service in crowd scenario

for example a festival in which there will be a concert. The location has to be in an open area,
for example a park, because of the necessities of space and free walking. The number of
people attending the event means that the place will be quite crowded, and it is needed that
the active users get the highest possible capacity when they are connected.

The configuration for the sites, transmitters and antennas are the ones that were analysed
in previous sections, for sites in Macro Cell as in Table 21, and for sites in Small Cell as in
Table 26, in general, the configuration of all transmitters is 26 GHz or mmWave band using
400MHz bandwidth carrier. The beamforming antennas used will be the Default Beamforming
showed in Table 27.

First, a new “RequiredService” has been created with the following properties, see Figure
41. For the new service, the Downlink capacity required is approximately 25 Mbps and the
Uplink capacity required is approximately 50 Mbps, as the 3GPP TS 22.261 document
recommends [20]. Therefore, those need to be the values of Average requested throughput.
For the values of Max throughput demand in Uplink and Downlink, we have considered that
all the capacity could be given for a single user in the case that only one needs to transmit or
receive data. Also, for the values of Min throughput demand in Uplink and Downlink, the
minimum possible demand is zero. With these parameters we prevent unnecessary resource
saturation which can increase the number of rejected users.

Figure 41. Properties of the new Required Service.

68
 A new user, named Basic user, has been modelled in order to specify its capacity
requirements, see Figure 42. Commonly, during a concert, people upload shorts videos in
some social networks, the average volume for these videos is 40 Mbps. The download volume
is lower in a concert or festival, due to people are generating social content and not consuming
it at that moment.

Figure 42. Properties of the new user.

Then, a new environment has to be created, its name is Ciutadella, and its properties are
shown in Figure 43. To determine the number of people per square kilometer or density in the
environment, we suppose 200000 Suscribers/km2, which is a lower value recommended in
[20] for this kind of crowd environment. For this density value, the number of assistants to the
event can be between 20000 to 30000, in this particular case it will be 26000 assistants. This
value divided by the surface area of the zone of interest (0.1356 km²) give to us the density,
200000 Suscribers/km2.

Figure 43. Properties of the new Environment.

5.2. Creation of the traffic map.

According to all the results obtained in the previous subsection 5.1 and the zone
requirement of the hypothetical scenario, we have selected an open area for capacity analysis,
the zone is shown in Figure 44 in grey.

69
 Figure 44. Traffic zone.

5.3. Justification for network deployments

One of the most important and difficult part for a network deployment is the correct location
for the sites, for both Macro and Small cell deployments.

To select the site locations, we are going to use Downlink Quality Prediction which allows
us to know the C/I+N level as was explained in subsection 3.4.2. Also, to choose the locations,
the conclusions obtained in Chapter 4 have been taken into account, and according to this
some possible deployments were tested, but only the best options will be explained.

5.3.1. Site locations for Macro cell deployments

Three deployments have been developed, in the first one, it will be just one site active and
in the other two, it will be two sites, because the surface area is physically enough for more
than one cell. For all the cases and in order to compare the results, the quality prediction
graphs and interference analysis are developed.

First, one site located over a high building is activated, as we can see in Figure 45, its
quality coverage prediction graph is more uniform than the one when two sites are activated.
In terms of interference, it increases when two sites are active as the results of the point
analysis show. However, in the case when Site 1 and Site 2 are active the interference level
is lower than in the case with Site 1 and Site 2_a.

70
Quality Prediction Legend
Site 1

Site 1 and Site 2

Site 1 and Site 2_a

71
 Figure 45. Quality predictions and interference results.

Also, in the results for the predictions in Table 35, the signal level is better with the second
case even when there is interference. Therefore, for the case when two active sites are
needed, we are going to choose the deployment with Site 1 and Site 2.

Table 35. Quality Prediction Results in Macro Cell.

Site 1 Site 1 and Site 2 Site 1 and Site 2_a
Mean value PDSCH C/(I+N) Level (DL)
15.9 16.01 13.96
(dB)

5.3.2. Site locations for Small cell deployments

Three deployments have been developed, with two three and four sites active, because
the surface area is physically enough for more than two small cells. For all the cases and in
order to compare the results, the quality prediction graphs and interference analysis are
developed.

First, the case with two active sites has been tested with three pairs of sites and the results
are shown in Figure 46.

72
 Quality Prediction Legend
Site 3 and Site 6

Site 3 and Site 7

Site 4 and Site 5

Figure 46. Quality predictions and interference results with two small cells.

73
 Also, in the results for the predictions in Table 36, the signal level is better with the first
case. Therefore, for the case when two active sites are needed, we are going to choose the
deployment with Site 3 and Site 6.

Table 36. Quality Prediction Results in Small Cell with 2 sites.

Site 3 and Site 6 Site 3 and Site 7 Site 4 and Site 5
Mean value PDSCH C/(I+N)
16.34 15.98 14.04
Level (DL) (dB)

Now, we have two fixed sites, for the third site, we are going to choose between the sites
4 and 5 due to they are physically well located. The first case is the better, so it will be used
when three sites are needed, see Table 37.

Table 37. Quality Prediction Results in Small Cell with 3 sites.

Site 3, Site 6 & Site 4 Site 3, Site 6 & Site 5
Mean value PDSCH C/(I+N)
17.41 15.29
Level (DL) (dB)

When, four sites are active the Mean value PDSCH C/(I+N) Level (DL) is 17.98 dB.

5.4. Simulation of network deployments

Network services and users were modelled and a traffic map has been created. Now, we
have set all the requirements needed to create a simulation.

In network tab, we can find the folder Simulations. When we select to create a New
simulation, it is needed to specify the traffic that will be used, as Figure 47 shows, we are
going to use the traffic map, Ciutadella, created before in subsection 5.1. This traffic zone is
using as parameters, the new Required Service, the new Basic User and the new Environment.

Figure 47. Selection of traffic map in a new Simulation.

For all the simulations, the configurations were explained in previous subsection 5.1. It is
important to notice that for the other services available in Atoll, such as: Broadband, Internet,

74
Video Call, Voice Call and Machine-Type, the number of users is zero because we are using
only the Required Service defined before.

As we explained in subsection 3.5.4, simulations follow some internal steps to achieve its
main objective: to simulate network traffic.

We are going to use the map to display the distribution of traffic that is generated by all
simulations according to activity status. The simulation will be done for the sites set as active
in its corresponding network (Macro or Small cell). It is important to take care about the
coherent settings. Some numerical and graphical results of simulations are showed in Table
38, Figure 48 and, Figure 49.

Table 38. Simulations results in Macro and Small Cell deployment.

Macro Cell Small Cell
Results \ Number of sites 1 2 2 3 4
Total users trying to connect 495 489 462 469 481
Percentage from the total number of users (%) 2.4 1.2 11 7.5 2.9
No Coverage 10 4 11 15 7
Rejected Users
No Service 2 2 40 20 7
Resource Saturation 0 0 0 0 0
Percentage from the total number of users (%) 97.6 98.8 89 92.5 97.1
Active Downlink 226 258 207 226 236
Connected Users Active Uplink 254 224 204 206 229
DL Peak RLC Cumulated Throughput (Mbps) 3404.06 8388.07 2591.32 5618.93 7999.42
UL Peak RLC Cumulated Throughput (Mbps) 4960.89 11764.74 3793.34 7614.06 11955.09
DL capacity per user (Mbps) 15.06 32.51 12.52 24.86 33.90
Experienced data rate
UL capacity per user (Mbps) 19.53 52.52 18.59 36.96 52.21

75
Active sites Simulation

Site 1

Site 1 and
Site 2

Figure 48. Traffic simulation in Macro Cell environment.

76
Active sites Simulation

Site 3 and 6

Site 3, 4 and
6

Site 3, 4, 5
and 6

Figure 49. Traffic simulation in Small Cell network.

77
5.4.1. Analysis with variations of parameters

To analyze the effects of user parameters in connectivity and capacity, we are going to
vary the number of videos per hour per user in every deployment. The range of variation is
from 1 to 10 videos/hour, to obtain the experienced uplink and downlink bit rate per user, the
number of active users and rejected users. It will be used a set of 10 simulations for every
configuration, in order to have the lower standard deviation in the results.

As Figure 50 and Figure 51 show that in all deployments, when the number of video/hour
increases the uplink and downlink bit rate experienced per user decreases. The desired
experienced bit rate of 50 Mbps for Uplink and 25 Mbps for Downlink is achieved:

o In Macro Cell deployments, when:

▪ Videos/hour <= 2 for the deployment with 1 site
▪ Videos/hour <= 5 for the deployment with 2 sites

o In Small Cell deployments, when:

▪ Videos/hour <= 2 for the deployment with 2 sites
▪ Videos/hour <= 3 for the deployment with 3 sites
▪ Videos/hour <= 5 for the deployment with 4 sites

Figure 50. Experienced Uplink bit rate from simulations results.

78
 Figure 51. Experienced Downlink bit rate from simulations results.

However, Figure 52 shows that the active users increase as capacity demand increases.

Figure 52. Active users from simulations results.

In Figure 53, for Macro Cell deployments, the percentage of rejected users is:
• between 3% and 4% in deployment with 1 site, and
• between 1.9% and 2.6% in deployment with 2 sites.
For Small Cell deployments, the percentage of rejected users is:
• between 12.7% and 15% in deployment with 2 sites,
• between 6.4% and 7.4% in deployment with 3 sites, and
• between 4% and 5.9% in deployment with 4 sites.

79
 Figure 53. Rejected users from simulations results.

With macro cell the number active users is little bit higher than with small cell case. Small
cell with 4 sites achieves DL and UL higher capacities per user. Macro cell network present
less percentage of rejected users.

80
6. Conclusions and future development

5G NR access will include mmWave frequencies (above 24.25 GHz) which are expected
to support data rates in the order of Gbps, however, they present some inconvenients. The
free space path loss will be increased because it is dependent on the squared of the carrier
frequency. The effect of blockage, also will be increased, as a result of weaker non-line-of-
sight paths. Another affection is the attenuation due to rain and atmospheric absorption.

This thesis studies the performance in terms of capacity and coverage of different macro
and small cell deployment options for the exploitation of the 26 GHz band in 5G. The study is
conducted using the ATOLL radio planning tool.

Considering the aforementioned, in the current work, representative sites have defined
with the aim that the cells cover geographically different areas to analyze the behavior in an
open area, in a residential area, and in a dense residential area. The representative macro
cell sites were located in an urban park zone, urban zone, and dense urban zone. For small
cell, also three representative sites have been determined. These sites are located as follows:
the first one in an urban park zone, the second one near to a corner in an urban zone, and the
last one located in the middle of the block street in a residential zone.

Then, the selection of a propagation model was needed. The comparison between two
propagation models, 3GPP 38.900 model and mmWave Aster model, supported in ATOLL to
cope with mmWave frequencies, were done. These models were intended to be used in both
outdoor to outdoor (O2O) and outdoor to indoor (O2I) scenarios, to choose the most robust
propagation model for 5G in mmWave band. The mmWave Aster propagation model with its
default values was selected because its level of characterization is higher in terms to consider
the different clutter classes and be able to take into account the O2I losses.

The results from coverage analysis prediction and from capacity prediction for macro and
small cell deployments with the representative sites have been obtained in two frequency
bands: 26 GHz and 3.5 GHz in order to compare their behaviors. The coverage analysis is
based on the prediction of the SS-RSRP in the downlink. For each site, the percentage of the
computation zone covered with values of SS-RSRP above a given threshold has been
computed. The prediction calculated to evaluate the capacity is Peak RLC channel throughput
(DL) (kbps), which is the maximum RLC channel throughput attainable using the highest
bearer available at the subscriber location in the downlink. Initially, the results were obtained
with a 50 MHz channel bandwidth. After, they were compared with the results obtained with
400 MHz. Also, different beamforming configurations were applied to prove that technologies
such as massive MIMO and beamforming that make use of hundreds of antenna elements,
help to get better signal level and consequently get better values of capacity and connectivity.
Therefore, these parameters configuration analysis were important to get the most accurate
macro and small cell network deployments and be able to obtain simulations of mobile traffic
with coherent results. Most relevant concluding remarks are:

For 5G macro cell deployments:

- In coverage analysis, the covered area when the site is in 3.5 GHz is higher than when
the site is in 26 GHz, in numerical terms, with macro cell open area site the covered
area is 2.22 times higher, with macro cell urban area site is 3.24 times higher, and with
macro cell dense area site is 6.84 higher. Nevertheless, the variation between the

81
 covered area with 3 sites using the same frequency is not significant. For example, in
26 GHz case, covered area with SS-RSRP ≥ -140dBm is: 0.647 km2 with macro cell
open area site, 0.618 km2 with macro cell urban area site, and 0.577 km2 with macro
cell dense area.
- In indoor coverage, with sites in 3.5 GHz, the percentage of coverage for indoor areas
at different heights achieves in average a 10% of indoor total area in the computation
zone, which is lower than the average percentage of coverage for outdoor areas of
91%. With sites in 26 GHz, the percentage of coverage for indoor areas at different
heights just achieve less than 2%, which is lower than the percentage of coverage for
outdoor areas of nearly 70%.
- In capacity coverage analysis, with carrier bandwidth set to 50 MHz in all the cases,
the covered area is from 2 to 6 times higher with sites in 3.5 GHz than with sites in 26
GHz. For example: the total area covered with values of Capacity ≥ 200 Mbps using
the macro cell open area site is: 0.129 km2 when 3.5 GHz is used and 0.06 km2 with
26 GHz.

For 5G small cell deployments:

- In coverage analysis, quantitatively the behavior for the covered area is similar for
frequencies 3.5 GHz and 26 GHz. Moreover, small cell open area site achieves the
highest amount of covered area with values SS-RSRP ≥ -140dBm: 52600 m2 using
3.5 GHz, and 47500 m2 using 26 GHz.
- In indoor coverage analysis, small cell sites in 3.5 GHz cover between 15 - 28% of the
total indoor area. In 26 GHz, they cover between 1.12 – 1.76% of the total indoor area.
- In capacity coverage analysis, with carrier bandwidth set to 50 MHz in all the cases,
the covered area is on average twice as high with sites in 3.5 GHz as with sites in 26
GHz. For example: the total area covered with values of Capacity ≥ 200 Mbps using
the small cell open area site are: 17300 m2 when 3.5 GHz is used and 10200 m2 with
26 GHz.
- When the channel bandwidth is set to 400 MHz, in all the sites, the data rate value
increases to more than 2000 Mbps instead of 200 Mbps achieved using 50 MHz
channel bandwidth. If we compare the amount of covered area where the Capacity ≥
200 Mbps, the value with 50 MHz channel bandwidth (9800 m2) is about the third part
of the area when the bandwidth is 400 MHz (29900 m2).

The impact of beamforming configuration was performed in small and macro cell scenarios,
in which the transmitter was configured with and without beamforming, working in the 26 GHz
band with 400MHz carrier bandwidth. When the transmitter configuration is set to the option
‘Without Beamforming’, a specific antenna must be chosen. In this study, the 3GPP Antenna
has been selected. When the transmitter is using beamforming, two configurations of the
antenna parameters have been studied. ‘Default beamformer’ is a 8X8 antenna with 60 beams,
and ‘beamformer 2’ is a 4X4 antenna with 20 beams.

- In Macro Cell deployments, when the site is using an antenna with ‘Default
Beamforming’ configuration, it covers from 3 to 16 times more area than the ‘without
beamforming’ configuration, and it covers from 2 to 8 times more area than the
‘beamformer 2’ configuration, for all capacity values. For example, the covered area
with for Capacity ≥ 2000 Mbps, using ‘Default Beamforming’ is 0.0558 km2, using
‘beamformer 2’ is 0.0287 km2 and ‘without beamforming’ is just 0.0035 km2.

82
 - In Small Cell deployments, when the site is using an antenna with ‘Default
Beamforming’ configuration, it covers from 2 to 9 times more area than the ‘without
beamforming’ configuration, and it covers from 2 to 5 times more area than the
‘beamformer 2’ configuration, for all capacity values. For example, the covered area
with Capacity ≥ 2000 Mbps, using ‘Default Beamforming’ is 11300 m2, using
‘beamformer 2’ is 6300 m2 and ‘without beamforming’ is just 1300 m2.

Finally, for both 5G Macro and Small Cell in mmWave band, it has been possible to get
five network deployment options as a result of the analysis of coverage and capacity
predictions. They have better results for outdoor situations when the site is located in an open
area.

Since some network deployment options emerged as answer to find new wireless
spectrum which allows greater capacity, the simulation of mobile traffic load was generated
as follows: in a crowd scenario like a concert with a required service of broadband with
downlink capacity of 25 Mbps and the uplink capacity of 50 Mbps as the 3GPP TS 22.261
document [20] recommends for this scenario, with users sending or posting videos in social
networks, and with the option of varying the traffic load (videos/hour).

In the simulation results, data rates and connectivity experienced by users were analysed.
The number of active users is little bit higher in macro cell than in small cell deployments. The
percentage of rejected users is: between 2-4% in macro cell deployments and between 4-
12.7% in small cell deployments.

Although, macro cell deployments have low values of rejected users, their experienced UL
bit rates per user are less than ones in the small cell deployments. For example, when the
traffic load is the lowest (1 video/hour), the experienced UL bit rates per user is between 25-
100 Mbps less. Also, their experienced DL bit rates per user are less than ones in the small
cell deployments. For example, for the 70% of variations of traffic load, the macro cell
deployments with 2 sites get less values of experienced DL bit rates than the small cell
deployment with 4 sites.

Small cell deployment with 4 active sites achieves UL and DL higher capacities per user.
The experienced UL bit rates per user is almost the double of the value get with small cell
deployment with 2 active sites and is nearly 1.5 times higher than the value get with small cell
deployment with 3 active sites. The experienced DL bit rates per user is nearly the triple of the
value get with small cell deployment with 2 active sites and is nearly 1.5 times higher than the
value get with small cell deployment with 3 active sites. Also, the percentage of rejected users,
with this deployment of 4 small cell sites, is between 4% and 5.9%, being the lowest values in
this cell deployment option.

The desired experienced UL bit rate ≥ 50 Mbps is reached by small cell deployment with
4 active sites even when there are 2-3 videos/hour more than in traffic load that allows that bit
rate with the other small cell deployments. The desired experienced DL bit rate ≥ 25 Mbps is
reached by small cell deployment with 4 active sites even when there are 4-6 videos/hour
more than in traffic load that allows that bit rate with the other small cell deployments.

Overall, the simulation results demonstrate an important increase of the performance of

the network as a function of 5G NR features, specially, in the use of a mmWave band and 400
MHz carrier bandwidth. Nevertheless, these improvements are present mainly in outdoor

83
scenarios when the sites are located or pointing to an open area. Thus, network operators
must analyse if the benefits prevail over the inconvenient for indoor cases and the cost
generated by possible additional equipment.

For future work, it would be interesting to analyze the network’s behavior when Macro and
Small cells are working together in order to improve the coverage and capacity results.

84
Bibliography

[1] Chen, S., & Zhao, J. “The requirements, challenges, and technologies for 5G of
terrestrial mobile telecommunication”. IEEE communications magazine, vol 52, issue
5, pp 36-43, May 2014. DOI: 10.1109/MCOM.2014.6815891.
[2] Zaidi, A., Athley, F., Medbo, J., Gustavsson, U., Durisi, G., & Chen, X. 5G Physical
Layer: Principles, Models and Technology Components. Academic Press, (2018).
[3] Rappaport, T. S., Sun, S., Mayzus, R., Zhao, H., Azar, Y., Wang, K., Gutierrez, F., et
al. “Millimeter wave mobile communications for 5G cellular: It will work!”. IEEE access,
vol 1, pp 335-349, May 2013. DOI: 10.1109/ACCESS.2013.2260813.
[4] 3rd Generation Partnership Project, “3GPP TR21.915 Release 15 Description”, Oct
2018.
[5] ITU-R, “Minimum requirements related to technical performance for IMT-2020 radio
interface(s)”, 2017.
[6] 3rd Generation Partnership Project, “3GPP TS 38.211 NR Physical channels and
modulation”, Oct 2018.
[7] 3rd Generation Partnership Project, “3GPP TS 38.215 5G NR Physical layer
measurements”, Oct 2018.
[8] L. Liu, R. Chen, S. Geirhofer, K. Sayana, Z. Shi and Y. Zhou, “Downlink MIMO in LTE-
Advanced: SU-MIMO vs. MU-MIMO”, IEEE Communications Magazine, Vol. 50, Issue
2, pp 140-147, February 2012. DOI: 10.1109/MCOM.2012.6146493.
[9] Johnson Chris, “5G New Radio in Bullets”. First Edition. Monee, IL. USA, 2019
[10] Forsk, “Atoll 3.4.0: User Manual for Radio Networks”, 2019.
[11] Forsk, “Atoll 3.4.0: Technical Reference Guide for Radio Networks”, 2018.
[12] Visicom, “Visicom”, 2018. [Online]. Available:
https://visicomdata.com/samples?lang=en.
[13] 3rd Generation Partnership Project; “3GPP TR 38.900. Study on channel
model for frequency spectrum above 6 GHz”, 2018.
[14] Forsk, “Aster 2.7.1: User Manual”, 2019.
[15] Forsk, “Aster 2.7.1: Technical Reference Guide”, 2019.
[16] Antenas GSM. Available:
https://antenasgsm.com/41.3897674697056/2.195751714706393/15/true,true,true,tr
ue,true.
[17] NR radio resources. Available: http://niviuk.free.fr/store_nr.php#nrb_fr
[18] 5G Tools for RF Wireless. Available: https://5g-tools.com/5g-nr-throughput-
calculator/
[19] Denver Public Works, “City and County of Denver Small Cell Infrastructure
Design Guidelines“, Nov 2019.
[20] 3rd Generation Partnership Project; “3GPP TS 22.261. Service requirements
for the 5G system”, Sept 2019.

85
 Abbreviations and acronyms

3GPP Third Generation Partnership Project

5G Fifth Generation
ARFCN Absolute Radio-Frequency Channel Number
BPSK Binary Phase Shift Keying
CDMA Code Division Multiple Access
CINR Carrier to Interference and Noise Ratio
CP Cyclic Prefix
CSI-RS Channel State Information Reference Signal
DL Downlink
DM-RS Demodulation Reference Signal
DTM Digital Terrain Model
eMBB Enhanced Mobile Broadband
EPRE Energy Per Resource Element
FDD Frequency Division Duplex
GSM Global System for Mobile communications
I2I Indoor to indoor
I2O Indoor to outdoor
IMT-2020 International Mobile Telecommunications 2020
ITU International Telecommunications Union
ITU-R International Telecommunications Union-Radio Communications Sector
ISD Intersite Distance
LTE Long-Term Evolution
MAC Medium Access Control
MIMO Multiple-Input Multiple-Output
mMTC Massive Machine Type Communication
MU-MIMO Multi-User MIMO
NB-IoT Narrow-Band Internet-of-Things
NR New Radio
O2I Outdoor to Indoor
O2O Outdoor to outdoor
OFDM Orthogonal Frequency-Division Multiplexing
PBCH Physical Broadcast Channel
PDCCH Physical Downlink Control Channel
PDCP Packet Data Convergence Protocol
PDSCH Physical Downlink Shared Channel
PHY Physical Layer
PRACH Physical Random-Access Channel
PSS Primary Synchronization Signal
PT-RS Phase Tracking Reference Signal
PUCCH Physical Uplink Control Channel
PUSCH Physical Uplink Shared Channel
QAM Quadrature Amplitude Modulation
QPSK Quadrature Phase Shift Keying
RB Resource Block
RLC Radio Link Control
SDAP Service Data Adaption Protocol
SISO Single-Input Single-Output
SNR Signal to Noise Ratio

86
SPM Standard Propagation Model
SRS Sounding Reference Signal
SS Synchronization Signal
SS-RSRP SS Reference Signal Received Power
SSS Secondary Synchronization Signal
SU-MIMO Single-User MIMO
TDD Time Division Duplex
UE User Equipment
UL Uplink
UMTS Universal Mobile Telecommunication System
URLLC Ultra-Reliable Low-Latency Communication

87