Published by John Wiley & Sons, Inc., Hoboken, New Jersey.
Published simultaneously in Canada.

No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission.

Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.

Library of Congress Cataloging-in-Publication Data is available.

ISBN: 978-1-118-85434-1

CONTENTS

ABOUT THE AUTHORS

FOREWORD

ACRONYMS

PART I INTRODUCTION

1 Small Cells—The Future of Cellular Networks
1. 1.1 Introduction
2. 1.2 The Industry Challenge
3. 1.3 Are Small Cells the Answer?
4. 1.4 A Brief History of Small Cells
5. 1.5 Small Cell Challenges and Outline of the Book
6. References
2 100× Capacity Scaling of Cellular Networks
1. 2.1 Introduction
2. 2.2 System Model
3. 2.3 Network Densification
4. 2.4 Higher Frequency Bands
5. 2.5 Multi-antenna Techniques and Beamforming
6. 2.6 Scheduling
7. 2.7 Energy Efficiency
8. 2.8 What Is Different in Ultra-Dense Small Cell Deployments
9. 2.9 Summary and Conclusions
10. References
3 Automation of Cellular Networks
1. 3.1 Introduction
2. 3.2 Self-Organizing Network Use Cases and Standardization
3. 3.3 Intelligent Techniques for Self-Organizing Networks
4. 3.4 Case Studies
5. 3.5 Summary and Conclusions
6. References

PART II COVERAGE AND CAPACITY OPTIMIZATION

4 Frequency Assignment and Access Methods
1. 4.1 Introduction
2. 4.2 Frequency Assignment Approaches
3. 4.3 Access Methods
4. 4.4 Co-channel Operation of Residential Femtocells
5. 4.5 Multi-carrier Soft-Reuse
6. 4.6 Summary and Conclusions
7. References
5 Coverage and Capacity Optimization for Indoor Cells
1. 5.1 Introduction
2. 5.2 Initial Configuration of Coverage
3. 5.3 Coverage Optimization of Residential Femtocells
4. 5.4 Coverage Optimization of Femtocell Groups
5. 5.5 Summary and Conclusions
6. References
6 Coverage and Capacity Optimization for Outdoor Cells
1. 6.1 Introduction
2. 6.2 Cell Range Expansion
3. 6.3 Cell Range Expansion Bias Optimization
4. 6.4 Coverage Optimization by Using Switched Multi-element Antenna Systems
5. 6.5 Summary and Conclusions
6. Appendix 6.A: Proof of Theorem 1
7. Appendix 6.B: Proof of Theorem 2
8. Appendix 6.C: Proof of Theorem 3
9. Appendix 6.D: Derivation of REB for a Desired Coverage Radius
10. Notes
11. References

PART III INTERFERENCE MANAGEMENT

7 Frequency-Domain Inter-cell Interference Coordination: Frequency-Domain Inter-cell Interference Coordination
1. 7.1 Introduction
2. 7.2 Carrier Aggregation in LTE-Advanced
3. 7.3 Protocol Architecture and Implementation Aspects
4. 7.4 HetNet Deployments
5. 7.5 Summary and Conclusions
6. Notes
7. References
8 Time-Domain Inter-cell Interference Coordination: Time-Domain Inter-cell Interference Coordination
1. 8.1 Introduction
2. 8.2 Almost Blank Subframes
3. 8.3 Configuration and Optimization
4. 8.4 Non-linear Relaxation Approach
5. 8.5 Stochastic Relaxation Approach
6. 8.6 Summary and Conclusions
7. Notes
8. References
9 The Sector Offset Configuration
1. 9.1 Introduction
2. 9.2 Macrocell Tier with Horizontal Sector Offset Configuration
3. 9.3 Macrocell Tier with Horizontal and Vertical Sector Offset Configuration
4. 9.4 Small Cell Tier with Sector Offset Configuration
5. 9.5 Scenario and System Model
6. 9.6 Performance Comparison for Macrocell-Only Scenario
7. 9.7 Performance Comparison for HetNet Scenario
8. 9.8 Summary and Conclusions
9. Notes
10. References
10 Control Channel Inter-cell Interference Coordination
1. 10.1 Introduction
2. 10.2 Concept of Orthogonally Filled Subframes
3. 10.3 Formal Definition
4. 10.4 Implementation
5. 10.5 Performance Evaluation
6. 10.6 Summary and Conclusions
7. Notes
8. References
11 Uplink-Oriented Optimization in Heterogeneous Networks
1. 11.1 Introduction
2. 11.2 3G CDMA Heterogeneous Networks
3. 11.3 4G LTE Heterogeneous Networks
4. 11.4 Summary and Conclusions
5. Notes
6. References

PART IV MOBILITY MANAGEMENT AND ENERGY EFFICIENCY

12 Mobility Management
1. 12.1 Introduction
2. 12.2 The Handover Process
3. 12.3 Radio Link failure, Handover Failure and Unnecessary Handovers
4. 12.4 Mobility Challenges in Heterogeneous Networks
5. 12.5 Mobility Performance in Heterogeneous Networks
6. 12.6 Mobility Optimization in Heterogeneous Networks
7. 12.7 Mobility State Estimation
8. 12.8 Summary and Conclusions
9. References
13 Dormant Cells and Idle Modes
1. 13.1 Introduction
2. 13.2 Potential Energy Saving gains of Deploying Small Cells
3. 13.3 Distributed Idle Mode Procedure for Small Cell Base Stations
4. 13.4 Centralized Idle Mode Procedure for Small Cell Base Stations
5. 13.5 Summary and Conclusions
6. References

PART V SMALL CELL DEPLOYMENT

14 Backhaul for Small Cells
1. 14.1 Introduction
2. 14.2 Wireless Backhaul
3. 14.3 Wired Backhaul
4. 14.4 Cost
5. 14.5 Summary and Conclusions
6. References
15 Optimization of Small Cell Deployment
1. 15.1 Introduction
2. 15.2 Overview of the Small Cell Site Selection Process
3. 15.3 UE Traffic Demand Map Generation
4. 15.4 Small Cell Deployment Optimization
5. 15.5 Summary and Conclusions
6. References

PART VI FUTURE TRENDS AND APPLICATIONS

16 Ultra-Dense Networks
1. 16.1 Introduction
2. 16.2 The Attocell
3. 16.3 Challenges in Attocell Deployments and Operation
4. 16.4 Simulation of an Attocell Deployment
5. 16.5 Simulations of Mobility
6. 16.6 Future Enhancements
7. 16.7 Summary and Conclusions
8. References
17 HetNet Applications
1. 17.1 Introduction
2. 17.2 Localized and Personalized Push Services
3. 17.3 Proximity and Presence Detection
4. 17.4 Indoor Localization
5. 17.5 Home Automation
6. 17.6 Local Access Control
7. 17.7 Summary and Conclusions
8. References

A SIMULATING HETNETS

A.1 Introduction
A.2 Network Layout Modeling
A.3 Antenna Gain Modeling
A.4 Path-Loss Modeling
A.5 Environment Loss Modeling
A.6 Shadow Fading Modeling
A.7 Multi-path Fading Gain Modeling
A.8 Received Signal Strength and Quality Modeling
A.9 Throughput Modeling
A.10 User Equipment Location and Traffic Modeling
A.11 Mobility Modeling
A.12 Summary and Conclusions
References

INDEX

IEEE Press Series on: Networks and Services Management

EULA

List of Tables

Chapter 2

Table 2.1
Table 2.2
Table 2.3

Chapter 3

Table 3.1
Table 3.2

Chapter 4

Table 4.1
Table 4.2

Chapter 5

Table 5.1
Table 5.2
Table 5.3

Chapter 7

Table 7.1
Table 7.2

Chapter 8

Table 8.1
Table 8.2
Table 8.3
Table 8.4

Chapter 9

Table 9.1
Table 9.2
Table 9.3
Table 9.4

Chapter 10

Table 10.1
Table 10.2
Table 10.3
Table 10.4

Chapter 11

Table 11.1
Table 11.2
Table 11.3

Chapter 12

Table 12.1
Table 12.2

Chapter 13

Table 13.1
Table 13.2

Chapter 14

Table 14.1
Table 14.2
Table 14.3

Chapter 15

Table 15.1
Table 15.2
Table 15.3

Chapter 16

Table 16.1

Table A.1
Table A.2
Table A.3
Table A.4
Table A.5
Table A.6
Table A.7
Table A.8
Table A.9
Table A.10

List of Illustrations

Chapter 1

Figure 1.1 Growth in capacity demand [7].
Figure 1.2 Capacity scaling with densification for different inter-site distances (ISDs).
Figure 1.3 Transmit power for increasing bandwidth, small cell densities, and inter-site distances (ISDs).
Figure 1.4 Increasing spectral efficiency through beamforming with multiple antennas for different small cell densities, and inter-site distances (ISDs).
Figure 1.5 Dimensions for scaling wireless capacity and achievable future gains at cellular bands.

Chapter 2

Figure 2.1 Horizontal dipole array.
Figure 2.2 (a) Average number of active BSs in the network per square km and (b) average number of active UEs per active BS. The UE densities are 600, 300, and 100 active UEs per km². The UE distributions are uniform and non-uniform within the scenario. The rest of the parameters are s = 1, f = 2 GHz, a = 1 and t = 12 dB.
Figure 2.3 (a) Transmit power per active BS and (b) transmit power of the network per km². The UE densities are 600, 300, and 100 active UEs per km². The UE distributions are uniform and non-uniform within the scenario. The rest of the parameters are s = 1, f = 2 GHz, a = 1, and t = 12 dB. It is important to note that in Fig. 2.3a there are six overlapping curves. This is because the power used by the cell if it is activated does not depend on the UE density.
Figure 2.4 SINR spatial distributions of several ultra-dense small cell deployments with ISD among small cell BSs of 100 and 50 m with or without idle mode capabilities. The rest of the parameters are d = 300 UE/km², ud = 1, s = 1, f = 2 GHz, a = 1 and t = 12 dB. The triangles represent BSs and the squares represent active UEs. (a) 100 ISD and idle mode capability deactivated, (b) 50 ISD and idle mode capability deactivated, (c) 100 ISD and idle mode capability activated and (d) 50 ISD and idle mode capability activated.
Figure 2.5 UE SINR CDF of several ultra-dense small cell deployments with ISD among small cell BSs of 200, 75, 50, 20, 10, and 5 m with or without idle mode capabilities. The rest of the parameters are d = 300 UE/km², ud = 1, s = 1, f = 2 GHz, a = 1 and t = 12 dB. (a) CDF of SINR for different ISDs with idle mode capability deactivated and (b) CDF of SINR for different ISDs with idle mode capability activated.
Figure 2.6 UE SINR CDF in ultra-dense small cell deployments with large BS densities, 35, 10, and 5 m, low UE densities, 100 and 600 active UE per square km and SNR targets, 15, 12, and 9. The rest of the parameters are ud = 1, s = 1, f = 2 GHz, and a = 1. (a) 100 active UEs per km² and (b) 600 active UEs per km².
Figure 2.7 Average (a) and 5th percentile UE (b) throughput for four different network configurations where the carrier frequencies are 2.0, 3.5, 5.0, or 10 GHz. The rest of the parameters are d = 300 UE/km², ud = 1, s = 1, a = 1 and t = 12 dB.
Figure 2.8 Transmit power per active BS (a) and transmit power of the network per km² (b) for different network configurations where the carrier frequencies are 2.0, 3.5, 5.0, or 10 GHz. The rest of the parameters are d = 300 UE/km², ud = 1, s = 1, f = 2 GHz, a = 1 and t = 12 dB.
Figure 2.9 Average (a) and 5th percentile UE (b) throughput for different network configurations where the carrier frequencies are 2.0 or 10 GHz and the number of antennas per BS are 1, 2 or 4. The rest of the parameters are d = 300 UE/km², ud = 1, s = 1, f = 2 GHz and t = 12 dB.
Figure 2.10 Average cell throughput for various ISDs with RR and PF schedulers.
Figure 2.11 CDF of UE throughput for different ISDs.
Figure 2.12 Energy efficiency in bps/W for different network configurations where the carrier frequency is 2.0 GHz, the bandwidth is 20 MHz, and the number of antennas per BS are 1, 2, or 4. The rest of the parameters are d = 300 UE/km², ud = 1, s = 1, f = 2 GHz and t = 12 dB.

Chapter 3

Figure 3.1 Block diagram of a typical fuzzy rule-based system.
Figure 3.2 Steps of an EA scheme.
Figure 3.3 Block diagram of (a) a classical reinforcement learning system and (b) a fuzzy logic assisted reinforcement learning scheme.
Figure 3.4 Block diagram of FARL scheme.
Figure 3.5 (a) Input fuzzy membership functions and (b) output fuzzy membership functions.
Figure 3.6 Average fitness for FARL and ELF schemes.
Figure 3.7 Variance of antenna tilt angles of different BSs over iterations.
Figure 3.8 Example of self-healing function in FARL and ELF.
Figure 3.9 High-level overview of centralized online evolution utilizing generalized network model.
Figure 3.10 Comparison of fitness obtained from the exact model and the generated model.

Chapter 4

Figure 4.1 Frequency allocation options for Macrocells (MC) and Small Cells (SC). (a) Separate frequency bands, (b) Shared frequency bands and (c) Partially shared frequency bands.
Figure 4.2 Coverage probability for un-registered macrocell UEs as a function of the distance from a small cell with closed and open access [2]. (a) Private access and (b) Public access.
Figure 4.3 Fraction of total user demand covered by small cells with open access in an urban environment for operators with different market share [3].
Figure 4.4 Residential femtocell deployment scenario.
Figure 4.5 Floor plan of modeled house (a) and an example of the resulting path gain (b) located at the center of the house.
Figure 4.6 CDF of downlink throughputs.
Figure 4.7 Distributions of downlink throughputs. (a) Example downlink throughput distribution in the simulated area and (b) Example downlink throughput distribution in the area around one femtocell.
Figure 4.8 CDF of uplink throughputs.
Figure 4.9 Distributions of uplink throughputs. (a) Example uplink throughput distribution in the simulated area and (b) Example uplink throughput distribution in the area around one femtocell.
Figure 4.10 Simulation scenario with a single small cell, showing the locations of the macrocell and small cell BSs and the maximum SINR.
Figure 4.11 Fraction of free physical resource blocks in a single small cell scenario for different P_MacroF2 reduction.
Figure 4.12 Number of UEs unable to achieve the 500 Kbps throughput target for different P_MacroF2 reduction.
Figure 4.13 Comparison between global and local power reduction.
Figure 4.14 Example scenario with small cells grouped to their respective macrocells.
Figure 4.15 Load metrics of macrocell/small cell groups during optimization.
Figure 4.16 Number of UEs unable to achieve the 500 Kbps in multiple small cell scenario for different P_MacroF2 reduction.

Chapter 5

Figure 5.1 Residential femtocell deployment scenario and resulting mobility problem.
Figure 5.2 Deployment scenario with eight enterprise femtocell BSs covering an office building.
Figure 5.3 Distance-based power configuration example.
Figure 5.4 Coverage optimization approaches with different antenna configurations [3].
Figure 5.5 Overview of measurement intervals (a) and flowchart of mobility event-based coverage optimization (b) that minimizes mobility events of passing users [1].
Figure 5.6 Antenna circuitry for switched multi-element antenna solution [2].
Figure 5.7 Physical design study of a Femtocell and schematic of patch antenna and IFA location within the housing [3].
Figure 5.8 Resulting measured antenna patterns achievable by selecting either one single antenna or any combination of two antennas [2].
Figure 5.9 Flowchart of the switched multi-element antenna configuration process [2].
Figure 5.10 Antenna circuitry for a tuneable multi-element antenna solution [3].
Figure 5.11 Coverage example for macrocell (a) and femtocell (b). The color-bar on the right shows the received pilot powers for covered areas in dBm. White areas are not covered by the corresponding BS [1].
Figure 5.12 Average number of mobility events per hour for different femtocell locations [1, 2, 3].
Figure 5.13 Average resulting indoor coverage for different femtocell locations [1, 2, 3].
Figure 5.14 CDF of configured pilot powers [1, 2, 3].
Figure 5.15 Snapshot of the update timeline illustrating the length of intervals for the two updates.
Figure 5.16 Algorithm for user load balancing.
Figure 5.17 The algorithm for minimizing coverage holes and overlap.
Figure 5.18 Load balancing performance of the proposed algorithm as a function of different update intervals. ΔP_inc = ΔP_dec = 3 dB.
Figure 5.19 Average pilot transmit power of the whole femtocell network, as a percentage of the maximum allowable pilot transmit power dBm. is set to 500 ms.
Figure 5.20 A snapshot of the radio coverage using the proposed algorithm, with the color-bar on the right showing received pilot channel powers in dBm. White areas are not covered by the small cells.

Chapter 6

Figure 6.1 Macrocell UE jamming the UL of a nearby small cell.
Figure 6.2 Small cell coverage shrinks in the proximity of a macrocell BS leading to poor offloading.
Figure 6.3 Range expansion mitigates UL interference and facilitates offloading at the expense of increasing DL interference.
Figure 6.4 Natural cell boundaries during (a) DL and (b) UL in a scenario with heterogeneous DL transmission powers.
Figure 6.5 (a) Average UL UE throughput; (b) average DL UE throughput; and (c) JFI values in the DL/UL with different cell range expansion bias values.
Figure 6.6 Small cell equal DL RSS boundary (ESB) and equal path-loss boundary (EPB).
Figure 6.7 Semi-axis lengths (s₁ and s₂) of the ESB and (s′₁ and s′₂) of the EPB ellipses, Also shown is the distance d′_{p, e} between the small cell BS and its EPB ellipse center using a typical HetNet scenario.
Figure 6.8 Plot of ESB and EPB ellipses using a typical HetNet scenario. The UE hotspot boundary indicates the area where small cell UEs are distributed, and it is discussed in Section 6.3.4. Even if we expand the small cell coverage from ESB to EPB, due to the shift of the EPB ellipse center from the small cell BS, an macrocell UE may still be closer to the small cell BS than a small cell UE (i.e., d₁ > d₂), causing severe UL interference to the small cell BS.
Figure 6.9 Plot of ΔER^HSB_p required for a targeted small cell DL coverage radius at several macro to small cell BS distances. The vertical green line denotes the 40 m hotspot radius specified in [24].
Figure 6.10 Two design schematics for the switched multi-element antenna system.
Figure 6.11 Simulated scenario. (a) Stachus square in central Munich, Germany where the location of small cell BSs and some of the neighboring macrocell BSs are represented by star and circle markers, respectively and (b) The WiSE simulator prediction representing the strongest received pilot power (in dBm) from small cell BSs.
Figure 6.12 Gain antenna pattern of the antenna elements used in small cell BSs.
Figure 6.13 Misclassification rate versus size of the training set for four different classification methods.
Figure 6.14 Misclassification rate versus size of the training set for four different classification methods when measurement errors are additionally considered.
Figure 6.15 Misclassification rate versus the number of visible cells.
Figure 6.16 Classification performance. (a) Expected misclassification rate, , when rejecting classifications with Cr_u smaller than T′_Cr and (b) Ratio of rejected classifications.
Figure 6.17 Performance gain of MEA scheme over existing ODA systems. (a) CDF of UE SINRs and (b) CDF of UE throughputs.

Chapter 7

Figure 7.1 Concept of fractional frequency reuse. Soft reuse is employed for the downlink while hard reuse is employed for the uplink.
Figure 7.2 Carrier aggregation arrangements for (a) intra-band contiguous allocation, (b) intra-band non-contiguous allocation, and (c) inter-band non-contiguous allocation.
Figure 7.3 Carrier aggregation use cases for (a) Scenario 1—co-located and overlaid cells with nearly identical coverage, (b) Scenario 2—co-located and overlaid cells with different coverage, (c) Scenario 3—co-located cells with different spatial antenna orientation, and (d) Scenario 4—dislocated cells (e.g., remote radio heads).
Figure 7.4 Recent advances in LTE carrier aggregation: (a) Aggregation of supplemental downlink bands, (b) joint aggregation of FDD-TDD carrier components, (c) carrier aggregation of secondary cells (SCell) with licensed assisted access from a primary cell (PCells), and (d) inter-site carrier aggregation (dual connectivity).
Figure 7.5 Dual connectivity architectures where (a) radio bearers terminate individually in MeNodeB and SeNodeB, (b) radio bearers terminate in MeNodeB only and data are forwarded over an RLC bearer to SeNodeB.
Figure 7.6 Principles of PCell and SCell selection.
Figure 7.7 Principles of PCell and SCell channel activity.
Figure 7.8 LTE protocol stack with configured carrier aggregation.
Figure 7.9 Exemplary implementation architectures: (a) eNodeB downlink transmitter for contiguous carrier aggregation, (b) UE uplink transmitter for non-contiguous carrier aggregation.
Figure 7.10 Scheduling options for carrier aggregation: (a) cross-scheduling by UE A, and (b) per-CC scheduling by UEs B and C.
Figure 7.11 Principle of carrier aggregation usage in heterogeneous networks. Significant path loss and reduced macrocell transmit power can be used to protect small cell transmissions in shared carrier components.
Figure 7.12 Carrier aggregation deployment strategies for HetNets availing of two carrier components: (a) Dedicated CCs, (b) full CC reuse, (c) asymmetric CC allocation, and (d) soft CC reuse.
Figure 7.13 Mobility management under carrier aggregation.
Figure 7.14 Performance evaluation of carrier aggregation deployment strategies in HetNets [50].

Chapter 8

Figure 8.1 Metrocell UEs in the expanded region are the victims of macrocell interference.
Figure 8.2 Mobile macrocell UEs are the victims of interference from a femtocell restricting its service to a closed subscriber group.
Figure 8.3 Concept of almost blank subframes.
Figure 8.4 Cumulative distribution function of UE throughputs with (i) RELAXED-ABS eICIC, (ii) no eICIC but with small cells, and (iii) no eICIC and no small cells, assuming dense urban macrocell UE density and 4 W small cell transmit power.
Figure 8.5 CDF of UE throughputs with (i) STOCHASTIC-ABS eICIC, (ii) small cells with no eICIC, and (iii) no small cells.
Figure 8.6 CDF of UE throughputs with STOCHASTIC-ABS eICIC, and with no eICIC for a first metrocell having 128 UEs.
Figure 8.7 CDF of UE throughputs with STOCHASTIC-ABS eICIC, and with no eICIC for a second metrocell having 45 UEs.

Chapter 9

Figure 9.1 Multi-carrier HSPA/LTE HetNet configuration for traditional cell structure with partial frequency reuse (a) and proposed configuration with sector offset (b).
Figure 9.2 Single-carrier LTE HetNet configuration for traditional cell structure with frequency reuse of 1 (a) and proposed configuration with sector offset (b).
Figure 9.3 UE channel quality and resource block allocation over different antennas.
Figure 9.4 Antenna design/implementation. (a) Traditional radome, (b) Sector offset radome and (c) Circular array.
Figure 9.5 RF chain design. (a) Tx RF chain and (b) Rx RF chain.
Figure 9.6 Proposed configuration for 3 sector eNodeB with horizontal-and-vertical sector offset configuration. (a) Different downtilt for carriers within one sector, (b) View from top on theee-sector cell and (c) Resulting cell structure.
Figure 9.7 Example of antenna patterns for the proposed configuration for three-sector eNodeB with horizontal-and-vertical sector offset configuration.
Figure 9.8 Multi-carrier HSPA/LTE HetNet configuration for traditional cell structure with partial frequency reuse (a) and proposed configuration with sector offset (b).
Figure 9.9 Single-carrier LTE HetNet configuration for traditional cell structure with frequency reuse of 1 (a) and proposed configuration with sector offset (b).
Figure 9.10 ABS load balancing algorithm diagram.
Figure 9.11 Sector offset RB load balancing algorithm diagram.
Figure 9.12 RSRP CDF of UEs moving at 30km/h over the CRS.
Figure 9.13 Wideband SINR CDF of UEs moving at 30km/h over the CRS.
Figure 9.14 SINR distribution for the three-sector eNodeB case. (a) Traditional configuration, (b) Offset configuration and (c) Offset + vBF configuration.
Figure 9.15 Spectrum fragment reuse for the three-sector eNodeB case. (a) Offset configuration and (b) Offset + vBF configuration.
Figure 9.16 SINR distribution and spectrum coverage for the three-sector eNodeB case. (a) Three-sector traditional conf. SINR distribution, (b) Three-sector offset conf. SINR distribution and (c) Three-sector offset conf. spectrum coverage.

Chapter 10

Figure 10.1 Exemplary time-frequency alignment of macrocell and small cell resources (identical CFI, subframe offset of one, identical DC carriers, variable bandwidths).
Figure 10.2 System model notation. The letters G^{p, s}, G^{p, m}, ΔG^{p, s}, ΔG^{p, m}, and P^tx_s, P_m^tx denote channel gains and transmit powers, respectively.
Figure 10.3 Outage probability of logical PDCCH candidates of a small cell, sorted on the order of (i) the relative aggregate overlap with the macrocell channels, and (ii) the outage probability itself. HetNet from Section 10.5.
Figure 10.4 Maximum number of simultaneously active macrocell LPCs at aggregation level 8 subject to each active small cell LPC achieving at least SINR_min = −4 dB as a function of the small cell REB. The macrocell uses 6 LPCs at the aggregation level 8.
Figure 10.5 Maximum number of simultaneously active macrocell LPCs at aggregation level 8 subject to each active small cell LPC achieving at least SINR_min = −4 dB as a function of the minimum small cell radius. The macrocell uses 6 LPCs at the aggregation level 8.
Figure 10.6 The average small cell UE capacity per subframe as a function of the offset of the small cell eNodeB from the center of a hotspot in an OFS-configured system and a non-OFS system.

Chapter 11

Figure 11.1 Average spectral radius Λ of the matrix A, and average number N of UEs associated to MC/SC BS as a function of the fraction F of the SC BS power budget allocated to the pilot signal.
Figure 11.2 Example of received pilot CPICH powers in the central part of the simulated network in Dublin, Ireland. SC downlink transmission power is 1 W. Simulated SC BS locations denoted as 1a, 1b, 2a, and 2b.
Figure 11.3 Cumulative probability density of single-UE coupling G_SC◃MC(1), G_MC◃SC(1), and in an MC-SC sub-network having the SC BS placed (i) in the hotspot center (SC BS 1a in Fig. 11.2), and (ii) 30 m off the hotspot center (cell SC 1b in Fig. 11.2).
Figure 11.4 Uplink capacity regions of a co-channel MC-SC sub-network with perfectly synchronized UE transmissions (worst-case scenario) for variable MC-SC BS distance and SC BS output (γ_MC = −15 dB, γ_SC = −12 dB). To indicate the transmit power raise as a function of the cell load, two cases are distinguished: (i) unlimited transmit powers, and (ii) transmit power limit of 23 dBm for 75% of the UE total, and no limit for the remaining 25%.
Figure 11.5 Total capacity loss of the MC-SC cluster due to inter-cell coupling, expressed as the percentage of the load-limited capacity N_SC^max of the small cell (i.e., the double of the “per-cell” loss) and shown as a function of D = E[G_MC(N_MC)]E[G_SC(N_SC)].
Figure 11.6 Uplink capacity regions of a co-channel MC-SC sub-network with perfectly synchronized UE transmissions (worst-case scenario) for variable SC BS offset and indoor coverage (γ_MC = −15 dB, γ_SC = −12 dB). To indicate the transmit power raise as a function of the cell load, two cases are distinguished: (i) unlimited transmit powers, and (ii) transmit power limit of 23 dBm for 75% of the UE total, and no limit for the remaining 25%.
Figure 11.7 Maximum UL target SINR for a co-channel SC for γ_MC = −15 dB assuming perfectly synchronized UE transmissions (worst-case scenario) and maximum power limit of 23 dBm. The SC BS is placed in hotspot center and covers both indoor and outdoor spaces (CPICH sensitivity − 115 dBm).
Figure 11.8 SON architecture for LTE uplink optimization.
Figure 11.9 Network-wide data rate comparison.
Figure 11.10 Macrocell UE data rate comparison.
Figure 11.11 Small cell UE data rate comparison.

Chapter 12

Figure 12.1 The HO stages.
Figure 12.2 The HO procedure and signaling.
Figure 12.3 L1 and L3 filtering procedures. RSRP is measured over one subframe every, for example, 40 ms and recorded as RSRP_L1(l). L1 filtering performs linear averaging over every 200 ms to provide , with RSRP_L1 and M(n) in linear units. Finally, L3 filtering performs averaging over every 200 ms to obtain F(n) = (1 − a)F(n − 1) + a10log ₁₀{M(n)}, where a is the L3 filter coefficient, and M(n) is in linear units and F(n) is in dB. Note that a = 1/2^k/4 and k is specified in Table 12.1.
Figure 12.4 Event A3 entry condition.
Figure 12.5 Timers in radio link monitoring and HO processes. Timer T310 is running while HO_COMMAND is received.
Figure 12.6 Timers in radio link monitoring and HO processes. RLF is declared while TTT timer is running.
Figure 12.7 Timers in radio link monitoring and HO processes. Example of unnecessary HO classified as ping-pong and two other cases, which cannot be classified as ping-pongs.
Figure 12.8 The HOF issue in HetNets.
Figure 12.9 Coverage areas of macrocell/picocell with and without range expansion and RLF locations from macrocell perspective (Q_out = −8 dB).
Figure 12.10 Coverage areas of macrocell/picocell with and without range expansion and RLF locations from picocell perspective (Q_out = −8 dB).
Figure 12.11 Simulated HOF and ping-pong rates with four randomly deployed picocells per macrocell sector, with and without an 8 dB range expansion bias, and ABS. (a) HOF rate and (b) Ping-pong rate.
Figure 12.12 HO parameter optimization based on UE velocity.
Figure 12.13 HOF performance with UE velocity-dependent HO parameter optimization.
Figure 12.14 UE velocity dependent HO.
Figure 12.15 HOF performance with mobility-based inter-cell interference coordination.

Chapter 13

Figure 13.1 Normalized traffic demand map of Wellington, New Zealand.
Figure 13.2 Intra-day variations of the traffic demand.
Figure 13.3 Power consumption model.
Figure 13.4 Percentage of the traffic demand that is served by small cell BSs.
Figure 13.5 Average load of small cell base stations.
Figure 13.6 Total power consumption as a function number of small cell base stations deployed.
Figure 13.7 Percentage of the total power consumed by macrocell base stations.
Figure 13.8 Optimal offload rate of small cells as the function of the α parameter.
Figure 13.9 Total power consumption of the network over the years and for different deployment configurations.
Figure 13.10 Flowchart of distributed idle mode procedure for small cell base stations.
Figure 13.11 Schematic hardware design for a typical femtocell base station.
Figure 13.12 The percentage of reduction in the power consumption, Ω, when using the idle mode.
Figure 13.13 Flowchart of centralized idle mode wake-up procedure for the small cell base stations.
Figure 13.14 Fingerprint measurement path in the city center of Dublin, Ireland.
Figure 13.15 The average misclassification error rate, μ, against the number of RF fingerprints available in the training set.
Figure 13.16 The misclassification performance of k-NN method with different values of k.

Chapter 14

Figure 14.1 High-level backhaul architecture.
Figure 14.2 Point-to-multipoint configuration using NLOS wireless backhaul.
Figure 14.3 Point-to-point line-of-sight wireless backhaul link.
Figure 14.4 Small cell BS connected to POP through an intermediate node.
Figure 14.5 Small cell BS connected to POP through an intermediate node that includes another small cell BS.
Figure 14.6 Specific attenuation by atmospheric gases at sea level.
Figure 14.7 Probability of line of sight between two antennas with heights of 5 m and 32 m, respectively.
Figure 14.8 Geometrical elements for diffraction loss calculation.
Figure 14.9 Coverage areas in an urban environment with LOS and wall reflections. (a) Areas with direct line-of-sight to a rooftop point of presence and (b) Extension of coverage using directional antenna and with wall reflections.
Figure 14.10 Received powers for NLOS point of presence.
Figure 14.11 Maximum antenna gain and half-power beamwidth for circular antenna apertures of different diameters. (a) Maximum antenna gain and (b) Total half-power beamwidth.
Figure 14.12 Illustration of busy time mean and quiet time peak cell throughputs.
Figure 14.13 Total cost of ownership after 5 years of a small cell backhaul with different capacities.
Figure 14.14 Total cost of ownership after 5 years of a small cell backhaul with different distances.

Chapter 15

Figure 15.1 Overview of the small cell placement process.
Figure 15.2 Outdoor small cell deployment scenario.
Figure 15.3 RSRP map of a macrocell (a) and small cell (b) generated using the path-loss models.
Figure 15.4 UE density map, shown in number of UEs per km².
Figure 15.5 Areas with wireless LOS backhaul availability.
Figure 15.6 Cumulative distribution function of backhaul and site leasing costs of candidate small cell sites.
Figure 15.7 Median UE throughput gain per cost of candidate sites calculated at the first greedy algorithm iteration.
Figure 15.8 Cumulative distribution function of UE throughputs for macrocell-only deployment, and with 20 small cells deployed using the greedy algorithm.
Figure 15.9 Cumulative distribution function of cost, and with 20 small cells deployed using the greedy algorithm.
Figure 15.10 Illustration of the disadvantages of greedy algorithm.
Figure 15.11 Example of a crossover operation.
Figure 15.12 Correction performed after crossover operation.
Figure 15.13 Illustrative example of the evaluation performed in one iteration of the steepest ascent approach.
Figure 15.14 Difference in median throughput and cost of S_Gcost and S_GAcost relative to S_Gtput.
Figure 15.15 Map of macrocell throughput, units are in Mbps.
Figure 15.16 Map of areas where macrocell throughput is less than 300 kbps and UE density is more than 600 UEs per km² (white).
Figure 15.17 High offload areas with the S_GAcost deployment with 100 small cells.

Chapter 16

Figure 16.1 Overview of the attocell concept.
Figure 16.2 Illustration of simulated office scenario. Green dots are attocell BS positions, red lines are walls, and white lines are cubicle partitions. The heat map showing the RSCP of a narrowbeam antenna attocell is overlaid.
Figure 16.3 Directional attocell antenna gain patterns.
Figure 16.4 Diagram of the single-element dipole antenna model used for the narrowbeam antenna gain calculation.
Figure 16.5 Mean UE SINR for different attocell antennas and transmit powers.
Figure 16.6 Percentage of UEs offloaded to attocells for different attocell antennas and transmit powers.
Figure 16.7 Gains in throughput achieved with attocells of different antennas and transmit powers, compared with a macrocell-only deployment.
Figure 16.8 UE call outage probabilities for narrowbeam attocell deployment.
Figure 16.9 UE throughput CDF for macrocell-only, femtocell, and 5 mW narrowbeam attocell deployments.
Figure 16.10 Map showing a portion of office scenario, with arrows indicating the UE path used in mobility simulations, and circles indicating attocell locations.
Figure 16.11 UE Layer 3 RSRP measurements moving along the path, with the dashed vertical lines indicating the times where handover occurs. Only measurements from BS the UE hands over to are shown.
Figure 16.12 Cumulative distribution function of SINR experienced by UEs in office simulation scenario with 5 mW narrowbeam attocells.
Figure 16.13 Spectral efficiency of different modulation schemes to achieve 1% BLER.
Figure 16.14 Cumulative distribution function of UE throughput for attocell deployments with up to 256-QAM available, for different carrier frequencies.
Figure 16.15 Cumulative distribution function of SINR experienced by UEs in office simulation scenario with 10 GHz attocells using narrowbeam and omni-directional antennas.

Figure A.1 Homogeneous macrocell layout following the hexagonal grid model.
Figure A.2 Example of different types of antennas. (a) Macrocell antenna, (b) Small cell dipole array and (c) LightRadio.
Figure A.3 Macrocell BS antenna downtilt.
Figure A.4 Macrocell BS antenna pattern and spatial antenna gains. (a) Macrocell BS horizontal and vertical antenna pattern and (b) Macrocell BS spatial antenna gains.
Figure A.5 Dipole array horizontal and vertical antenna pattern (a) and spatial antenna gains (b).
Figure A.6 LightRadio array horizontal and vertical antenna pattern (a) and spatial antenna gains (b).
Figure A.7 Macrocell BS spatial path loss.
Figure A.8 Macrocell BS LOS probability (a) and path gain (b) as a function of distance.
Figure A.9 City of Dublin environment map.
Figure A.10 Transmission losses for different materials.
Figure A.11 Example of path gains using an explicit wall model based on transmission losses for a residential environment.
Figure A.12 Auto- and cross-correlated shadow fading map (a) and its statistics (b).
Figure A.13 Multi-path fading map.
Figure A.14 City of Dublin received signal strength map of one urban macrocell BS sector.
Figure A.15 City of Dublin SINR map of an hexagonal macrocell layout with D_ISD = 800 m and 1 tier of interferers with and without small cells. (a) Macrocell-only scenario and (b) Heterogeneous network (HetNet) scenario.
Figure A.16 LUT for channel performance assessment in HSPA and LTE systems. (a) Receiver SINR to throughput in HSPA and (b) Receiver SINR to bits per data symbol in LTE.
Figure A.17 Generated routes over the city of Dublin.
Figure A.18 Waypoints for mobility model in residential scenario.
Figure A.19 Probability of the time a UE spends at an area in the residential scenario.
Figure A.20 Waypoints for mobility model in enterprise scenario.

ABOUT THE AUTHORS

Dr. Holger Claussen is the leader of the Small Cells Research Department of Nokia Bell Labs located in Ireland and the United States. In this role, he and his team are innovating in all areas related to future evolution, deployment, and operation of small cell networks to enable exponential growth in mobile data traffic. His research in this domain has been commercialized in Nokia’s (formerly Alcatel-Lucent’s) Small Cell product portfolio and continues to have significant impact. He received the 2014 World Technology Award in the individual category Communications Technology for innovative work of the “greatest likely long-term significance”. Prior to this, Holger was the head of the Autonomous Networks and Systems Research Department at Bell Labs Ireland, where he directed research in the area of self-managing networks to enable the first large-scale femtocell deployments from 2009 onward. Holger joined Bell Labs in 2004, where he began his research in the areas of network optimization, cellular architectures, and improving energy efficiency of networks. Holger received his Ph.D. degree in signal processing for digital communications from the University of Edinburgh, UK, in 2004. He is the author of more than 100 publications and 120 filed patent families. He is a fellow of the World Technology Network, a senior member of the IEEE, and a member of the IET. Dr. David López-Pérez is a member of technical staff in the Small Cells Research Department of Nokia Bell Labs in Ireland. Prior to this, David received B.Sc. and M.Sc. degrees in telecommunication from Miguel Hernandez University, Spain, in 2003 and 2006, respectively, and Ph.D. degree in wireless networking from the University of Bedfordshire, UK, in 2011. David was also an RF engineer with Vodafone, Spain, from 2005 to 2006, and a research associate with King's College London, UK, from 2010 to 2011. David has authored the book Heterogeneous Cellular Networks: Theory, Simulation and Deployment (Cambridge University Press, 2012), as well as more than 100 book chapters, journal articles, and conference papers, all in recognized venues. He also holds more than 35 patent applications. David received the Ph.D. Marie-Curie Fellowship in 2007 and the IEEE ComSoc Best Young Professional Industry Award in 2016. He was also a finalist for the Scientist of the Year prize in the Irish Laboratory Awards in 2013 and 2015. He is an editor of IEEE Transactions on Wireless Communications since 2016 and he was awarded Exemplary Reviewer of IEEE Communications Letters in 2011. He is or has also been a guest editor of a number of journals, for example, IEEE Journal on Selected Areas in Communications and IEEE Communication Magazine.

Dr. Lester Ho is a distinguished member of technical staff in the Small Cells Research department of Nokia Bell Labs in Ireland. He obtained his B.Eng. degree in electronic engineering in 1999, and his Ph.D. degree on self-organizing wireless networks in 2003, both from the University of London. He joined Bell Labs, which was then part of Lucent Technologies, in 2003, where he performed research in various topics in wireless communications, particularly in small cells, self-organizing networks, and network optimization techniques, many of which can be found in commercial deployments today. He has over 40 patents granted, as well as more than 40 publications in journals, conference papers, and book chapters. He is a senior member of the IEEE, and the recipient of the Eckermann-TJA Prize in 2008, and a finalist for the Scientist of the Year prize in the Irish Laboratory Awards in 2014.

Dr. Rouzbeh Razavi received a master’s degree with distinction in information systems and a Ph.D. in computer science, both from the University of Essex, UK. He is currently a faculty member in the Department of Management and Information Systems at the Kent State University, OH, USA. Prior to this, Dr. Razavi was a director in the Group Decision Sciences at the Commonwealth Bank of Australia in New York. Before joining the Commonwealth Bank of Australia, Rouzbeh served as a senior research data scientist at SAP and as a member of technical staff in the Small Cells Research Department of Nokia Bell Labs in Ireland for several years. He has been supervising a number of Ph.D. students and post-doctoral researchers and has published more than 60 technical papers in peer-reviewed journals and conferences. In addition, he has authored five book chapters and filed more than 30 patent applications. Rouzbeh is a senior member of the IEEE and recipient of best paper awards from two international conferences.

Dr. Stepan Kucera is a member of technical staff in the Small Cells Research Department of Nokia Bell Labs in Ireland. In this role, he serves as a principal investigator and technical lead of multiple projects aiming to create novel disruptive technologies for wireless networking with current focus on multi-connectivity and gigabit wireless technologies. His expertise lies mainly in the area of wireless and IP networking technology, including both proprietary solutions and 3GPP/IEEE/IETF standards. Stepan has filed more than 50 patents and published over 30 book chapters, transactions, and conference papers in peer-reviewed IEEE venues. He is also the (co)-recipient of several professional awards. Between 2008 and 2011, he was a research scientist at the New Generation Wireless Communications Research Center in the Keihanna Research Laboratories, NICT, Japan. He received his Ph.D. degree in informatics from the Graduate School of Informatics at Kyoto University, Kyoto, Japan, in 2008. He is a senior member of the IEEE, and actively serves on technical boards of major IEEE journals and conferences.

Giancarlo Fortino	Xiaoou Li	Ray Perez
Dmitry Goldgof	Andreas Molisch	Linda Shafer
Don Heirman	Saeid Nahavandi	Mohammad Shahidehpour
Ekram Hossain	Jeffrey Nanzer	Zidong Wang

SMALL CELL
NETWORKS

Deployment,
Management, and
Optimization

ABOUT THE AUTHORS