d1 7 Final Project Report

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT
SEVENTH FRAMEWORK PROGRAMME (FP7)

THEME 3
Information and Communication Technologies
FINAL PROJECT REPORT D1.7
Project acronym:
Project full title:
Grant agreement no.:
HARP
High capacity network Architecture with Remote
radio heads & Parasitic antenna arrays
318489
Version number: 1
Date of preparation of final project report (latest version): 16/11/2015
Date of approval of scientific report by Commission: ......................................
Authors:
Tharmalingam Ratnarajah, Faheem Khan, Jiang Xue, Fahd Khan, Sudip Biswas, Paula Aquilina, Lin Zhou,
S. Razavi, Constantinos Papadias, Konstantinos Ntougias, Dimitrios Ntaikos, Bobby Gizas, David
Gesbert, Laura Cottatellucci, Haifan Yin, Bruno Clerckx, Borzoo Rassouli, Yueping Wu, Ralf Muller,
Mohamamd Ali Sedaghat, Lars Dittmann, Henrik Christiansen, Matteo Artuso, Bjarne Skak Bossen,
Aleksandra Checko, Laurent Roullet, Francois Taburet, Dora Boviz, Stephane Senecal, Jean-Marc Kelif
Author Contact Information

THARMALINGAM RATNARAJAH, UEDIN, T. RATNARAJAH@ED.AC.UK
FAHEEM KHAN, UEDIN, FAHEEM.KHAN@ED.AC.UK
JIANG XUE, UEDIN, J.XUE@ED.AC.UK
FAHD KHAN, UEDIN, F.KHAN@ED.AC.UK
SUDIP BISWAS, UEDIN, SUDIP.BISWAS@SMS.ED.AC.UK
PAULA AQUILINA, UEDIN, P.ACQUILINA@ED.AC.UK
LIN ZHOU, UEDIN, L.ZHOU@ED.AC.UK
S. RAZAVI, UEDIN, MORTEZA.RAZAVI@ED.AC.UK
CONSTANTINOS PAPADIAS, AIT, CPAP@AIT.GR
KONSTANTINOS NTOUGIAS, AIT, KONTOU@AIT.GR
DIMITRIOS NTAIKOS, AIT, DINT@AIT.GR
BOBBY GIZAS, AIT, BOGI@AIT.GR
DAVID GESBERT, EUR, DAVID.GESBERTS@EURECOM.FR
LAURA COTTATELLUCCI, EUR, LAURA.COTTATELLUCCI@EURECOM.FR
HAIFAN YIN, EUR, HAIFAN.YIN@EURECOM.FR
BRUNO CLERCKX, IMPERIAL, B.CLERCKX@IMPERIAL.AC.UK
BORZOO RASSOULI, IMPERIAL, B.RASSOULI12@IMPERIAL.AC.UK
YUPING WU, IMPERIAL, YUEPING.WU@IMPERIAL.AC.UK
RALF MLLER, NTNU, RALF.MUELLER@LNT.DE
MOHAMMAD ALI SEDAGHAT, NTNU, MOHAMMAD.SEDAGHAT@IET.NTNU.NO
LARS DITTMANN, DTU, LADIT@FOTONIK.DTU.DK
HENRIK CHRISTIANSEN, DTU, HLCH@FOTONIK.DTU.DK
MATTEO ARTUSO, DTU, MATART@FOTONIK.DTU.DK
BJARNE SKAK BOSSEN, RADIOCOMP, BJARNE.BOSSESN@MTIGROUP.COM
ALEKSANDRA CHECKO, RADIOCOMP, ALEKSANDRA.CHECKO@MTIGROUP.COM
LAURENT ROULLET, ALBLF, LAURENT.ROULLET@ALCATEL-LUCENT.COM
FRANCOIS TABURET, ALBLF, FRANCOIS.TABURET@ALCATEL-LUCENT.COM
DORA BOVIZ, ALBLF, DORA.BOVIZ@ALCATEL-LUCENT.COM
STEPHANE SENECAL, ORANGE, STEPHANE.SENECAL@ORANGE.COM
JEAN-MARC KELIF, ORANGE, JEANMARC.KELIF@ORANGE.COM
Table of Contents
1. Final Publishable Summary report ...................................................................................................... 4
1.1 HARP Vision ................................................................................................................................... 4
1.2 Work Packages .............................................................................................................................. 7
1.3 Executive Summary ........................................................................................................................... 8
1.3.1 Technical Highlights ................................................................................................................... 8
1.3.2 Key findings .............................................................................................................................. 10
1.3.3 Collaboration in the HARP Project ........................................................................................... 12
1.3.4 Summary of Year III Conclusions.............................................................................................. 12
1.4 Technical Work................................................................................................................................ 14
1.4.1 Network Requirements and Fundamental Limits (WP3 lead by Orange) M1-M36 .............. 14
1.4.2 Multi-Point ESPAR Enabled RRH Access (WP5 lead by Imperial) M13-M30 ............. 16
1.4.3 Aggregation Network (WP6 lead by DTU) M9-M33 ..................................................... 23
1.4.4 Experimentation and Demonstration (WP7 lead by ALU) M13-M36 ......................... 31
2. Dissemination and Use of Foreground ............................................................................................. 48
2.1 Dissemination Activities ....................................................................................................... 48
2.1.1 Journal Papers ................................................................................................................. 49
2.1.2 Conference Papers ............................................................................................................. 50
2.1.3 Industry White Papers ....................................................................................................... 55
2.1.4 Papers Submitted or in Preparation ................................................................................. 55
2.1.5 Multi-partner joint publications ....................................................................................... 57
2.1.6 Invited Talks, Tutorials, Poster Presentations ................................................................. 59
2.1.7 Workshops .......................................................................................................................... 61
2.1.8 Special Sessions .................................................................................................................. 61
2.2 Use of foregrounds: Exploitable foreground and plan for exploitation ..................................... 62
2.2.1 Education ............................................................................................................................ 62
2.2.2 Description of exploitable foreground................................................................................. 64
2.2.3 Path towards standardization ........................................................................................... 69
3. Report on societal implications ......................................................................................................... 71
1. Final Publishable Summary report

1.1 HARP Vision
The current state of the art of dense radio networks follows two common assumptions.
From the antenna design and the air interface point of view, multiple radio frequency (RF)chains are required to drive the available antenna elements that should also retain a large
enough inter-element spacing in order to provide the necessary fading properties.
From the network infrastructure point of view, the intelligence that coordinates the
transmission among adjacent cells could be located either in the central processing unit or at
the individual base station itself.
The HARP project envisions to overcome these two limitations by combining advantageously two
recently emerged technology innovations:
Remote radio heads (RRHs)
RRHs are compact radio units that are connected with a central base station via optical fibres and have
been recently proposed as a way to share multiple geographically dispersed sets of wireless users.
RRH reduce substantially the network cost and also allow the implementation of a distributed
multiple-input multiple-output (MIMO) transmission, thus improving network throughput.
Single RF compact array radiators
This term refers to antenna arrays with very small inter-element spacing that are fed with only a single
RF chain. The apparent cost, hardware and size savings make those lightweight arrays an ideal
candidate for RRHs radio equipment. This entails an important increase of the network throughput
without making the base stations bulkier and without increasing the cost and implementation
complexity.
An overview of the architecture that benefits from the combination of these two approaches is shown
in Figure 1. HARPs objectives can be listed as follows:
To define an end-to-end architecture, including both wireless access and backhaul network,
based on RRHs and single-RF compact arrays that can deliver substantially higher over-the air
capacity.
To explore the fundamental limits (bandwidth / power efficiency) of distributed multi-antenna

wireless access networks as the one shown in Figure 1.
To develop efficient transceiver techniques (both single and multi-user) for single-RF compact
arrays that enable them to emulate spatial multiplexing / beamforming / diversity
communication, which comply with the requirements of the LTE Advanced and related next
generation wireless systems
To develop appropriate channel estimation and channel feedback techniques, as required for
the developed transceiver techniques and as dictated by the distributed mobile access
network parameters. This will be an important catalyst for the envisioned intelligent
interference.
To extend the above-mentioned transceiver and channel feedback techniques to

corresponding cooperative transmission and channel sharing techniques for efficient
distributed wireless access.
To define, model and optimise the aggregation network between the RRH-based wireless
access nodes and the central processing stations (building upon the existing common public
radio interface (CPRI) & OBSAI protocol structures) so as to best serve the requirements in
both data traffic and channel distribution of the developed wireless access techniques.
To demonstrate the key HARP concept by building an end-to-end demonstration that will
include both LTE-compliant base stations and prototyped electronically steerable parasitic
antenna radiator (ESPAR)-equipped radio heads, thus showcasing the validity of the approach
over the air towards the end user.
Figure 1: HARPs ESPAR-based RRH wireless network architecture.

The Consortium partners are:
University of Edinburgh (UEDIN) (UK, Co-ordinator)
Athens Information Technology (AIT) (Greece, Technical Leader)
Eurecom (EUR) (France)
Imperial College London (IMPERIAL) (UK)
Norwegian University of Science and Technology (NTNU) (Norway)
Technical University Denmark (DTU) (Denmark)
MTI RadioComp (RADIOCOMP) (Denmark)
Alcatel-Bell-Lucent France (ALBLF) (France)
Orange Labs (ORANGE) (France)
1.2 Work Packages

HARPs work plan consists of seven inter-connecting WPs, which follow a natural progression allowing
for research and development, focussed on the spectrum and interference management, intelligent
signal processing and information theoretic study of IA networks. The high level WPs for the project
are:
WP1: Management of the Project (Leader UEDIN)
WP2: Dissemination Activities and Exploitation Plan (Leader MTI RADIOCOMP)
WP3: Network requirements and Fundamental Limits (Leader ORANGE)
WP4: Single-point ESPAR-enabled RRH Access (Leader AIT)
WP5: Multi-point ESPAR-enabled RRH Access (Leader IMPERIAL)
WP6: Aggregation Network (Leader DTU)
WP7: Experimentation and Demonstrations (Leader ALBLF)
1.3 Executive Summary

This document details the main scientific results obtained throughout the Year III of the HARP project.
In order to guide the reader regarding the core contents of the document, we provide herein a short
executive summary, containing the technical highlights of the Year III of the project. Then, an overview
of the main scientific results obtained throughout the Year III follows.
1.3.1 Technical Highlights

The core technical components of the document with respect to the HARP projects description of
work are listed below:
WP3 - Network Requirements and Fundamental Limits: During Y3 of the project, only T3.2 was
ongoing.
T3.2 Fundamental Limits
In this Task, we investigate the fundamental limits of the ESPAR-enabled multi-antenna systems
proposed in the project. The main focus of our work in T3.2 during Y3 of the project was on the
investigation of the performance limits of single-RF ESPAR systems with bounded sum-power and
constant sum-power constraints.
WP5 - Multi-point ESPAR-enabled RRH Access: During Y3 of the project, T5.2 and T5.3 were ongoing.
T5.2 Cooperative Techniques
In this Task, we address the problem of applying cooperative techniques and evaluating the associated
gains when ESPAR-enabled remote radio heads (RRH) are utilized. In Y3 of HARP,
we devised a novel beamforming-based approach that enables us to perform arbitrary

channel-dependent precoding with single-RF ESPAR antennas. This techniques is based on the
corresponding signal model that was developed in Y2 of HARP.
we studied the application of this solution in multi-cell Cloud-RAN-based ESPAR-enabled

systems. A radio transmission protocol was devised for that purpose and several multi-cell
coordinated transmission variants were investigated.
a beamforming design for MISO broadcast channels that exploits both statistical and delayed
CSI has been described and evaluated.
a rate splitting approach for the case where CSI is imperfect was studied.
user-centric interference nulling in multi-antenna heterogeneous networks was investigated.
T5.3 Cloud-RAN Implementation of Cooperative Techniques

The role of this Task is to describe how the ESPAR-based cooperative techniques for RRH-enabled
wireless access can be implemented on a Cloud-RAN setup. In Y3 of the HARP project, the
implementation of user-selective user detection was investigated. Namely, we focused on the actual
implementation in Cloud-RAN of uplink and downlink cooperative techniques selected among the
ones defined in Task 5.2.
WP6 - Aggregation Network: During Y3, T6.2-T6.4 were ongoing.
T6.2 Aggregation Network Optimisation
In this Task, we investigate the optimisation of the aggregation network in order to meet the
requirements specified by the HARP project. During Y3 of the project,
we studied both through simulations and using testbeds the use of protocols and techniques
that will allow a packet-based aggregation network to meet the stringed inter-cell cooperation
delay requirements of Cloud-RAN setups that employ coordinated transmission.
we investigated the affect of specific aggregation network dimensioning approaches to

backhaul capacity as well as the delay requirements for ensuring user-satisfaction.
T6.3 Protocol Extensions Design and Implementation

The role of this Task is the development of protocols for the aggregation network. By the end of Y2,
the end-to-end protocol stack was defined and implemented in OPNET (using MATLAB for the PHY
layer software implementation). In Y3 of the project, the focus was mainly on studying the
performance of higher-layer protocols that run on the developed protocol stack.
T6.4 Cloud-RAN for Multi-cell Network Aggregation
In this Task, we investigate how a Cloud-RAN could be integrated in the RRH aggregation network. The
focus is mainly on the design and implementation of an Eth2CPRI gateway which will bring together
the CPRI-based mobile fronthaul with the Ethernet-based aggregation network. In Y3, we evaluated
Ethernet based fronthaul network characteristics, defined communication protocol betweek network
front-end and BBU-pool and realised implementation of Eth2CPRI gateway based on our previous
results obtained in this task.
WP7 Experimentation and Demonstration: During Y3, all Tasks of this WP were ongoing.
T7.1 ESPAR Prototyping
In this Task, ESPAR design and fabrication is considered. The implementation should take into account
the transmission techniques that were developed during the HARP project as well as the setup of the
final over-the-air demo. In Y3 of the project, we designed, fabricated and tested load-controlled
parasitic antenna arrays that can be used to apply the beamforming-based cooperative arbitrary
channel-precoding technique devised in the same year. The design of this antenna array meets the
implementation requirements of the final demo.
T7.2 Radiohead Prototyping
This task is devoted on radiohead prototyping for the ESPAR-enabled Cloud-RAN setup envisioned in
HARP. During the project, we opted for a solution based on a commercial 4G RRH and we emphasized
on the integration of this wireless access component with the ESPAR-based transceiver technologies
developed in HARP. In Y3 of the project, we have developed a replacement of the planned RRH
prototype made of custom equipments (PC and USRP board) that we call the RRH gateway using the
OAI code base. It can receive IQ samples from the cloud based eNB and forward them to the USRP
board. The end to end validation has been performed with the possibility to connect a 4G phone. This
RRH gateway offers a 3rd option to the fronthaul (in addition to the direct USB3 and the 10GbE
interfacing the CPRI gateway).
T7.3 Verification of the Aggregation Network Protocols
This Task focuses on the design, implementation and verification of the protocols used in the
aggregation network. In Y3 of the project, we implemented several possible designs that came up in
the study of 6.3 in the end-to-end network model and we then performed a comparative analysis of
their performances. As a result, we then chose the protocol configuration that was able to ensure the
best performances for the applications running in the UEs.
T7.4 System Integration and Final Demo
This Task focuses on the integration of the various system components (ESPAR-based transmitters,
RRH nodes, Eth2CPRI gateways, Cloud-RAN, OpenAir receivers etc.) and the over-the-air (OTA)
demonstration of the end-to-end system. In Y3 of the project, we have built an end to end
demonstrator that is composed of the following elements: RRH prototypes with ESPAPR antennas, CSI
based downlink cooperative beamforming, uplink user selective joint detection and Cloud-RAN
integration with network and protocol simulator.
1.3.2 Key findings

The key findings of the research work conducted in the Year III of the HARP project are summarized
as follows:
Results regarding the optimum (capacity-achieving) input signal format in single-RF MIMO
systems (point-to-point and multiple access) with bounded or constant sum-power have been
derived.
Asymptotic capacity analysis for the case where the number of antennas grows towards
infinity has been conducted for the aforementioned setups when the optimum input signal
format is applied.
The optimal input distribution for 22 MIMO systems with arbitrary channel matrix has also
been derived.
Mapping strategies to construct PMH signals have been proposed.
A novel beamforming-based technique for performing arbitrary channel-dependent

precoding with single-RF ESPAR antennas has been proposed.
A radio transmission protocol that exploits the aforementioned precoding technique to

facilitate the application of multi-cell coordinated transmission in single-RF ESPAR-enabled
Cloud-RAN setups has been also proposed.
Several variants (including reduced feedback overhead methods) of transmission schemes

that incorporate the aforementioned beamforming-based precoding and radio protocol have
been described and evaluated in terms of the achieved sum-rate in a single-RF ESPAR-enabled
Cloud-RAN setup through numerical simulations.
Linear beamforming design and power allocation strategies for ergodic rate optimization in
MISO broadcast channels with statistical and delayed CSIT have been proposed and their
performance gains have been evaluated through numerical simulations.
The performance of rate-splitting (RS) techniques in MIMO wireless networks with imperfect
CSIT has been analyzed and evaluated through numerical simulations.
The optimum RS precoding design for the aforementioned setup has been studied and
evaluated through numerical simulations.
A hierarchical RS scheme that exploits a two-tier precoding structure to mitigate multiuser

interference in massive MIMO systems has been proposed and evaluated through numerical
simulations.
A user-centric interference nulling (IN) scheme for the downlink of two-tier HetNets that can
improve network performance by improving each users SIR has been proposed. Analysis and
simulations focused on the outage probability and the asymptotic DoFs of the system.
User-selective joint detection for the uplink has been studied.
The use of IEEE 1588 for aggregation network synchronization has been studied.
A source scheduling algorithm for the aggregation network has been proposed.
The effect of different functional splits on MFH capacity has been studied.
The effect of MFH dimensioning at application-level satisfaction has been quantified.
TCP performance has been studied in the context of the developed end-to-end protocol stack.
ESPAR-enabled precoding-based joint Zero-Forcing Beamforming has proven to outperform

the omni-directional setup over the entire SNR range.
Load-controlled parasitic antenna arrays that support the application of the developed
cooperative transmission techniques and meets the implementation requirements of the final
OTA demo have been designed, fabricated and tested.
A replacement of the planned RRH prototype made of custom equipments that uses the OAI
code has been developed.
OTA transmission with and without CSI feedback has been demonstrated.
An end-to-end over the air system has been implemented with a beam selection optimization
based on downlink CQI using the RRH prototype.
1.3.3 Collaboration in the HARP Project

The requirements of the project have lead naturally to collaboration in one form or another between
various partners in several levels (intra-WP, intra-Task; intra-WP, inter-Task; and inter-WP). This is
indicated, for example, by the joint journals and conference papers (see D2.1.3).
1.3.4 Summary of Year III Conclusions

In Y3 of the HARP project, several theoretical and practical contributions on every aspect of the
considered research study have emerged. More specifically,
Results on optimal signal design and corresponding performance evaluation for single-RF
multi-antenna systems with bounded and constant sum-power have been presented.
A novel precoding technique for single-RF ESPAR antennas has been devised.
A radio communication protocol for performing multi-cell cooperative precoding transmission

in single-RF ESPAR-enabled Cloud-RAN setups has been proposed.
Several variants of multi-cell coordinated transmission schemes (including reduced feedback

overhead methods) incorporating the aforementioned precoding technique and radio
communication protocol have been described and evaluated.
Design and performance results of limited beamforming and power allocation schemes for
MISO broadcast channels with statistical and delayed CSIT have been provided.
Results on rate-splitting (RS) performance in MIMO wireless networks with imperfect CSIT
have been derived.
The optimum RS precoder design has been presented and evaluated.
A hierarchical RS scheme for massive MIMO setups has been proposed and evaluated.
A user-centric interference nulling (IN) technique for the downlink of HetNets has been
proposed and evaluated.
User-selective joint detection schemes have been implemented in a Cloud-RAN prototype.
The use of IEEE 1588 for aggregation network synchronization has been evaluated through
both numerical simulations and lab testbed-based experimentation.
A source scheduling algorithm for the aggregation network has been proposed and evaluated.
Results on the effect of various functional split schemes on the MFH capacity have been
provided.
The effect of dimensioning on application-level latency penalties has been studied.
TCP performance over the devised end-to-end protocol stack has been evaluated.
ESPAR-enabled precoding-based joint Zero-Forcing Beamforming has been shown to

outperform the omni-directional setup over the entire SNR range due to the power gain
related to the application of transmit beamforming.
Load-controlled parasitic antenna arrays for applying the proposed coordinated transmission
/ precoding techniques and for using in the final OTA demo have been designed, fabricated
and tested.
A replacement of the planned RRH prototype made of custom equipments that uses the OAI
code has been developed.
OTA transmission with and without CSI feedback has been demonstrated.
An end-to-end over-the-air system has been implemented with a beam selection optimization
based on downlink CQI using the RRH prototype.
1.4 Technical Work

1.4.1 Network Requirements and Fundamental Limits (WP3
lead by Orange) M1-M36
Fundamental Limits (Task 3.2 lead by NTNU) M4-M36
The standard implementation of multiple-antenna transmitters uses one RF-chain per transmit
antenna. This means that the power on each antenna is generated by a separated power amplifier
and the peak to average power ratio of the signal on each antenna determines the required amount
of back-off and linearity for the power amplifiers. Power amplifiers usually have a maximum power
which is determined by the saturated power level. This power constraint should be considered in
addition to the average power constraint which limits the amount of power radiated into the air. In
classical multi-antenna transmitters, the maximum power constraint appears per antenna branch.
This limitation has been considered in [1] for real signals on single-antenna and in [2] for complex
signals in order to investigate fundamental information theoretic limits. In single-RF multiple-antenna
transmitters, the story is totally different. There is a central power amplifier which feeds all the
antenna elements. Therefore, the peak to average ratio of the sum power matters. The central power
amplifier has higher power efficiency if the Peak to Average Sum Power Ratio (PASPR) is lower. The
maximum power limitation appears for the sum power and not for every antenna branch. This basic
difference in multiple-antenna systems with single-RF transmitters creates some new fundamental
limits which had not been addressed and we have addressed some of them in HARP.
During this project, we have mainly investigated two different problems. The first one is to analyze
MIMO systems with bounded sum power. The second problem is about MIMO systems with constant
sum power. In the following subsections, these two problems are explained and a summary of results
is given.
MIMO systems with bounded sum power
In single-RF MIMO systems, the central power amplifier has a maximum power which should be taken
into account. Lets consider a point to point MIMO link with
receive antenna elements with the following channel model
M transmit antenna elements and N
y = Hx + n,
where y is the receive vector,
is the channel matrix between the transmit antenna elements and
the receive antenna elements, x is the input vector and nis the additive white Gaussian noise vector.
In single-RF multiple-antenna transmitters, we have a limitation on the sum power as
x x P.
We are interested to find the channel capacity of such a channel. The simplest case is when the
channel matrix is a unitary matrix and
given in the following theorem.
M=N. This case is investigated in [3] and the main result is
Theorem: The channel capacity is obtained by the input vector distributed uniformly on discrete and
finite number of sphere shells.
The higher the SNR of the channel, the larger the number of the sphere shells. The above theorem
states that the optimal distribution is uniform in phases and discrete in amplitude. The result is
generalized form of the results in [1] and [2]. For very low SNR, one sphere shell is enough to achieve
the capacity.
MIMO systems with constant sum power
Next, we explain the results in the case that the sum power is constant. In this case we have
x x = P.
This is motivated by the result of bounded sum power. In fact, it is observed that for typical values of
SNR, the input vector distributed on one sphere shell achieves a large portion of the channel capacity.
We call the class of signals with fixed norm as Phase Modulation on the Hypersphere (PMH) signals.
We have analyzed PMH in different MIMO channels. The derivation and analysis are given in [4], [5]
and a summary of the results is presented here.
First, the system is analyzed when the channel matrix is identity or unitary. We have shown in [4] that
the capacity is acheived by uniformly distributed PMH and the channel capacity is plotted in Figure 2
compared to the normalized channel capacity of the same channel with average power constraint. It
is shown that by increasing the number of antennas, the channel capacity becomes closer to the
mutual information of Gaussian input.
Figure 2: The normalized channel capacity of PMH in channels with unitary matrices versus SNR.
Next, we investigated the mutual information of uniform PMH in a multiple access MIMO channel. To
this end, the uplink channel of a wireless system is considered. The analysis is done using the replica
method from statistical physics [4]. Lets assume that the number of antennas at the base station and
the aggregated number of antennas at the user terminals go to infinity and they are equal. M is
assumed to be the number of antennas at each user terminal. The channel is assumed to be iid
Gaussian channel. The mutual information versus SNR is plotted in Figure 3. Once again, it is observed
that increasing the number of antennas makes PMH very attractive in terms of channel capacity. Few
antennas are enough to obtain mutual information very close to the mutual information of Gaussian
input.
Figure 3: The mutual information of PMH in a multiple-access iid Gaussian channel for different
number of antenna elements at the user terminals.
We have also analyzed 2 by 2 MIMO channels with arbitrary channel matrix (see [5] for more details).
In this case, it has been shown that the optimal distribution is some discrete points on the circle. The
number of points is finite and depends on SNR and the channel matrix.
Next, we have also proposed some mapping strategies to construct PMH in [4]. The first option is to
distribute some points on the hypersphere randomly with maximum possible distance. This can be
done using spherical codes which is discussed in [6]. Spherical codes are not known for many
dimensions. Therefore, we proposed a new mapping method in [4] which is called Minimum Shift
Keying on the Hypersphere (MSKH). In MSKH, the mapping is with memory and the symbols are not
independent. Finally, we have also proposed a new spectral shaping method called spherical filtering.
Spherical filtering keeps the symbols on the hypersphere in order to fix PASPR to 0dB [7].
1.4.2 Multi-Point ESPAR Enabled RRH Access (WP5 lead by Imperial) M13M30
Cooperative Techniques (Task 5.2 lead by Imperial) M13-M30
Several techniques have been investigated to deal with multi-user / multi-cell transmission, consider
both cases of perfect and imperfect channel state information at the transmitter. The latter term
refers, for instance, to the case where delayed, quantized or statistical information about the CSI is
known to the transmitter.
Cooperative Multi-Cell Beamforming in ESPAR-enabled Cloud-RAN Setups with perfect CSIT
Arbitrary Precoding Utilizing Single-RF ESPAR Antennas
From an antenna perspective, the signal model of a
( M, N) MIMO link is [8]
y = Hi + n ,
where y is the vector of open-loop voltages at the receive antennas, i is the vector of currents that
run on the transmit antennas, H is the channel matrix that relates the input currents with the output
voltages, and n is zero-mean circularly symmetric complex Gaussian (ZMCSCG) additive noise vector
{ }
with covariance matrix R n = E nn = n2I .

On the other hand, the conventional signal model for closed-loop MIMO transmission is
y = HWs + n,
where W is the precoding matrix and s is the transmit signal vector. Thus, by mapping the precoded
symbols to the antenna currents, that is, by setting [9]
i = Ws,
we can perform channel-depending precoding with single-RF ESPAR antennas. For that purpose, we
have first to calculate the required currents for the desired precoding design and input signal
according to i = Ws and then compute the corresponding loading values that will generate these
currents according to the generalized Ohms law [8]
1
i = ( Z + Z L ) v,
where Zis the mutual coupling matrix, Z L is a diagonal matrix that holds the source resistance and the
impedances of the tunable loads, and
is a vector that hods the feeding voltage.
The computed loads should lie within a reasonable range of values. Moreover, the following design
condition should be met [10]:
Re{Zin } > 0
This is because if the input resistance is negative, the ESPAR antenna feeds back power and, therefore,
exhibits unstable behavior. The input resistance depends on the antenna currents (that is, on the input
signal) and, therefore, on the corresponding loading values and on the mutual coupling matrix
(which, however, is fixed for a given array geometry).
Unfortunatelly, this design condition cannot be met for any given input constellation. On the other
hand, it is well-known that adaptive transmit beamforming with single-RF ESPARs admits any input
signal format. That is because the radiation pattern (and, therefore, the currents and the
corresponding loading values required to shape the desired beam) does not depend on the signal to
be transmitted.
Based on that remark, we have devised a novel beamforming-based approach which enables us to
overcome the aforementioned implementation challenges and perform arbitrary channel-dependent
precoding with single-RF ESPARs. This method consists of two steps:
1. First, we perform beamforming using any valid method.
2. Then, we perform precoding-based transmission over the beams.
This technique is universal, i.e., it applies on any input signal format (and, in fact, on any type of
antenna array technology as well).
Radio Transmission Protocol for Multi-Cell Cooperative Beamforming with ESPAR Antennas
Now, consider the downlink transmission in a Cloud-RAN setup where there are K RRHs equipped
with single-RF ESPAR antennas and K single-antenna UE. At each timeslot (TS), each RRH wishes to
serve its UE. All transmissions take place simultaneously and over the same frequency band. Each RRH
is able to generate at each TS one out of Ldistinct predetermined beams, corresponding to a set of
fixed loading configurations. Thus, there are L beam combinations in total. RRHs can cooperate (i.e.,
exchange control information) due to the Cloud-RAN architecture, but the UEs operate independently
from each other.
System operation is divided in three phases [11]:
1. Learning phase: Each UE returns to its RRH either (a) its SINR or (b) its estimates of the
direct and cross channels.
2. Beam-selection phase: (a) If SINR values have been fed back to the RRHs by the UEs, then
K
the beam combination that results in the maximum sum-rate throughput R = Rk is

k =1
selected, where
Rk = log2 (1+ SINRk ) is the data rate of
U E k and SINR k is its SINR
( k =1,2,K,K) . (b) If channel estimates have been fed back to the RRHs by the UEs, then
the RRHs exchange the channel estimations to compile the composite ( K K ) channel
matrix
H for each beam combination. Then, the metric T = Tr ( HH )1
is calculated and
the beam combination with the minimum metric value is selected.

3. Transmission phase: Non-precoding-based or precoding-based transmission over the
selected beams takes place. In the latter case, the RRHs perform jointly zero-forcing
beamforming (ZFBF).
Transmission Variants and Numerical Results
In Figure 4 is illustrated the average sum-rate throughput achieved by using various transmission
schemes in the aforementioned setup with K = 2 and
1000 channel realizations in a target receive SNR range
L=4. These curves have been obtained over
[ 20,30] dB. A single-bounce scattering model
has been incorporated in the simulation program to capture the effects of radiation beams and small-
/large-scale fading. The beams have been generated with the help of antenna design software for a
load-controlled parasitic antenna array with 5 elements. The half-power beam width (HPBW) is 30 .
Moreover, the channel matrices have been normalized appropriately to facilitate performance
evaluation.
Figure 4: Average sum-rathe throughput over SNR for cooperative multi-cell techniques in ESPARenabled Cloud-RAN setups.
The following transmission schemes have been considered:
1. Non-precoding-based transmission. In this case, beam-selection is based on SINR
feedback.
2. Reduced feedback overhead precoding-based transmission. In this case, beam-selection
is based on SINR feedback. For the selected beam combination, the UEs report to the RRHs
their channel estimates, in order for joint ZFBF precoding to be performed.
3. Standard multi-cell beamforming transmission. In this case, beam-selection is based on
CSI.
Also, for comparison purposes, it has been considered an equivalent single-RF scenario where each
RRHs is equipped with a single antenna and joint ZFBF transmission takes place.
We note that:
The curves related with precoding-based transmission schemes do not floor, since the
application of ZFBF results in null inter-cell interference (ICI).
The ESPAR-enabled setup performs better than the omni-directional-antennas setup over the
entire SNR range due to the power gain related with the application of transmit beamforming.
Precoding-based transmission in which beam-selection is CSI-based outperforms significantly

the case where beam-selection is SINR-based, since in the former case beam-selection is ZFBFdriven while in the latter case the procedures of beam-selection and precoder design are
decoupled.
MISO Broadcast Channels with Statistical and Delayed CSIT

As a follow-up to year 2 activities in MISO BC with delayed CSIT, some improvements have been
performed by integrating the knowledge of delayed CSIT and statistical CSIT. This work focuses on
linear beamforming design and power allocation strategy for ergodic rate optimization in a two-user
Multiple-Input Single-Output (MISO) system with statistical and delayed channel state information at
the transmitter (CSIT). We propose a transmission strategy, denoted as Statistical Alternative MAT
(SAMAT), which exploits both channel statistics and delayed CSIT. Firstly, with statistical CSIT only, we
focus on statistical beamforming (SBF) design that maximizes a lower bound on the ergodic sum-rate.
Secondly, relying on both statistical and delayed CSIT, an iterative algorithm is proposed to compute
the precoding vectors of Alternative MAT (AMAT), originally proposed by Yang et al., which maximizes
an approximation of the ergodic sum-rate with equal power allocation. Finally, via proper power
allocation, the SAMAT framework is proposed to softly bridge between SBF and AMAT for an arbitrary
number of transmit antennas and signal-to-noise ratio (SNR). A necessary condition for the power
allocation optimization is identified from the Karush-Kuhn-Tucker (KKT) conditions. The optimum
power allocation to maximize an ergodic sum-rate approximation is computed using Sequential
Quadratic Programming (SQP). Simulation results show that the proposed SAMAT scheme yields a
significant sum-rate enhancement over both SBF and AMAT. More information is available in [12].
A Rate-Splitting Approach in MIMO wireless networks with imperfect CSIT
A class of Rate-Splitting (RS) approaches has been proposed recently, which divides one receivers
message into a common and a private part, and superposes the common message on top of precoded
private messages.
Rate Analysis of rate-splitting strategies
To enhance the multiplexing gain of two-receiver Multiple-Input-Single-Output Broadcast Channel
with imperfect channel state information at the transmitter (CSIT), a Rate-Splitting approach is
analyzed. With quantized CSIT, we study the ergodic sum rate of two schemes, namely RS-S and RSST, where the common message(s) are transmitted via a space and space-time design, respectively.
Firstly, we upper-bound the sum rate loss incurred by each scheme relative to Zero-Forcing
Beamforming (ZFBF) with perfect CSIT. Secondly, we show that, to maintain a constant sum rate loss,
RS-S scheme enables a feedback overhead reduction over ZFBF with quantized CSIT. Such reduction
scales logarithmically with the constant rate loss at high Signal-to-Noise-Ratio (SNR). We also find that,
compared to RS-S scheme, RS-ST scheme offers a further feedback overhead reduction that scales
with the discrepancy between the feedback overhead employed by the two receivers when there are
alternating receiver-specific feedback qualities. Finally, simulation results show that both schemes
offer a significant SNR gain over conventional single-user/multiuser mode switching when the
feedback overhead is fixed. More information is available in [13].
Robust beamforming for rate-splitting

This work focuses on robustly designing the RS precoders. However, the actual sum-rate cannot be
considered as an optimization metric at the BS due to the CSI uncertainty. Alternatively, precoders are
designed such that the Average SR (ASR) is maximized, where averaging is taken over the error
distribution. The ASR problem is transformed into a closely related Average Weighted Sum Mean
Square Error (AWSMSE) problem, solved using an Alternating Optimization (AO) algorithm which
converges to a stationary point. Numerical results show that in addition to the gains in the high SNR
regime, the benefits of incorporating a common symbol also include reduced CSI quality requirements
and enhanced achievable rate regions compared to conventional MU transmission. More information
is available in [14].
Two-tier rate splitting for Massive MIMO
Massive multiple-input multiple-output (MIMO) is a key enabling technology for 5G cellular systems,
offering substantial benefits in energy and spectral efficiency. The major challenge to realize those
benefits lies in imperfect channel state information at the transmitter (CSIT). In a multiuser MIMO
broadcast channel, the rate performance is limited by multiuser interference due to the imperfection
in CSIT. By clustering users into groups based on their channel covariance matrices, we propose a
hierarchical rate splitting (RS) scheme that exploits a two-tier precoding structure to mitigate
multiuser interference. The inner-tier RS mitigates intra-group interference by packing part of one
users message into a common codeword that can be decoded by users in that group. The outer-tier
RS can overcome inter-group interference by further packing part of one users message into a
common codeword that is be decodable by all users. Then, the private messages are superimposed
over the common messages and broadcast with linear precoding. We analyse the asymptotic rate of
hierarchical RS (HRS) and optimize the precoder of the common messages. We also derive a closedform power allocation solution that provides insights into the effect of system parameters. Finally,
simulation results validate the significant gain of HRS over various state-of-the-art baselines. More
information is available in [15].
User-Centric Interference Nulling in Downlink Multi-Antenna Heterogeneous Networks
Distributed remote radio heads are subject to strong interference due to spectrum reuse. This affects
the signal-to-interference ratio (SIR) of each user, and hence is one of the limiting factors of network
performance. However, in previous works, interference management approaches in HetNets are
mainly based on interference level, and thus cannot effectively utilize the limited resource to improve
network performance. In this work, we propose a user-centric interference nulling (IN) scheme in
downlink two-tier HetNets to improve network performance by improving each users SIR. This
scheme has three design parameters: the maximum degree of freedom for IN (IN DoF), and the IN
thresholds for the macro and pico users, respectively. Using tools from stochastic geometry, we first
obtain a tractable expression of the coverage (equivalently outage) probability. Then, we characterize
the asymptotic behavior of the outage probability in the high reliability regime. The asymptotic results
show that the maximum IN DoF can affect the order gain of the asymptotic outage probability, while
the IN thresholds only affect the coefficient of the asymptotic outage probability. Moreover, we show
that the IN scheme can linearly improve the outage performance, and characterize the optimal
maximum IN DoF which minimizes the asymptotic outage probability. More information is available
in [16], [17].
Cloud-RAN Implementation of Cooperative Techniques (Task 5.3 lead by ALBLF)

M13-M30
Cooperative Downlink Transmission with ESPAR
Using Cloud-RAN architecture, we can efficiently control information (e.g., channel state information
(CSI), signal-to-interference-plus-noise (SINR) values, power allocation vectors, precoding matrices
etc.) in the BBU-pool. We consider a setup with cells, one active single-antenna user per cell and
one RRH equipped with a single-RF ESPAR antenna. At each timeslot, each BS wishes to serve its user.
Each RRH is able to generate distinct predetermined beams, corresponding to fixed sets of loading
values. Thus, there are possible beam combinations (-tuples of beams) in total. K simultaneous
transmissions take place over the same frequency band.
We defined a three-phase transmission protocol. In first phase, we perform learning of direct and
cross link channels or SINR of every UE in the BBU-pool through feedback. Based on this information,
in the second phase, BBU-pool select jointly the best set of beams. Finally joint transmission based on
precoding takes place.
Figure 5: Equivalent circuit diafram of a single-fed ESPAR antenna

The system setup for the considered radio transmission protocol is shown in Figure 5 for = 2 and
= 4.
Our results in this setup prove that ESPAR-enabled precoding-based joint Zero-Forcing Beamforming
outperforms the omni-directional setup over the entire SNR range due to the power gain related to
the application of transmit beamforming.
User-selective Joint Detection

We implemented in our Cloud-RAN prototype with co-located BBUs uplink joint detection for selected
users. This cooperative scheme improves spectral efficiency and guarantees QoS of critical users in
case of network congestion in an area covered by several antennas, when severe interference for celledge users cannot be avoided by Coordinated Scheduling. Giving access to this feature to a subset of
users only reduces fronthaul bandwidth usage, since for high-bandwidth MIMO transmission available
fronthaul infrastructure would not allow forwarding I/Q symbols of all users.
Obviously, joint detection improves spectral efficiency, since multiple users (one/cell for each JD PRB
set) are sharing the same PRBs. As one can observe in Figure 6, energy efficiency is also improved with
respect to single user transmission.
Our implementation benefits from BBU co-localisation to share received data from every cell in the
cooperation cluster with a low latency. In practice, data from subframes allocated for JD is stored in a
shared memory segment of the BBU-pool server, so that any BBU can get needed data when needed.
Each BBU does the part of joint processing for the user associated to its own cell, which makes BBUpool architecture scalable and reduces processing time through parallel computing in BBUs.
Figure 6: Required overall SNR for achieving 10-1 symbol error rate during the joint detection and
improvement of energy efficiency with respect to the 1 user/PRB case.
1.4.3 Aggregation Network (WP6 lead by DTU) M9-M33

Aggregation Network Optimisation (Task 6.2 lead by RADIOCOMP) M13M27
The overall network architecture considered here is a Cloud-RAN-based mobile network equipped
with load-controlled parasitic antenna arrays. Cloud-RAN implies separating the baseband and the RF
and then pooling all baseband processing. The main area of investigation in WP6 has been the network
interconnecting the Remote Radio Heads (RRHs) and the baseband proccessing pool. This network is
here denoted as the RRH aggregation network. Traditionally, the RRH aggregation network would
consist of direct fibers running the CPRI protocol to each RRH. In HARP, however, we move to a packet
switched, Ethernet-based RRH aggregation network as shown in Figure 7.
Figure 7: The HARP network illustrating the Ethernet-based fronthaul.

The benefit of using a packet based RRH aggregation network is flexibility. Capacity in the aggregation
network can be shared among RRHs and thus the shift of capacity between different cells the socalled tidal effect can be exploited.
However, the packet based approach also implies challenges. These are briefly outlined below. More
details can be found in D6.2 and D6.3.
1. Synchronization
Achieving high bit rates in mobile networks requires inter-cell cooperation which in turn
increases the requirements to synchronization of the base stations. This, however, is going
against the trends in the fronthaul/backhaul where synchronization is being taken out of the
technologies used (e.g., Ethernet). Obviously, something must be done, and this is to put some
synchronization mechanism back in to compensate for jitter in the packet switched network.
This can also be used to deliver the clock to the RRHs.
2. Protocols
Protocols are required to control the information exchange in networks. There is a bunch of
protocols running on top of the RRH aggregation network. However, they were all designed
to be running in the basestation (eNB in the case of LTE). In the Cloud-RAN architecture where
processing is moved away from the RF, additional delay is introduced and this will impact
protocol and application behavior. So the challenge here is to investigate if the Cloud-RAN
architectures imposes any limitations to the network and if that is the case, if they can be
alleviated by proper protocol design / configuration.
3. Dimensioning
Finding the required RRH aggregation network capacity is crucial for a well-functioning radio
access network. The capacity obviously depends on the number of RRHs attached to the
fronthaul but also on the so-called functional split, i.e., which functionality to place in the
baseband pool and in the RRH, respectively, impacts the required capacity. With a packet
based RRH aggregation network we get a lot of flexibility in terms of resource sharing. The
challenge is how to exploit this flexibility without sacrificing high bit rates and low latency
from a user point of view.
These issues along with the key findings will be elaborated on subsequently. The elements (1) and (3)
are dicussed in this section, whereas (2) is studied in the next Section which is about T6.3.
Synchronization
Albeit being very flexible, Ethernet is far from ideal for use in fronthaul networks. The challenge is
keeping the RRHs synchronous despite jitter occurring in the packet based fronthaul. Figure 9 presents
an overview of synchronization requirements end-to-end. What was investigated here was ways of
mitigating this problem. This was investigated theoretically using simulation as well as experimentally
using a lab setup (see Figure 8). Both indicated that the required level of synchronization is achievable.
Figure 8: Implemented lab setup used to find the delay measurements shows to the right.

3ms Round Trip Time latency allowed, 100-400us RTT left for fronthaul link
1588
Slave
or
GPS
1588
Transparent Clock
1588
Master
Switch
RRH
Radio, LTE-A TDD
LTE
CPRI2Eth
GW
CPRI
LTE
CPRI
CPRI
IQ
1588
Frequency error
requirement:
50ppb (Wide Area BS).
250ppb (Home BS).
Cell phase
synchronisation
requirement:
5s between to
overlapping cells
65ns between MIMO
signals
R-18: Maximum jitter

contribution from CPRI
link to RRH: 2ppb.
R-20: Round trip absolute
accuracy: 16.276ns.
1588
Performance affected by
Variable queueing delay
Timestamping error: +1ns/
+4ns
1588
Ethernet
CPRI Data
1588 packets
IQ
1588
Eth
IQ Data
BBU
Pool
Note: switch may not

support 1588, leading to
bigger timing recovery
errors
Ethernet clock accuracy

defined as 100ppm.
Realistic values are 1ppm.
Eth
Eth
Inhibiting factors in nodes:

Local Oscillator phase &
frequency noise, up to 100ppm
Queueing in Eth switches
Figure 9: Overview of synchronization in the various domains and a mapping from RF level
requiremnts to fronthaul transport synchronization requirements.
Mainly two things were investigated:
Using IEEE 1588 for synchronization. It was verified using simulation that the requiements
could be met.
Using source scheduling and preemption. There is currently a large focus on this within the
Ethernet community. The basic idea is to avoid delay variation by eliminate the need for
queueing in the intermediate switches. We have proposed a source scheduling algorithm
especially for use in packet based fronthaul.
Dimensioning
At the end of the day what matters is the application level delay experienced by the users of mobile
terminals attached to the network. The huge benefit of using a packet based fronthaul is that statistical
multiplxing can be exploited. However, this is only possible in case more functionality is moved to the
RRHs. The traditional functional split will generally cause a constant bitrate between the BBU and RRH,
i.e. no multiplexing gain in the fronthaul. However, other functional splits load the fronthaul
proportionally to the cell load, meaning that the tidal effect can now be exploited to also reduce the
requied fronthaul capacity. This is interesting because the required fronthaul capacity directly impacts
the overall cost of the network.
One of the most interesting findings was that the fronthaul latency originating from the physical size
of the network should be kept below 0.1 ms in order to avoid severe penalties (delays) at application
level.
Protocol Extensions Design and Implementation (Task 6.3 lead by DTU)

M16-M32
One of the major achievements in WP6 was the development of a full, end-to-end protocol model of
the HARP network. This model, depicted in Figure 10, can be used for investigations/verification of
protocol modifications and extensions. A few highlights of the model:
Built using OPNET modeler.
Full LTE protocol stack, covers physical- to application layer protocols.
Wireless physical layer processing modelled in MATLAB.
Realistic processing delays in the BBU pool extracted from OpenAirInterface.
Figure 10: The comprehensive protocol model encompassing the entire HARP network.
The issue here is that the entire fronthaul network is in between BBU and the end user. These are the
two devices that terminate the RLC and MAC protocols. These two protocols can do retransmission
which is governed by a set of timers, i.e., they will be impacted by a change in fronthaul latency, which
again changes with the physical size of the network. One investigation carried out was to look into
TCP, which is running on top of RLC/MAC and which can also do retransmissions.
TCP is interesting because a lot of widely used applications (e.g., web browsing) are using it and hence
delay at TCP level directly means delay imposed to the end users applications. The optimal TCP
settings depend on the properties of the network and the interesting thing is that they can be adjusted
from the end users devices, i.e., in this case the mobile phones.
Figure 11: Example results showing the end-to-end application delay for vasious TCP settings.
Cloud-RAN for Multi-Cell Network Aggregation (Task 6.4 lead by ALU)

M19-M33
Ethernet fronthaul evaluation
In this study we performed measurements on ethernet fronhaul connection in order to evaluate
various aspects of this one:
Frequency error comparison between Distributed RAN and Cloud-RAN setup including CPRIto-Ethernet Gateway. We have highlighted the necessity of thight synchronization between
network front-end and central processing unit.
Delay and queueing measurements on Ethernet based fronthaul network containing several
switches. We used the laboratory setup represented in the following Figure 12.
1588
Slave
1588
Master
Eth
RRH
BBU
Pool
Measurements
CPRI2Eth
Eth
switch
SDN
Figure 12: Laboratory setup for measuring delay and queueing on Ethernet-based fronthaul
network.
The results of these measurements are included in Deliverable 6.4. They provide a detailed
understanding of the behavior of Ethernet fronthaul network and allowed us to design network
protocol consequently.
Definition of Radio over Ethernet protocol
This protocol defines two aspects of data transfer between the RRH and the BBU-pool:
The split in the physical layer between network front-end and BBU-pool.
Transport of I/Q samples in Ethernet packets.
We have several possibilities to perform different stages of uplink and downlink physical layer
processing either in the front-end components (RRH, CPRI-to-Ethernet Gateway) or in the BBU-pool.
The following Figure 13 shows different interfaces (PHY layer) where this split can take place.
Figure 13: Different interfaces with splits in the physical layer
As indicated on Figure 13 above, depending on split interface the required fronthaul bandwidth
changes. Scaling values are understood with respect to time-domain I/Q samples transmission
(interface 4).
Radio-over-Ethernet protocol that we have defined allows establishement of communication and data
transfer over the fronthaul interface as represented on Figure 14 below.
NEW C-RAN FRONTHAUL: E-RRH

Central Office
New Ethernet RRH
Metrocell
clock
Small Cell
Wi-Fi
Small cell
Backhaul
EPC
LTE Fronthaul
LTE Backhaul
RADIO OVER ETHERNET
S1
EPC
Ethernet RRH
at Macrocell site
BBU
at Central Office
Figure 14: Implementation of CPRI-to-Ethernet Gateway

We have completed the implementation of CPRI-to-Ethernet gateway interfacing the RRH and the
fronthaul Ethernet network which connects it to the BBU-pool. To realize this indispensible element
of our final validation prototype we needed to solve several challenges.
Sychronization between the CPRI-to-Ethernet Gateway and the BBU-pool. We realized the
setup of the Figure 15 below to meet network requirements.
A
L
M
1588 slave A
C FPGA
P
Software BBU
B
el
l
Clock
1588
Reference
Figure 15: Setup for sychronization between the CPRI-to-Ethernet Gateway and the BBU-pool
Implementation of Xilinx FPGA platform realizing low-delay forwarding. We used Xilinx ZC706
board, Zinq System-on-Chip (FPGA+800 MHz dual core arm processor). Its characteristics
allow high performance execution of functional blocks included in the gateway.
1.4.4 Experimentation and Demonstration (WP7 lead by ALU) M13-M36

ESPAR Prototyping (Task 7.1 lead by AIT) M13-M30
Introduction
Electronically Steerable Parasitic Antenna Radiator (ESPAR) antennas have a sole active element that
is fed by the radio frequency (RF) unit and a number of parasitic elements placed in close vicinity to
that element. The close inter-element spacing results in strong electromagnetic coupling among the
antennas. As a consequence, the feeding voltage induces currents to the parasitics, thus enabling
them to participate in the formation of the far-field radiation pattern. The parasitic elements are
terminated to tunable analog loads. By adjusting electronically the parasitic loads with the help of a
low-cost control circuit, we can vary the magnitude and phase of the currents that run on the parasitic
elements. In other words, we can exploit the strong mutual coupling among the antenna elements we
can perform beam-shaping / beam-steering in an adaptive and controllable manner.
Description of the Antenna
In our design, the active element resonates at 2.6 GHz while its input resistance is matched to the RF
generator (50 Ohm in our case for maximum power transfer). The parasitics are terminated to either
inductive or capacitive loads.
The proposed ESPAR antenna consists of five patches which form a cross. It is built on commercially
available FR-4 substrate material with dielectric constant r = 4.4 and loss tangent tan = 0.0002 . The
front side of the board is etched with the active and the four parasitic patches, while the back side is
the common ground. Each square patch is 26mm by 26mm. The distance between all patches is 2mm.
The active element (which is at the center) is fed through feeding point. It has to be noted that in order
for the active element to have input impedance equal to 50 Ohms, the feeding point has to be offcentered downwards by 8mm. The overall FR-4 board dimensions are 100mm by 100mm. The ESPAR
antenna design is presented in Figure 16 and Figure 17. The proposed antenna design is driven by the
needs of the HARP demo. During this demo, two RRHs will be used. Each one will be connected to an
ESPAR antenna. Every RRH will switch between three predefined beams. In the demo, the cooperating
RRHs will each choose one of the predefined beams so that the chosen beam combination maximizes
the sum rate of the links to their intended receivers, based on SINR feedback. Furthermore, it is
possible to precode over the chosen beams, based on the fed-back channels seen at the receivers
through the beams.
Figure 16: ESPAR antenna design.
Figure 17: ESPAR antenna design (tilted view) where hole positions are shown.
Simulation Results
The ESPAR antenna resonates at 2.6GHz. Its reflection coefficient versus frequency is presented in
Figure 18.
Figure 18: Reflection coefficient of the proposed ESPAR antenna.

As mentioned earlier, each parasitic element has to be terminated with an inductive or reactive load
of the form Z = R + jX , where
R is the ohmic part of the impedance and represents the energy loss
and X is the imaginary part and represents the phase shift. This energy loss and phase shift are in
turn used to change the amplitude and phase of the reflected waves, thus leading to beam
manipulation (reshaping and/or beam steering). In our proposed design the carefully selected load
values are presented in the following Table Table 1: ESPAR parasitic elements load termination values..
Table 1: ESPAR parasitic elements load termination values.
Zvalue
Load 1
Load 2
Load 3
Load 4
50 j122, 427
50 j 68 , 015
50 j76, 517
50 j68, 015
In order to calculate the capacitive reactance we use the following formula:

X = 1
( 2 fC ) ,
where f is the resonant frequency (in Hz) and C is the capacity (in Farad). The resulting values for the
termination capacitors are presented in Table Table 2.
Table 2: ESPAR parasitic elements termination capacitor values.
X value
C1
0,5 pF
C2
0,9 pF
C3
0,8 pF
C4
0,9 pF
The final version of the ESPAR antenna has been implemented using varactors as termination loads,
since they can be easily controlled using a Digital-to-Analog (DAC AD5504) module.
Figure 19, Figure 20 and Figure 21 present the E-plane, the H-plane and the 3D form of the far field
radiation pattern, respectively, as they have been simulated using a full wave EM analysis software. It
is clear that there is one main lobe, while all other side lobes have been suppressed. This lobe has a 3dB bandwidth of about 35 , thus ensuring very low interference with neighboring lobes (high
isolation). Its gain was found to be of the order of 12dB, while its radiation efficiency was about 0.49,
which is a very typical value for planar via fed patch antennas.
Figure 19: E-plane far field radiation pattern.
Figure 20: H-plane far field radiation pattern.
Figure 21: 3D form of the far field radiation pattern.
Load Control Circuit

In order to be able to change the values of the four loads (varactors) we implement a load control
circuit, based on a DAC module. Its schematic block diagram is presented in Figure 22. By changing the
applied voltage at each varactor, its capacitance changes accordingly, thus leading to different values
of reactive impedance. If needed, we can switch from inductive to capacitive resistance load values.
This can be implemented with the use of lumped LC loads. Provided that the varactor has sufficient
range, it can either be set to its minimum value, thus leading to inductive loads, or to its maximum
value, thus cancelling out the effects of the inductor and leading to capacitive load values.
Figure 22: Schematic block diagram of the proposed load control circuit.
Additional Load-Controlled Parasitic Antenna Array Solution
Since the last EuCNC meeting in Paris (June 2015), we experienced quite a few problems with the
voltage-controled capacitors. Namely, it has been proven that these components are extremely
sensitive to slight voltage variations, thus making very hard to control their exact capacity value with
high presicion. Being unable to precisely set the desired capacity value for each of the four capacitors
connected the parasitic elements, led to the unpleasant situation of not being able to steer/reshape
the far field radiation pattern as planned.
Trying to find ways to overcome this situation we ended up with an alternative solution, which is
equivalent in terms of functionality with the original one. More specifically, we built three parasitic
arrays in three different FR-4 boards. These three boards are placed at a 45 degree angle one next to
the other as presented in Figure 23. The parasitic elements of each antenna board are connected to a
set of fixed loads. These different load configurations, in conjuction with the physical displacement of
the board, generate three different beam patterns. By switching to a different antenna board each
time, we select the radiation beam pattern that we want to use. At each time, only a single antenna
board is active, that is, a single-RF is utilized.
This alternative solution which enables us to overcome the aforementioned implementation

challenges has been selected for two reasons: (a) It enables us to apply the beamforming-based
precoding scheme that has been devised in Y3 of the project; and (b) and meets the needs of the final
OTA demo.
Figure 23: Back and front view of the layout of the antennas.
The generated beams of the above layout are presented in Figure 24.
Figure 24: The generated beams according to simulation program.
Radiohead Prototyping (Task 7.2 lead by RADIOCOMP) M13-M30

Antenna Controller
The controller is the main interface between the antenna and the rest of the system. Its role is to
control the properties of the radiation pattern. Depending on the implementation, this is done by (1)
tuning the reactive loads of the antennas parasitic elements or (2) by selecting a desirable beam angle
using RF switches.
Initially, the controller architecture was planned to follow the first approach, as described in T7.1.
However, due to implementation challenges and given the new emphasis on the arbitrary precoding
over pre-selected beams which emerged in Y3 of the project, we opted eventually a more robust
method for producing the beams in the HARP final demo. We have chosen to use microstrip parasitic
antenna arrays in a pre-defined geometry configuration with each parasitic element terminated to
fixed loads. By switching to a different antenna board, we are able to perform beam selection. Hence,
the properties of the antenna system do not deviate from the initial plan, since we still use one RF
input and achieve the proposed radiation characteristics, while at the same time this architecture
serves more efficiently the needs of the final HARP demo.
The complete antenna system consists of a number of hardware modules, including a microcontroller
unit, an Ethernet controller, an RF switch and fixed loads. Among these, the Microcontroller Unit
(MCU) is the core module that provides the functionality to control the antenna radiating properties
through a secure shell connection (SSH).
A number of MCUs were investigated including the STM32F4 from ST Microelectronics but since most
evaluation boards including the aforementioned module do not come with an Ethernet controller, we
had decided to build our system around a raspberry pi board. This readily available solution runs a
Linux-based operating system (OS), is Ethernet enabled and is compatible with many interfaces such
as SPI, I2C, etc.
The complete system consists of three antenna sectors placed in such a way that is capable of
producing directive beams in the desired angles of 45 , 0 and 45 degrees (see T7.1). The direction
of a beam is selected via a highly isolated, low insertion loss 4-port RF switch which is controlled
through a raspberry pi. The antenna control board is shown in Figure 23 (left).
The integration of the antenna into the final end to end system is simple enough, since it only requires
two connections. The RF signal from the external antenna connector (EAC) port is fed to the RF input
of the switch and the controller (RPi) is connected to the 1Gbit Ethernet switch. Once the raspberry pi
module acquires a static IP, the beam can be controlled through an SSH session.
The end-to-end system architecture is shown in Figure 25.
Figure 25: End-to-end System Architecture with Load-Controlled Parasitic Antennas Arrays
RRH prototype description
We have developed a replacement of the planned RRH prototype made of custom equipments (PC
and USRP boards) that we call the RRH gateway using the OAI code base. It can receive IQ samples
from the cloud based eNB and forward them to the USRP board. The end to end validation has been
performed with the possibility to connect a 4G phone. This RRH gateway offers a 3rd option to the
fronthaul (in addition to the direct USB3 and the 10GbE interfacing the CPRI gateway). These options
are used in the final demonstration.
An ESPAR antenna has been developed on a Raspberry Pi chassis that can control the beam selection
via a SSH command interface.
We have then developed a log collector that connects to the eNB to retrieve the CSI and an analytics
engine that can process them.
Figure 26: Overview of RRH prototype elements
Verification of Aggregation Network Protocols (Task 7.3 lead by DTU)

M21-M32
To verify aggregation network protocols, we implemented several possible designs that came up in
the study of Task 6.3 in the end-to-end network model and we then performed a comparative analysis
of their performances. As a result, we then chose the protocol configuration that was able to ensure
the best performances for the applications running in the UEs.
System Integration and Final Demo (Task 7.4 lead by ALU) M25-M36
Parasitic Antenna Arrays

The cross-like patch parasitic antenna array design has been described in T7.1. The antenna design
was driven by the needs of the HARP demo. For this intermediate demo, two of these antennas were
used in order to demonstrate CSI-based precoding over preselected beams.
WARP Platform
For the purpose of setting up a MIMO testbed to demonstrate techniques of precoding over arbitrary
beams we used the WARP v3 boards [18]. WARP stands for Wireless Open-Access Research Platform.
It is a scalable and extensible programmable wireless platform, built from the ground up to prototype
advanced wireless networks. The WARP v3 modules are the latest generation of WARP hardware,
integrating a high performance FPGA, two flexible RF interfaces and multiple peripherals to facilitate
rapid prototyping of custom wireless designs.
The central controller consists of a single host PC, which uses MATLAB to send data and control
commands to the radio modules. WARP conveniently provides an open-source MATLAB-based
framework called WARPLab, which allows MATLAB to control and configure the WARP boards and
process transmit and receive data samples. This baseline framework is used for rapid physical layer
prototyping that allows the coordination of arbitrary combinations of single and multi-antenna
transmit and receive nodes. The extensible framework gives users the flexibility to develop and deploy
large arrays of nodes to meet any application or research need.
CSI-based Transmission over Preselected Beams: Estimation, Feedback and Precoding
For the CSI-based precoding testbed we use two WARP modules, having one configured as a two port
transmitter node and the other as the receiver node with the same number of ports ( 22 MIMO). A
parasitic antenna array was connected at each of the four ports.
The operation of the system is divided into four phases:
1. CSI Estimation
2. CSI Feedback
3. Precoding and Transmission
4. Data Analysis and Conclusions
The final setup is configured as a 22 MIMO OFDM network with two of the antennas transmitting
two independent spatial streams of data at the same time and in the same frequency band while two
antennas at the receiver are used to disentangle the two spatial streams.
The antennas are connected to various sets of fixed load configurations. By switching to a different
set of loading values, a different beam combination is generated at the transmit antennas,
corresponding to a different 22 beam-channel
H.
To accomplish channel estimation, explicit MIMO channel training is added to the preamble of the
transmission. The second antenna transmits during the first burst of Short Training Symbols (STS) in
order to ensure that proper gains are selected at the receive node radio interfaces. To avoid
beamforming accidentally the transmission to an arbitrary direction, the STS symbols transmitted
from the second antenna are cyclically shifted by a user-changeable number of samples. Long Training
Symbols (LTS) are used by the receiver to estimate and correct for Carrier Frequency Offset (CFO) as
well for timing alignment via a cross correlator.
Figure 27: MIMO channel training sequence.

Finally, the preamble was concluded with time-orthogonalized transmission of channel training
symbols from each of the two transmit antennas. These two symbols were used by the receiver to
generate the 22 channel matrix for each of the data-bearing subcarriers of the OFDM waveform.
The channel for each selected beam pair is estimated through least-squares (LS) estimation. Denoting
the training sequence vector as p, the received signal vector in the training phase is given by
y = Hp
Then, the channel matrix can be estimated as follows:
H = yp ( pp )
In general, receiver CSI feedback has to be quantized and fed back to the transmitter but in this case
and since the two WARP modules are controlled through the same PC/Controller, we provide channel
state information feedback directly back to the transmitter through MATLAB code.
Figure 28: MIMO setup and signal model.

Based on the CSI feedback, the transmitters perform jointly zero-forcing beamforming (ZFBF)
precoding. The precoding matrix
W is calculated as follows:
F = H + = H ( HH )
W=
F (:, k )
F (:, k )
, k = 1, 2
The second step has to do with power normalization purposes.

Thus, the transmitted signal becomes s = Ws and the signal model is given by
= HWs
.
y = Hs
Note that we have assumed perfect CSI at the transmitter (CSIT), so that H = H . This is a realistic
assumption for the considered setup.
Under this assumption, ZFBF diagonalizes the channel matrix and nulls co-channel interference (CCI).
The effective channel matrix is given by
= Q,
H eff = HW
where Q is a scaled version of the identity matrix I . Hence, we can rewrite the signal model as
y = H eff s = Qs.
That is, by plugging also into the previous equation the noise components and focusing on the symbolsampled complex-baseband received signal at each Rx, we have
yk =
1
sk + nk , k = 1, 2.
F (:, k )
Then, the SINR at R X k is given by
SINRk =
Pk
, k = 1, 2
F (:, k ) n2
where P k is the transmit power of T X k and n2 is the noise variance. The data rate associated with
R X k is
Rk =log2 (1+SINRk ) , k =1,2

and the sum-rate throughput is given by
R = R1 + R2
[b its/channel use]
Getting the channel estimations from the WARP testbed configuration for each one of four different
pre-determined beam pairs over 100 channel realizations, we have evaluated the average sum-rate
throughput of the system for a target SNR range
[ 10,10] dB. The channel matrices have been
normalized appropriately in the simulation program to facilitate performance evaluation. Also, a

power allocation algorithm has been applied, so that the power vector
P = [P1 P2 ] results in the
maximum possible sum-rate throughput under a sum-power constraint P1 + P2 P . The simulation

results are shown in Figure 29.
Average Sum Rate [bits/channel use]
2.5
Beam pair 1
Beam pair 2
Beam pair 3
Beam pair 4
1.5
0.5
0
-10
-8
-6
-4
-2
10
Average SNR [dB]
Figure 29: Sum-rate throughput for 4 different beam pairs when joint ZFBF transmission is applied
over the beams.
The final demonstrator including major findings during the project illustrates improvements that we
have realized during the HARP project. The demonstator is composed of four elements:
RRH prototypes with ESPAR antennas

CSI based downlink cooperative beamforming
Uplink user selective joint detection
Cloud-RAN integration with network and protocol simulator
RRH prototypes with ESPAR antennas

Remote radio head prototypes with ESPAR antennas have been described in T7.1.
CSI based downlink cooperative beamforming
We have developed a beam selection optimization based on downlink CQI using the RRH prototype.
The demonstration is made of 2 cells with 2 ESPAR antennas and 2 UEs. The CQIs are collected from
the 2 UEs by the log collector and used by the analytics engine to derive best configuration.
Figure 30: System setup (USB case)

In particular, the analytics engine activates the training (learning) phase, then retrieves the data to
perform the analytics, then performs the execution phase. The extreme flexibility of the solution
enables various input data and KPI formulas to be freely chosen. This demonstration showcases the
reprogrammability of the infrastructure.
Uplink user selective joint detection
We have implemented this joint detection scheme in OpenAirInterface eNodeBs deployed in Docker
containers. This implementation lets us benefit from a separate environment for each BBU and from
the co-localization of these ones in the same server. Thus, we can deploy two identical software eNBs,
each using the data coming from both of them corresponding to the linear combination of user signals
scheduled on the same PRB. This data is stored in a shared memory segment of the server and each
eNB uses it to jointly detect the user located in its own cell. This makes the implementation
symmetrical and scalable, since for any number of cooperating cells the processing to do is the same,
just the size of shared data changes. The following Figure 31 shows the architecture of our
implementation of joint detection in Cloud-RAN.
Figure 31: Arcitecture of our implementation of joint detection in Cloud-RAN

Cloud-RAN integration with network and protocol simulator
The implementation of uplink joint detection in OpenAirInterface allows us to reuse the multi-cell
MMSE function in OpenAir simulator. In order to provide a validaltion of this cooperative scheme with
realistic traffic data, we combined results of these simulations with OpNet network model where
system-level enhancements were implemented. This integration shows the joint benefit that the
uplink JD gives when integrated in the optimised network architecture. Thanks to this investigation
we can provide a prototype albeit simulated software of the overall HARP network, with realistic
traffic profiles and protocol behaviours. The following Figure 32 illustrates how system components
are connected.
Figure 32: Interconnection of system components

Concretely, we provide SINR/BER statistics of jointly detected users having different locations all over
the cooperation cluster. From each cell one user in included in the JD processing. These statistics are
than used in OpNet uplink model which is implemented similarly to downlink simulator described in
Deliverables 6.2 and 6.3. The combination of cooperative transmission in Cloud-RAN and aggregation
network optimization provides overall validation of HARP system.
References
[1]
J. G. Smith, The information capacity of amplitude-and variance constrained scalar gaussian channels,
Information and Control, vol. 18, no. 3, pp. 203219, 1971.
[2]
S. Shamai and I. Bar-David, The capacity of average and peak-power-limited quadrature gaussian
channels, IEEE Trans. Inf. Theory, vol. 41, no. 4, pp. 10601071, July 1995.
[3]
Rassouli, Borzoo, and Bruno Clerckx. "On the capacity of vector gaussian channels with bounded
inputs." arXiv preprint arXiv:1410.7189 (2014).
[4]
M. A. Sedaghat, R. Mueller, (Continous) Phase Modulation on the Hypersphere, submitted to IEEE

Transaction on Wireless Communications.
[5]
B. Rassouli and B. Clerckx, "Constant Envelope Signaling in MIMO Channels," submitted to IEEE Trans. Inf.
Theory.
[6]
Sedaghat, Mohammad A., and Ralf R. Mueller. "A novel modulation scheme for user devices equipped with
a single-RF MIMO transmitter." WSA 2015; 19th International ITG Workshop on Smart Antennas;
Proceedings of. VDE, 2015.
[7]
M. A. Sedaghat, and R. R. Mller. "Multi-Dimesional Continous Phase Modulation in uplink of MIMO

systems", EUSIPCO 2015, Nice, France.
[8]
V. Barousis, C. B. Papadias and R. Muller, A new signal model for MIMO communications with compact
parasitic arrays, Proc. IEEE International Symposium on Communications, Control and Signal Processing,
Athens, Greece, 2123 May 2014.
[9]
G. Alexandropoulos, V. Barousis and C. B. Papadias, Precoding for multiuser MIMO systems with single-fed
parasitic antenna arrays, IEEE Globecom 2014, Austin, TX, USA, Dec. 8-12, 2014.
[10]
V. Barousis and C. Papadias, Arbitrary precoding with single-fed parasitic arrays: Closed-form expressions
and design guidelines, IEEE Wireless Communications Lett., vol. PP, no. 99, pp. 14, 2014.
[11]
K. Ntougias, D. Ntaikos and C. B. Papadias, Reducing Complexity in Next-Generation MU-MIMO Systems.

arXiv preprint arXiv: 1507.04050 (2015).
[12]
M. Dai and B. Clerckx, Transmit Beamforming for MISO Broadcast Channels with Statistical and Delayed
CSIT, IEEE Trans. On Comm., vol. 63, no. 4, pp. 12021215, April 2015.
[13]
C. Hao, Y. Wu and B. Clerckx, Rate Analysis of Two-Receiver MISO Broadcast Channel with Finite Rate
Feedback: A Rate-Splitting Approach, IEEE Trans. on Comm., vol. 63, no. 9, pp. 3232-3246, September 2015.
[14]
H. Joudeh and B. Clerckx, Sum-Rate Maximization for Linearly Precoded Downlink Multiuser MISO Systems
with Partial CSIT: A Rate-Splitting Approach, submitted to IEEE Trans. on Wireless Comm.
[15]
M. Dai, B. Clerckx, D. Gesbert and G. Caire, "A Rate Splitting Strategy for Massive MIMO with Imperfect
CSIT," submitted to IEEE Trans. on Wireless Comm.
[16]
Y. Wu, Y. Cui and B. Clerckx, Performance Analysis and Optimization of Inter-Tier Interference Coordination
in Downlink Multi-Antenna HetNets with Offloading, IEEE ICC 2015.
[17]
Y. Wu, Y. Cui and B. Clerckx, User-Centric Interference Nulling in Downlink Multi-Antenna Heterogeneous
Networks, IEEE ISIT 2015.
[18]
http://mangocomm.com/products/kits/warp-v3-kit.
2. Dissemination and Use of Foreground

2.1 Dissemination Activities
Throughout HARPs project duration, we pursued targeted opportunities for disseminating the projects core
ideas and results to the scientific, engineering and regulatory communities. As a first step, project website
(http://www.fp7-harp.eu/) was established for exchange of information between the project members and as
an interface for communication with the general public. This site has received more than 5000 visitors to date.
It contains an overview of the project, the consortium partner universities and industry partners, and the key
member personnel involved in the project. The webpage also has a list of all papers produced by the project in
widely accessible journal and conference publications. The project has published 23 journals and 57 conference
papers and more are underway. The results have been published in the following prestigious journals:
IEEE Transactions on Information Theory,
IEEE Transactions on Wireless Communications,
IEEE Transactions on Vehicular Technology,
IEEE Transactions on Communications,
IEEE Communications Letters,
IEEE Wireless Communications Letters,
IEEE Transactions on Signal Processing,
IEEE Journal of Selected Topics in Signal Processing,
IEEE Journal on Selected Areas in Communications,
IEEE Communications Surveys and Tutorials Journal.
The conference papers have been presented at internationally well renowned conferences such as
IEEE International Conference on Communications (ICC),
IEEE Global Telecommunications Conference (Globecom),
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP),
IEEE International Conference on Systems, Signals and Image Processing,
IEEE International Symposium on Wireless and Mobile Communication Systems,
IEEE International Symposium on Information Theory,
IEEE Wireless Communications and Networking Conference,
European Signal Processing Conference,
IEEE International Workshop on Computer Aided Modelling and Design of Communication

Links and Networks,
International Symposium on Communications, Control and Signal Processing,
IEEE International Workshop on Signal Processing Advances in Wireless Communications,
ITG Workshop on Smart Antennas,
European Wireless conference,
IEEE Vehicular Technology Conference,
IEEE International Symposium on Personal, Indoor and Mobile Radio Communications,
European Conference on Networks and Communications,
Asilomar Conference on Signals, Systems and Computers,
Advanced International Conference on Telecommunications (AICT).
The details of all dissemination activities in the project are given below:
2.1.1 Journal Papers

1. A. Checko, H. L. Christiansen, Y. Yan, L. Scolari, G. Kardaras, M. S. Berger, and L. Dittmann,
"Cloud RAN for Mobile Networks - a Technology Overview". IEEE Communications Surveys
and Tutorials Journal, vol. 17, no. 1, pp. 405-426, Mar. 2015.
2. B. Rassouli, C. Hao, and B. Clerckx, A New Proof for the DoF Region of the MIMO Networks
with No CSIT, IEEE Communications Letters, vol. 19, no. 5, pp. 763-766, May 2015.
3. B. Clerckx and D. Gesbert, Space-Time Encoded MISO Broadcast Channel with Outdated
CSIT: An Error Rate and Diversity Performance Analysis IEEE Trans. on Communications, vol.
63, no. 5, pp. 1661-1675, May 2015.
4. M. Dai and B. Clerckx, Transmit Beamforming for MISO Broadcast Channels with Statistical
and Delayed CSIT, IEEE Trans. On Communications, vol. 63, no. 4, pp. 1202-1215, Apr. 2015.
5. P. Aquilina and T. Ratnarajah, Performance analysis of IA techniques in the MIMO IBC with
imperfect CSI, IEEE Trans. On Communications, vol. 63, no. 4, pp. 1259-1270, Apr. 2015.
6. F. A. Khan, H. He, J. Xue and T. Ratnarajah, "Performance Analysis of Cloud Radio Access
Networks with Distributed Multiple Antenna Remote Radio Heads" IEEE Trans. on Signal
Processing, vol. 63, no. 18, pp. 4784-4799, Sept. 2015.
7. A. Papazafeiropoulos and T. Ratnarajah, "Deterministic Equivalent Performance Analysis of
Time-Varying Massive MIMO Systems", IEEE Trans. on Wireless Communications, vol. 14, no.
10, pp. 5795-5809, Oct. 2015.
8. Y. Luo and T. Ratnarajah, "Robust Stochastic Optimization For MISO Broadcasting Channel
With Delayed CSIT and Limited Transmitting Antenna," IEEE Trans. on Vehicular Technology,
vol. 64, no. 8, pp. 3547-3559, Aug. 2015.
9. Y. Wu, Y. Cui and B. Clerckx, Analysis and Optimization of Inter-tier Interference
Coordination in Downlink Multi-Antenna HetNets with Offloading IEEE Trans. on Wireless
Communications, in press.
10. C. Hao, Y. Wu and B. Clerckx, Rate Analysis of Multiuser MISO Systems with Finite Rate
Feedback: A Rate Splitting Approach, IEEE Trans. on Communications, vol. 63, no. 9, pp.
3232-3246, Sept. 2015..
11. A. L. Zhou, F. A. Khan, T. Ratnarajah and C.B. Papadias, "Achieving Arbitrary Signals
Transmission Using a Single Radio Frequency Chain, IEEE Trans. Communications, in press.
12. C. Masouros, M. Sellathurai and T. Ratnarajah, "Vector Perturbation Based on Symbol Scaling
for Limited Feedback MISO Downlinks," IEEE Trans. on Signal Processing, vol. 63, no. 3, pp.
562-571, Feb., 2014.
13. C. Zhong, T. Ratnarajah, Z. Zhang, K-K Wong and M. Sellathurai, "Performance of Rayleigh
Product MIMO Channels with Linear Receivers", IEEE Trans. on Wireless Communications,
vol. 13, No. 4, pp. 2270-2281, Apr. 2014.
14. C. Masouros, M. Sellathurai and T. Ratnarajah, "Maximizing Energy-Efficiency in the Vector
Precoded MU-MISO Downlink by Selective Perturbation," IEEE Trans. on Wireless
Communications, vol. 13, no. 9, pp. 4974-4984, Sept. 2014.
15. X. Yi, S. Yang, D. Gesbert, and M. Kobayashi, "The degrees of freedom region of temporally
correlated MIMO networks with delayed CSIT," IEEE Trans. on Information Theory, vol. 60,
no. 1, pp. 594-614, Jan. 2014.
16. Ralf Mller, Laura Cottatellucci, and Mikko Vehkaper, Blind pilot decontamination, IEEE
Journal of Selected Topics in Signal Processing, vol. 8, no. 5, pp. 773-786, Oct. 2014.
17. Haifan Yin, David Gesbert, and Laura Cottatellucci, Dealing with Interference in Distributed
Large-scale MIMO Systems: A Statistical Approach, IEEE Journal of Selected Topics in Signal
Processing, vol. 8, no. 5, pp. 942-953, Oct. 2014.
18. V. I. Barousis and C. B. Papadias, Arbitrary precoding with single-fed Parasitic arrays: Closedform expressions and design guidelines, IEEE Wireless Communications Letters, vol. 3, no.
2, pp. 229-232, Apr. 2014.
19. M. Matthaiou, C. Zhong, M. R. McKay, and T. Ratnarajah, ``Sum Rate Analysis of ZF Receivers
in Ditributed MIMO Systems", IEEE Journal on Selected Areas in Communications, (Special
issue on Large-scale multiple antenna wireless systems), vol. 31, no. 2, pp. 180-191, Feb. 2013.
20. H. Yin, D. Gesbert, M. Filippou and Y. Liu, A Coordinated Approach to Channel Estimation in
Large-Scale Mulitple-Antenna, IEEE Journal on Selected Areas in Communications, (Special
issue on Large-scale multiple antenna wireless systems), vol. 31, no. 2, pp. 264-273, Feb. 2013.
21. C. Masouros, M. Sellathurai and T. Rantarajah, Large-Scale MIMO Transmitters in Fixed
Physical Spaces: the Effect of Correlation and Mutual Coupling, IEEE Trans. on
Communications, vol. 61, no. 7, pp. 2794-2804, July 2013.
22. C. Masouros, M. Sellathurai and T. Rantarajah, Computationally Efficient Vector
Perturbation
Precoding
Using
Thresholded
Optimization, IEEE
Trans.
on
Communications, vol. 61, no. 5, pp. 1880-1890, May 2013.
23. C. Masouros, M. Sellathurai, T. Ratnarajah, A Low Complexity Sequential Encoder for
Threshold Vector Perturbation, IEEE Communications Letter, vol.17, no.12, pp.2225-2228,
Dec. 2013.
2.1.2 Conference Papers

1. B. Rassouli and B. Clerckx, On the Capacity of Vector Gaussian Channels with Bounded
Inputs, in Proc. of IEEE International Conference on Communications, London, June 2015.
2. B. Rassouli , C. Hao, and B. Clerckx, DoF Analysis of the K-user MISO Broadcast Channel with
Hybrid CSIT, in Proc. of IEEE International Conference on Communications, London, June
2015.
3. Y. Wu, Y. Cui and B. Clerckx, Performance Analysis and Optimization of Inter-Tier
Interference Coordination in Downlink Multi-Antenna HetNets with Offloading, in Proc. of
IEEE International Conference on Communications, London, June 2015.
4. A. Checko, A. Juul, H. Christiansen and M S. Berger, "Synchronization Challenges in Packetbased Cloud-RAN Fronthaul for Mobile Networks", in Proc. of IEEE ICC 2015 - Workshop on
Cloud-Processing in Heterogeneous Mobile Communication Networks (IWCPM), London, June
2015.
5. M. Artuso, H. Christiansen, Fronthaul Dimensioning in C-RAN with Web Traffic for
Coordinated Multipoint Joint Transmission, in Proc. Of IEEE ICC 2015 - IEEE Workshop on
Next Generation Backhaul/Fronthaul Networks BackNets 2015, London, June2015.
6. Y. Wu, Y. Cui and B. Clerckx, User-Centric Interference Nulling in Downlink Multi-Antenna
Heterogeneous Networks, in Proc. of IEEE ISIT 2015, Hong Kong, June 2015.
7. L. Zhou, F. A. Khan, and T. Ratnarajah, "Arbitrary Signal Transmission Using An ESPAR
Antenna", in Proc. of IEEE International Conference on Communications, London, June, 2015.
8. P. Aquilina and T. Ratnarajah, Topological Interference Management for Interference
Broadcast Channels with Alternating Connectivity in Proc. Of IEEE ISIT, Hong Kong, June 1419, 2015.
9. S. Biswas, J. Xue, F. Khan and T. Ratnarajah, "On the Capacity of Correlated Massive MIMO
Systems using Stochastic Geometry" in Proc. of IEEE ISIT, Hong Kong June 14-19, 2015.
10. A. Avramova, H. Chrisansen and V. Iversen, Cell Deployment Opmization for Cloud Radio
Access Networks using Teletraffic Theory, The Eleventh Advanced International Conference
on Telecommunications, AICT 2015, June 21 - 26, 2015.
11. P. Aquilina and T. Ratnarajah, "Topological Interference Management for Two Cell
Interference Broadcast Channels with Alternating Connectivity," in Proc. of IEEE 40th
International Conference on Acoustics, Speech and Signal Processing, Brisbane, Australia, April
19-24, 2015.
12. C. Masouros, M. Sellathurai, and T. Rantarajah, "A Scalable Performance-Complexity
Tradeoff for Vector Precoding by Partial Perturbation," in Proc. of 16th IEEE International
Workshop on Signal Processing Advances in Wireless Communications, Stockholm, Sweden,
June 2015.
13. H. Du, T. Ratnarajah, M. Sellathurai, and J. Chambers, "Mixed Norm Minimization for MIMO
Cellular Interference Channel" in Proc. of 16th IEEE International Workshop on Signal
Processing Advances in Wireless Communications, Stockholm, Sweden, June 2015.
14. H. Holm, A. Checko, R. Al-obaidi and H. Christiansen, Optimizing CAPEX of Cloud-RAN

Deployments in Real-life Scenarios by Optimal Assignment of Cells to BBU Pools in Proc. of
EuCNC 2015, Paris, July 2015.
15. R. Al-obaidi, A. Checko, H. Holm and H. Christiansen, Optimizing Cloud-RAN Deployments in
Real-life Scenarios Using Microwave Radio(awarded Best Student paper award), in Proc.
of EuCNC 2015, Paris, July 2015.
16. M. A. Sedaghat, and R. R. Mller, "A novel modulation scheme for user devices equipped
with a single-RF MIMO transmitter" in Proc. of 19th International ITG Workshop on Smart
Antennas, Ilmenau, Germany3-5 Mar. 2015.
17. M. A. Sedaghat, R. R. Mller, "Multi-Dimensional Continuous Phase Modulation in Uplink of
MIMO Systems", in Proc. of EUSIPCO, Nice, France, Sept. 2015.
18. M. Dai, B. Clerckx, D. Gesbert and G. Caire, A Hierarchical Rate Splitting Strategy for FDD
Massive MIMO under Imperfect CSIT, Invited paper to IEEE CAMAD 2015, Guildford, UK,
Sept. 7-9, 2015.
19. A. C. Cirik, K. Rikkinen, Y. Rong and T. Ratnarajah, "A subcarrier and power allocation
algorithm for OFDMA full-duplex systems", in Proc. of European Conference on Networks and
Communications (EuCNC'2015), Paris, France, June 29-July 2, 2015.
20. V. Barousis, C. Papadias, R. Mller, A New Signal Model for MIMO communication with
Compact Parasitic Arrays, in Proc. of International Symposium on Communications, Control
and Signal Processing, Athens, Greece, May 2014.
21. A. Checko, H. Holm, and H. Christiansen, Optimizing small cell deployment by the use of CRANs, in Proc. of European Wireless 2014 (EW 2014), Barcelona, May 2014.
22. A Papazafeiropoulos and T Ratnarajah, "Uplink Performance of Massive MIMO Subject to
Delayed CSIT and Anticipated Channel Prediction," IEEE 39th International Conference on
Acoustics, Speech and Signal Processing, Florence, Italy, May 4-9, 2014.
23. C. Masouros, M. Sellathruai and T. Ratnarajah, "Bridging the Gap Between Linear and NonLinear Precoding in Small- and Large-Scale MIMO Downlinks," In Proc. of IEEE International
Conference on Communications, Sydney, Australia, June 10-14, 2014.
24. Haifan Yin, D. Gesbert, and L. Cottatellucci, A Statistical Approach to Interference Reduction
in Distributed Large-Scale Antenna Systems, in Proc. of IEEE International Conference on
Acoustics, Speech, and Signal Processing (ICASSP), May 2014, Florence, Italy.
25. A. Papazafeiropoulos and T. Ratnarajah, "Precoding for Downlink Massive MIMO with
Delayed CSIT and Channel Prediction," In Proc. of IEEE Wireless Communications and
Networking Conference, Istanbul, Turkey, April 6-9, 2014.
26. A. Papazafeiropoulos and T. Ratnarajah, "Ergodic Channel Capacity for Generalized Fading
channels," In Proc. of IEEE Wireless Communications and Networking Conference, Istanbul,
Turkey, April 6-9, 2014.
27. A. Sedaghat, R. R. Muller, and G. Fischer, A Novel Single-RF Transmitter for Massive MIMO,
in Proc. of ITG Workshop on Smart Antennas (WSA), March 2014.
28. M. Sedaghat, R. Mller, and G. Fischer, Broadcast precoding for massive MIMO subject to
an instantaneous total power constraint, in Proc. of IEEE Global Commun. Conf.
(Globecom), Austin, TX, USA, Dec. 2014.
29. Laura Cottatellucci, Spectral Efficiency of Extended Networks with Randomly Distributed
Transmitters and Receivers, in proc. of 2nd IEEE China Summit and International Conference
on Signal and Information Processing (ChinaSIP14), Xi'an China, July 9--13, 2014.
30. M. Artuso, H. Christiansen, Discrete-Event Simulation of Backhaul Impact on Coordinated
Multi-Point in LTE-Advanced, in Proc. of 11th International Symposium on Wireless
Communication System, Barcelona, August 2014.
31. M. Artuso and H. Christiansen, Modeling and Event-Driven Simulation of Coordinated MultiPoint in LTE-Advanced with Constrained Backhaul, in proc. of 2014 Winter Simulation
Conference, Savannah, GA, Dec. 2014.
32. G. C. Alexandropoulos, V. I. Barousis, and C. B. Papadias, "Precoding for multiuser MIMO
systems with single-fed parasitic antenna arrays," in Proc. of IEEE Global Communications
Conference, Austin, USA, 812 Dec. 2014.
33. Jean-Marc Kelif, Stephane Senecal, Constant Bridon, and Marceau Coupechoux, Quality of
Service and Performance Evaluation: a Fluid Approach for Poisson Wireless Networks, in
Proc. of IEEE NOF 2014, Paris, France, 3-5 Dec. 2014.
34. Jean-Marc Kelif, Stephane Senecal, Marceau Coupechoux, Constant Bridon , Analytical
Performance Model for Poisson Wireless Networks with Pathloss and Shadowing
Propagation, Invited to WONC Workshop, Proc. of IEEE GLOBECOM 2014, Austin/USA, 2 Dec.
2014
35. L. Dittmann, H. L. Christiansen, and A. Checko, "Meeting fronthaul challenges of future
mobile network deployments - the HARP approach", Invited to WONC Workshop, Proc. of
IEEE GLOBECOM 2014, Austin/USA, 2 December 2014
36. M. A. Sedaghat, R. R. Mller,and G. Fischer, Broadcast Precoding for Massive MIMO Subject
to an Instantaneous Total Power Constraint, in Proc. of IEEE Global Communications
Conference (GLOBECOM), Dec. 2014.
37. Laura Cottatellucci, Capacity per Unit Area of Distributed Antenna Systems with
Centralized Processing, in Proc. of IEEE Global Communications Conference (GLOBECOM),
Dec. 2014.
38. R. R. Muller, M. A. Sedaghat, and G. Fischer, Load modulated massive MIMO, in Proc. of
IEEE Global Conf. Signal & Inform. Proc. (Global SIP), Atlanta, GE, USA, Dec. 2014.
39. A. Zakrzewska, A. P. Avramova, H. Christiansen, Y. Yan, A. Checko, A. Dogadaev, S. Ruepp, M.
S. Berger and L. Dittmann, A Framework for Joint Optical-Wireless Resource Management
in Multi-RAT, Heterogeneous Mobile Networks, in Proc. of IEEE International Conference on
Communications (ICC), Budapest, Hungary, 9-13 June 2013.
40. A. Dogadaev, A. Checko, A. P. Avramova, A. Zakrzewska, Y. Yan, S. Ruepp, M. S. Berger, and L.
Dittmann and H. Christiansen, Traffic Steering Framework for Mobile-Assisted Resource
Management in Heterogeneous Networks, in Proc. of IEEE 9th International Conference on
Wireless and Mobile Communications (ICWMC) 2013, Nice, France 21-26 July 2013.
41. C. Hao and B. Clerckx, Imperfect and Unmatched CSIT IS Still Useful for the Frequency
Correlated MISO Broadcast Channel, In Proc. on IEEE International Conference on
Communications (ICC), Budapest, Hungary, 9-13 June 2013.
42. C. Masouros, T. Rantarajah, M. Sellathurai, J. Chen and T. Kenneth, Towards Massive-MIMO
Transmitters: On the Effects of Deploying Increasing Antennas in Fixed Physical Space, in
Proc. of Future Network and Mobile Summit, Lisbon, Portugal, 3-5 July 2013.
43. M. A. Sedaghat, and R. R. Mller, "Large System Analysis of Low-Cost MIMO Transmitters", in
Proc. of ITG Workshop on Smart Antennas (WSA), Mar. 2013.
44. H. Yin, D. Gesbert, M. Filippou and Y. Liu, Decontaminating Pilots in Massive MIMO
Systems, In Proc. of IEEE International Conference on Communications (ICC), Budapest,
Hungary, 9-13 June 2013.
45. P. de Kerret, J. Hoydis, and D. Gesbert, Rate Loss Analysis of Transmitter Cooperation with
Distributed CSIT, in Proc. of IEEE International Workshop on Signal Processing Advances in
Wireless Communications (SPAWC), Darmstadt, Germany, 16-19 June 2013.
46. L. Cottatellucci, R. Muller, and M. Vehkapera, Analysis of Pilot Decontamination Based on
Power Control, in Proc. of IEEE Vehicular Technology Conference (VTC Spring), pp. 15, Dresden, Germany, June 2013.
47. P. de Kerret, X. Yi and D. Gesbert , On the Degrees of Freedom of the K-user Time Correlated
Broadcast Channel with Delayed CSIT, in Proc. of IEEE International Symposium on
Information Theory (ISIT), Istanbul, Turkey, 7-12, July 2013.
48. A. Papazafeiropoulos and T. Ratnarajah, Degrees of Freedom of Multiple-Antenna
Interference Channel with General CSIT", In Proc. of IEEE 24 Annual International Symposium
on Personal, Indoor and Mobile Radio Communications, London, UK, Sep. 08-11, 2013.
49. C. Masouros, T. Ratnarajah, M. Sellathurai, "Low Complexity Vector Precoding for Fast Fading
MIMO Downlinks", In Proc. of IEEE Global Communications Conference, Atlanta, USA, 9-13,
Dec. 2013.
50. B. E. Godana and D. Gesbert, Coordinated Beamforming in Multicell Networks with Channel
State Information Exchange Delays, In Proc. of IEEE International Symposium on Personal,
Indoor and Mobile Radio Communications (PIMRC), London, UK, 8-11 Sep. 2013.
51. C. Hao and B. Clerckx, MISO broadcast channel with imperfect and (un)matched CSIT in the
frequency domain: DoR region and transmission strategies, In Proc. of IEEE International
Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), London, UK, 811 Sept. 2013.
52. A. Checko, H. Christiansen and M. S. Berger, Evaluation of energy and cost savings in mobile
Cloud Radio Access Networks, In Proc. of OPNETWORK 2013 conference, Washington D.C.,
USA, August 26-30 2013.
53. J. M. Klif, S. Senecal and M. Coupechoux, Impact of Small Cells Location on Performance
and QoS of Heterogeneous Cellular Networks, in Proc. of IEEE International symposium on
personal, indoor and mobile radio communications (PIMRC), 8-11 Sep., London UK, 2013.
54. R. Mller, M. Vehkaper, and L. Cottatellucci: Blind pilot decontamination in Proc. of ITG
Workshop on Smart Antennas (WSA), Stuttgart, Germany, Mar. 2013.
55. R. Mller, M. Vehkaper, and L. Cottatellucci Analysis of blind pilot decontamination in
Proc. of 47th Asilomar Conference on Signals, Systems and Computers, Pacific Grove, CA, USA,
Nov. 2013.
56. C. Masouros, T. Ratnarajah, and M. Sellathurai, "Complexity Reduction for Vector Precoding
Using QoS Requirements, in Proc. of IEEE 38th International Conference on Acoustics, Speech
and Signal Processing, Vancouver, Canada, May 26-31, 2013.
57. C. Masouros, M. Sellathurai, and T. Ratnarajah, "A Performance-Complexity Tradeoff for
Vector Perturbation Precoding", In Proc. the IEEE International Conference on
Communications, Budapest, Hungary, June 9-13, 2013.
2.1.3 Industry White Papers

1. Checko, Aleksandra; Kardaras, Georgios; Lanzani, Christian Fabio Alessandro; Temple, Dan;
Mathiasen, Carsten; Pedersen, Lars A.; Klaps, Bert, Transport of Baseband Radio Serial
Protocols in C-RAN Architecture for Mobile Network Applications White Paper OTN, Altera,
2014.
2.1.4 Papers Submitted or in Preparation

1. A. Sedaghat, V. Barousis, R. R. Muller, and C. Papadias, Load Modulated Arrays: a Low
Complexity Antenna Technology for Massive, Distributed and Small Cell Wireless Networks,
second round review to IEEE Comms. Magazine.
2. A. Sedaghat, R. R. Mller: "(Continuous) Phase Modulation on the Hypersphere Submitted
to IEEE Trans. on Wireless Communications, Sept. 2015.
3. B. Rassouli, C. Hao and B. Clerckx, DoF analysis of the K-user MISO broadcast channel with
hybrid CSIT, first round review to IEEE Trans. on Info. Theory, 2015.
4. B. Rassouli and B. Clerckx, "On the Capacity of Vector Gaussian Channels With Bounded
Inputs" first round review to IEEE Trans. on Info Theory, 2015.
5. M. Dai, B. Clerckx, D. Gesbert and G. Caire, "A Rate Splitting Strategy for Massive MIMO with
Imperfect CSIT," submitted to IEEE Trans. on Wireless Comm.
6. H. Joudeh and B. Clerckx, Sum-Rate Maximization for Linearly Precoded Downlink Multiuser
MISO Systems with Partial CSIT: A Rate-Splitting Approach, submitted to IEEE Trans. on
Wireless Comm.
7. C. Hao, B. Rassouli, and B. Clerckx, Achievable DoF Regions of MIMO Networks with
Imperfect CSIT, submitted to IEEE Trans. on Info Theory
8. B. Rassouli and B. Clerckx, Constant Envelope Signaling in MIMO Channels, submitted to
IEEE Trans. on Info Theory
9. S. Biswas, C. Masouros and T. Ratnarajah, Performance Analysis of Large Multi-User MIMO
Systems with Space-Constrained 2D Antenna Arrays, Submitted to IEEE Trans. on Wireless
Communications, under 3nd round of review.
10. K. Ntougias, D. Ntaikos, and C. B. Papadias, "Reducing Complexity in Next-Generation MUMIMO Systems". Online: http://arxiv.org/pdf/1507.04050v1.pdf
11. A. Checko, A.P. Avramova, M. S. Berger and H. L. Christiansen, Evaluating C-RAN fronthaul
functional splits in terms of network level energy and cost savings Submitted to IEEE Journal
Of Communications And Networks
12. L. Zhou, T. Ratnarajah and J.Xue, "Energy Efficiency of Cloud Radio Access Network with
ESPAR Antenna", IEEE Journal on Selected Areas in Communications (Special Issue on EnergyEfficient Techniques for 5G Wireless Communication Systems), under 2nd round of review.
13. H. He, J. Xue, F. A. Khan, T. Ratnarajah, and C. B. Papadias, "Modeling and Analysis of Cloud
Radio Access Networks Using Mat\'ern Hard-core Point Processes," IEEE Trans. on Wireless
14. P. Aquilina and T. Ratnarajah, "On the Degrees of Freedom of Interference Broadcast
Channels with Topological Interference Management," IEEE Trans. Communications, under
2nd round of review.
15. M. Artuso, H. Christiansen, Optimizing TCP for Cloud-Based Mobile Networks, submitted
to IEEE Vehicular Technology Conference 2016 - Spring, Nanjing 2016.
2.1.5 Multi-partner joint publications

Joint Journal Papers
1. Ralf Mller, Laura Cottatellucci,and Mikko Vehkaper, Blind pilot decontamination, IEEE
Journal of Selected Topics in Signal Processing, vol. 8, October 2014.
2. A. Checko, H. L. Christiansen, Y. Yan, L. Scolari, G. Kardaras, M. S. Berger, L. Dittmann, "Cloud
RAN for Mobile Networks - a Technology Overview". IEEE Communications Surveys and
Tutorials Journal, Vol. 17, no. 1, March 2015
3. B. Clerckx and D. Gesbert, Space-Time Encoded MISO Broadcast Channel with Outdated
CSIT: An Error Rate and Diversity Performance Analysis accepted to IEEE Trans. on
Communications, 2015
Transmission Using a Single Radio Frequency Chain, accepted for IEEE Trans.
Communications
Joint Conference Papers

1. A. Zakrzewska, A. P. Avramova, H. Christiansen, Y. Yan, A. Checko, A. Dogadaev, S. Ruepp, M.
S. Berger and L. Dittmann, A Framework for Joint Optical-Wireless Resource Management
in Multi-RAT, Heterogeneous Mobile Networks, In Proc. on IEEE International Conference
on Communications (ICC), Budapest, Hungary, 9-13 June 2013.
2.
A. Dogadaev, A. Checko, A. P. Avramova, A. Zakrzewska, Y. Yan, S. Ruepp, M. S. Berger, and L.

Dittmann and H. Christiansen, Traffic Steering Framework for Mobile-Assisted Resource
Management in Heterogeneous Networks, In Proc. on IEEE 9th International Conference on
Wireless and Mobile Communications (ICWMC) 2013, Nice, France 21-26 July 2013.
3. L. Cottatellucci, R. Muller, and M. Vehkapera, Analysis of Pilot Decontamination Based on

Power Control, in Proc. IEEE Vehicular Technology Conference (VTC Spring), pp. 15, Dresden, Germany, June 2013.
4. A. Checko, H. Christiansen and M. S. Berger, "Evaluation of energy and cost savings in mobile
Cloud Radio Access Networks", In Proc. of OPNETWORK 2013 conference, Washington D.C.,
USA, August 26-30 2013.
5. R. Mller, M. Vehkaper, L. Cottatellucci: Blind pilot decontamination ITG Workshop on
Smart Antennas (WSA), Stuttgart, Germany, Mar. 2013.
6. R. Mller, M. Vehkaper, L. Cottatellucci Analysis of blind pilot decontamination 47th
Asilomar Conference on Signals, Systems and Computers, Pacific Grove, CA, USA, Nov. 2013.
7. V. Barousis, C. Papadias, R. Mller, A New Signal Model for MIMO communication with
Compact Parasitic Arrays. International Symposium on Communications, Control and Signal
Processing, Athens, Greece, May 2014.
8. A. Checko, H. Holm, and H. Christiansen, Optimizing small cell deployment by the use of CRANs, Proc. of European Wireless 2014 (EW 2014), May 2014, Barcelona.
9. L. Dittmann, H. L. Christiansen, A. Checko, "Meeting fronthaul challenges of future
mobile network deployments - the HARP approach", Invited to WONC Workshop, Proc. of
IEEE GLOBECOM 2014, Austin/USA, 2 December 2014
10. A. Checko, A. Juul, H. Christiansen and M S. Berger, "Synchronization Challenges in Packetbased Cloud-RAN Fronthaul for Mobile Networks", in Proc. of IEEE ICC 2015 - Workshop on
Cloud-Processing in Heterogeneous Mobile Communication Networks (IWCPM)
11. H. Holm, A. Checko, R. Al-obaidi and H. Christiansen, Optimizing CAPEX of Cloud-RAN
Deployments in Real-life Scenarios by Optimal Assignment of Cells to BBU Pools EuCNC 2015
Paris, July 2015
12. R. Al-obaidi, A. Checko, H. Holm and H. Christiansen, Optimizing Cloud-RAN Deployments in
Real-life Scenarios Using Microwave Radio awarded Best Student paper, EuCNC 2015,
Paris, July 2015.
13. M. Dai, B. Clerckx, D. Gesbert and G. Caire, A Hierarchical Rate Splitting Strategy for FDD
Massive MIMO under Imperfect CSIT, accepted to IEEE CAMAD 2015 (invited paper).
Joint Industry White Papers

1. Checko, Aleksandra; Kardaras, Georgios; Lanzani, Christian Fabio Alessandro; Temple, Dan;
Mathiasen, Carsten; Pedersen, Lars A.; Klaps, Bert, Transport of Baseband Radio Serial
Protocols in C-RAN Architecture for Mobile Network Applications White Paper OTN, Altera,
2014. 9 p.
Joint Journal Articles Submitted or in Preparation

1. A. Sedaghat, V. Barousis, R. R. Muller, and C. Papadias, Load Modulated Arrays: a Low
Complexity Antenna Technology for Massive, Distributed and Small Cell Wireless Networks,
Submitted to ComSoc magazine.
2. M. Dai, B. Clerckx, D. Gesbert and G. Caire, "A Rate Splitting Strategy for Massive MIMO with
Imperfect CSIT," submitted to IEEE Trans. on Wireless Comm.
Transmission Using a Single Radio Frequency Chain, submitted to IEEE Trans.
2.1.6 Invited Talks, Tutorials, Poster Presentations

1. T. Ratnarajah delivered a talk on HARP project at Future Network and Mobile Summit,
Lisbon, July 2013.
2. Senecal and Kelif delivered a talk on HARP project at:
-
Computation and Simulation Institutre of Pierre&Marie Curie University, Paris, March 2014.
Orange internal TOOS project seminars, Paris, Dec. 2013
3. D. Gesbert delivered a talk on A statistical approach to interference management in large

scale antenna systems, KTH, Stockholm, Mar. 2014.
4. L. Cottatellucci presented Low complexity pilot decontamination via blind signal subspace
estimation at CEFIPRA workshop, IIS Bangalore, January 2014
5. M. Sedaghat: Poster, IEEE Workshop on Advanced Information Processing for Wireless
Communication Systems, Aalborg
6. M. Sedaghat and R. R. Mller presented Load Modulated Massive MIMO, at Communication
Theory Workshop 2014
7. Michail presented Traffic Adaptive Base Station Management Scheme for energy-aware
Mobile networks, EUCNC conference, Bologna, June 2014
8. C. Papadias presented Load-Modulated Antenna Arrays for Small & Large Scale MIMO
Systems HARP project, 2014 Smart Antennas Workshop, at Stanford, CA, USA, Aug. 1, 2014
9. C. Papadias presented Compact Antenna Array Designs for Remote Radio Head (RRH) 5G
Network Architectures, - Results from the FP7 Project HARP (RAS Cluster), Future Internet
Assembly (FIA ATHENS 2014), Greece, March 17, 2014
10. Alexandropoulos presented "Transceiver techniques for multi-user multi-antenna systems,"
IEEE Greece Signal Processing Chapter and UoP's Network on Research and Applications of
Wireless Communication and Sensor Networks, UoP, Patras, Greece, June 26, 2014.
11. Alexandropoulos presented "Some recent results in signal processing techniques for multiuser wireless system, FORTH, Heraklion, Crete, June 10, 2014.
12. B. Rassouli and B. Clerckx presented On the Capacity of Vector Gaussian Channels with
Bounded Inputs
-
iCore Inauguration Workshop, Imperial College, April 2015.

European School on Information Theory, April 2015.
13. Ralf Mller, Information Theory and Applications Workshop (ITA) San Diego, Feb 2015.
14. Ralf Mller, ``Load Modulated MIMO: A New Hardware Concept to Reduce Cost, Size, and
Amplifier Back-off', keynote speech at WONC collocated with GLOBECOM 2015, Austin, US,
Dec 2014
15. Ralf Mller, Workshop on Smart Antennas (WSA) Ilmenau, Germany, Mar. 2015.
16. M. A. Sedaghat poster presentation in SPcoding school in Campinas, Brazil (17-31 Jan. 2015):
Load Modulated Massive MIMO
17. Mohammad A. Sedaghat and Ralf Muller. A novel modulation scheme for user devices
equipped with a single-RF MIMO transmitter, In Proc. ITG Workshop on Smart Antennas
(WSA), Ilmenau, Germany, Mar. 2015.
18. M. A. Sedaghat participated in RF & Communications Round Table in Lund University, Sweden
(11-13 Nov. 2013) on massive MIMO demo
19. Aleksandra Checko, C-RAN fronthaul: Options, benefits and challenges. At http://www.ictijoin.eu/ijoin-winter-school-2015/
20. Joint Poster: Dimitrios Ntaikos, Bobby Gizas, Constantinos Papadias, Laurent Roullet, Franois
Taburet. ``Over-the-air demonstration for Remote Radio Head (RRH) based LTE access with the
use of parasitic antenna arrays: Results from the FP7 project HARP, EUCNC 2015
21. H.Chrisansen: Architectures of C-RAN, workshop at the COST 1044 training school.
Luxembourg, April 2015
22. Ralf Mller, Continuous Phase Modulation on the Hypersphere, Future of Wireless
Workshop, Stockholm, June 2015
23. Bruno Clerckx MIMO Wireless Networks: A Promising Rate Splitting Transceiver
Architecture Nanyang Technological University (NTU), July 2015 (sponsored by IEEE ComSoc
Singapore Chapter).
24. C. Papadias (AIT), Single-RF Transmission: An Emerging Technology for both Link and MultiUser MIMO Systems, keynote speech at Emerging MIMO Technologies and Millimeter-waves
for 5G Networks Workshop (mmW5G-WS) collocated with VTC 2015, Glasgow, UK, May 2015
Architecture I2R ASTAR, Singapore, July 2015.
Architecture LGE, Korea, August 2015
27. C. B. Papadias, "Antenna Arrays with less RF Chains than Elements: An Emerging Technology
for Multi-Antenna Wireless Systems, invited talk presented at
-
2015 SEU-UESTC Workshop on Advances in Wireless Communications (WAWC 2015),

UEST, Chengdu, China, July 7, 2015
Southeast University, Nanjing, China, July 9, 2015.
28. Navid Nikaein, Thrasyvoulos Spyropoulos, and Aleksandra Checko Backhaul and Fronthaul:
Principles, Challenges, and Technologies, submitted proposal to WCNC 2016
2.1.7 Workshops
1. WONC: Workshop on Wireless Optical Network Convergence in Support of Cloud Architectures
- Globecom 2014
Co-chairs: Laura Cottatellucci, Anna Tzanakaki
http://www.fp7-harp.eu/globecom-2014wonc.html
Contributions from HARP:
1. Keynote speech: ``Load Modulated MIMO: A New Hardware Concept to Reduce Cost, Size, and
Amplifier Back-off by Ralf Muller.
2. Invited paper: Analytical Performance Model for Poisson Wireless Networks with Pathloss and
Shadowing Propagation, by Jean-Marc Kelif, Stephane Senecal, Marceau Coupechoux, Constant
Bridon.
3. Invited paper: "Meeting fronthaul challenges of future mobile network deployments - the HARP
approach,L. Dittmann, H. L. Christiansen, A. Checko.
2. mmW5G-WS on Emerging MIMO Technologies and Millimeter-waves for 5G Networks - VTCSpring 2015
TPC-chairs from HARP: Bruno Clerckx, Constantinos Papadias, and Tharmalingam Ratnarajah
http://www.miwaves.eu/mmW5G-WS.html
Contributions from HARP:
1. Keynote speech: ``Single-RF Transmission: An Emerging Technology for both Link and Multi-User
MIMO Systems by Constantinos Papadias
Other contributions from the related European projects MiWaveS and MiWEBA
2.1.8 Special Sessions

1. EUSIPCO, Nice, France (2015, 31 Aug. 4 Sept. 2015)
Title: Massive and cloud-based virtual MIMO: alternative, complementary or merging wireless
network technologies?
Organizers: Laura Cottatellucci, Petros Elia
http://www.eusipco2015.org/content/special-sessions#SS12
Contributions from HARP
1. Invited paper: ``Multi-dimensional Continuous Phase Modulation in uplink of MIMO systems by M.
A. Sedaghat, R. R. Muller
2. GdR-ISIS, Paris, France (20 Nov. 2015)
Title: Cloud- and fog-based PHY communications in 5G: performance, feedback and complexity
Organizers: Laura Cottatellucci, Petros Elia
http://www.gdr-isis.fr/index.php?page=reunion&idreunion=291
Contributions from HARP
1. CRAN challenges: the architecture and solutions in HARP by Laurent Roullet.
2. C-RAN fronthaul enhancements using Software Defined Networking by Dora Boviz.
2.2 Use of foregrounds: Exploitable foreground and plan for

exploitation
The results of HARP are carried on into the future in several ways. One such way is when people
active in project move on to other workplaces and projects. Accordingly, we list in Section 2.2.1
graduate students, master theses and post-docs who have been active in the project. In Section 2.2.2,
we list some specific foreground items and how they will be exploited. Section 2.2.3 finally outlines a
path towards standardization provided by our industrial partner Alcatel-Lucent.
2.2.1 Education
PhD Students and Post Docs
PhD students
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Haifan Yin (EURECOM)

Borzoo Rassouli (Imperial College)
Sedaghat (NTNU)
Lin Zhou (UEDIN)
S. Biswas (UEDIN)
H. He (UEDIN)
P. Aquilina (UEDIN)
Aravinthan (ALU Bell Labs)
D. Boviz (ALU Bell Labs)
A. Checko (Radiocomp & DTU)
M. Artuso (DTU)
K. Ntougias (AIT)
Marcano (DTU)
Post Docs and Researchers

1.
2.
3.
4.
5.
Dr J. Xue (UEDIN)
Dr Fahd Ahmed Khan (UEDIN)
Dr Papazafeiropoulos (UEDIN)
Dr Faheem Ahmad khan (UEDIN)
Dr Ebtihal Yousif (UEDIN)
6.
7.
8.
9.
10.
11.
Dr Miltiades C. Filippou (UEDIN)

Dr Satyanarayana Vuppala (UEDIN)
Dr Ali Cagatay Cirik (UEDIN)
Dr Yueping Wu (Imperial College)
Dr Dimitrios Ntaikos (AIT)
Bobby Gizas (AIT)
Students benefitting from HARP

1. Mingbo Dai (Imperial College)
2. Hamdi Joudeh (Imperial College)
3. Chenxi Hao (Imperial College)
PhD, Master and Bachelor Students Projects

PhD Theses
1. Aleksandra Checko, "Cloud Radio Access Network (C-RAN) architecture", Expected: February
2016
2. Mohammad Ali Sedaghat, Load-Modulated Single-RF MIMO Transmitters, Expected: March
2016
3. Haifan Yin Interference Mitigation in Massive MIMO Systems, Expected: December 14, 2015
Master Theses
1. H. Bro, Backhaul for 5G mobile networks, supervised by H. Christiansen, M. Artuso (DTU)
2. F. H. Panah, Radio over Fibre Fronthaul design for 5G mobile networks, supervised by H.
Christiansen, M. Artuso (DTU)
3. L. Salim, Millimetre wave and 5G for Mobile Networks supervised H. Christiansen, A.
Marcano (DTU)
4. M. Rodriguez, 5G Technologies: Challenges for Next Generation Mobile Networks,
supervised by H. Christiansen (DTU)
5. Talu, Wireless interfaces for 5G mobile networks, supervised H. Christiansen (DTU)
6. Mixi Han Transceiver design for very large scale MIMO systems, supervised by Bruno
Clerckx (Imperial)
7. MIMO systems, supervised by Bruno Clerckx (Imperial)
8. Zongze Lu A new look at interference and feedback in MIMO wireless networks: overhear
the interference!, supervised by Clerckx (Imperial)
9. Ge,Y., MIMO Transmission with Single RF Chain, supervised by Clerckx (Imperial)
10. Cao,Q., Statistical and delayed CSIT-aided Multiuser MISO systems, supervised by Clerckx
(Imperial)
11. Xu,S., Space-time/frequency multi-user MIMO MS thesis, supervised by Clerckx (Imperial)
12. Ning, Yan, MIMO Transmission with Single RF Chain, supervised by Clerckx (Imperial)
13. Shi, Ruibo, Rate-Splitting for MISO Broadcast Channels, supervised by Clerckx (Imperial)
14. Gupta, Siddhant, Rate-Splitting for MISO Interference Channels, supervised by Clerckx
(Imperial)
15. Arslan Ali, Evaluation of Passive Load-Modulator Circuits for Single-RF Massive MIMO
Systems, supervised by Ali Sedaghat and Mller (NTNU)
Bachelor Thesis Projects

1. C.Y.J. Blumensaadt, Cloud RAN for 5G Mobile Networks, supervised H. Christiansen (DTU)
2.2.2 Description of exploitable foreground
1. Phase Modulation on the Hyposphere Technique

Description:
In Task 3.2.3, we developed the main results on MIMO systems with constant envelop signals. In LoadModulated Single-RF (LMSRF) MIMO transmitters, there is a Central Power Amplifier (CPA) which
feeds all the antennas via some load modulators. The power efficiency of the CPA is mostly determined
by the Peak to Average Sum Power Ratio (PASPR), i.e., peak to average ratio of the aggregated power
on all the antennas. The lower PASPR, the higher the power efficiency. In the case that the aggregated
power on the antennas becomes constant, the CPA can be very efficient. In this case, the CPA can be
a nonlinear amplifier with no back-off, e.g., class F amplifier with around 80% power efficiency. Having
highly efficient central CPA motivates us to introduce a novel modulation scheme for MIMO
communications called Phase Modulation on the Hypersphere (PMH). In PMH, the aggregated power
an all the antennas before pulse shaping is fixed. There are still many cases which their channel
capacities are unknown including the problem of finding the channel capacity of PMH signaling in
arbitrary MIMO channel for more than two antennas. The channel capacity and the performance in
some cases were investigated. Related research results have been presented in the 2015 International
ITG Workshop on Smart Antennas, Ilmenau, Germany, 2014 IEEE Global Commun. Conf. (Globecom),
Austin, TX, USA and are presented in the second year scientific report.
Exploitation:
The results can be exploited to solve advanced problems as there are still many cases for which their
channel capacities are unknown including the problem of finding the channel capacity of PMH
signaling in arbitrary MIMO channel for more than two antennas. Furthermore, the problem of
investigating channel capacity of PMH in multi-user scenario is unknown in general.
Market:
Research projects and telecommunications sector
Further Development:
PMH will be developed further to solve the problem of finding the channel capacity of PMH signaling
in arbitrary MIMO channel for more than two antennas and the problem of investigating channel
capacity of PMH in multi-user scenario is unknown in general.
Expected impact:
We expect the proposed PMH technique to solve advanced problems for which their channel
capacities are unknown including the problem of finding the channel capacity of PMH signaling in
arbitrary MIMO channel for more than two antennas and the problem of investigating channel
capacity of PMH in multi-user scenario is unknown in general.
IPR:
No patent submitted
Contact details:
RALF MLLER ( RALF.MUELLER@LNT.DE)
2. Remote Radiohead Prototype

Description: In Task 7.2, RRH prototype has been developed. It is made of the following components:
The dataplane: The RRH gateway able to receive IQ samples over Ethernet and convert them to/from
the radio interface; The control plane: Espar antenna controller board able to select the beam; Load
collector and data analytics able to, respectively, log all CQI (among others) and control the
parameters of the antennas. The antenna controller was built as a standalone module, so it could be
tested as a separate system in the lab or to be easily implementable in an end-to-end system for the
final demonstration. The controller is the main interface between the antenna and the rest of the
system that controls the properties of the radiation pattern by tuning the reactive loads of the
antennas parasitic elements or by selecting a desirable beam angle using RF switches.
Exploitation:
The prototype can be used as the starting point for advanced implementations of RRHs in Cloud RAN
networks.
Market:
Research projects and telecommunications sector.
Further development:
The antenna controller was built as a standalone module, so it could be tested as a separate system
in the lab or to be easily implementable in an end-to-end system for the final demonstration. The RRH
prototype will be developed further.
Expected impact:
The RRH prototyping will elevate the knowledge and skills of the participating students and
researchers. The software framework will enable rapid prototyping of new schemes which will result
in excellent and influential publications.
IPR:
No Patent submitted. The prototype will be used in further development of RRHs.
Contact details:
Laurent Roullet (laurent.roullet@alcatel-lucent.com)
3. Prototype of a cloud RAN cluster implementing selected techniques

Description: In Task 5.3.2, we presented the cooperative features developed in the task 5.3 using
results from previous tasks of Work Package 5. To realize downlink cooperation we benefit from the
ability of singe-RF chain parasitic antennas to realize beamforming toward specific users and we
coordinate beam parameters from the central processing unit. Uplink physical layer cooperation is
implemented in our C-RAN cluster thanks to the multi-cell MMSE function that we added to the
eNodeBs deployed in the BBU-pool.
Exploitation: The uplink cooperative technique implemented in our prototype will be exploited a first
step to the realization multi-cell user-selective joint detection in case of cell congestion.
Market: The developed algorithm is intended to be used in future cellular systems. The market, in
this case, consists of all the telecommunication equipment manufacturing companies that offer,
among other products, PHY layer solutions for 4G and beyond communication systems.
Further development: We will follow C-RAN control evolution to improve cooperative processing
control and dynamicity. The SDN controller connected to all RAN elements will allow realizing realtime optimization of cooperative features using measurements from the access network as well as
from the BBU-pool. In addition optimization over the fronthaul network will ensure high rate
transmissions between the BBU-pool and the RRH necessary for cooperative schemes in future
networks with higher bandwidth.
IPR:
No patents submitted.
Contact details:
4. Eth2CPRI prototype implementation

Description: In C-RAN, The LTE BBU sends the digitalized RF signal to a RRH under the form of digital
IQ samples. This requires a higher data rate than the one which would be seen in the corresponding
backhaul network, making cost-savings in the fronthaul network particularly difficult. As a result, the
network has to work under great constraints both in terms of bandwidth and latency. The IQ Samples
are generally transported on fiber by a well known protocol called CPRI (Common Protocol for Radio
Interface). It is based on high-quality synchronous transmission schemes that are optimal over short
distances, but which can be extremely difficult and expensive over bigger and more complex networks.
So far no viable optimized solution exists to solve the problem. In Task 6.4, we presented a solution
that is addressing these problems. The innovation of the solution is twofold: First, it replaces CPRI by
a new radio transport protocol based on Ethernet. This protocol is called sometimes CPRI over
Ethernet, which, we will see is an error, since it does not transport exactly the same content. The
second is a new functional breakdown of the LTE PHY layer between the Server Unit (SU) and the
Remote Unit (RU). Such a change will allow generating the constant deterministic cell traffic locally in
the RU and as a consequence, there will be less information to transmit between the RU and the SU.
This solution can adapt different C-RAN deployments and ultimately, the RRH itself preprocess the IQ
samples and send/receive the results over Ethernet. This solution can be also relevant for the small
cells market on copper.
Exploitation: The proposed technique implemented in our prototype will be exploited in different CRan deployments and in small cell market.
Market: The developed implementation will target the telecommunication equipment manufacturing
companies.
Further development: Future work will cover addressing the delay and jitter constraints by means of
prioritization, preemption and source scheduling.
IPR:
No patents submitted.
Contact details:
5. ESPAR Prototype
Description: Electronically Steerable Parasitic Antenna Radiator (ESPAR) antennas have only one
active element that is fed with the radio frequency while the others are parasitic. These parasitic
elements are in a close vicinity to the active one, in such a topology that allows them to interact with
it. This interaction is in essence strong electromagnetic (EM) coupling between the closely spaced
elements. We can take advantage of this strong EM coupling and by varying the parasitic elements we
can steer or reshape the far-field radiation pattern (beam) of the antenna. The basic concept of ESPAR
antennas is that the active element resonates at the desired frequency (which in our case is 2.6 GHz)
while its input resistance is matched to the RF generator (50Ohm in our case for maximum power
transfer) and the parasitics are terminated with either inductive or capacitive loads. This way, by
varying the values of the loads of the parasitic elements, we can reshape or steer the beam. Our
proposed ESPAR antenna consists of five patches which form a cross. It is built on commercially
available FR-4 substrate material with dielectric constant (r = 4.4) and loss tangent (tan = 0.002).
The front side of the board is etched with the active and the four parasitic patches, while the back side
is the common ground. Each square patch is 26mm by 26mm. The distance between all patches is
2mm. The active element (which is at the center) is fed through via feeding point. It has to be noted
that in order for the active element to have input impedance equal to 50Ohms, the via feeding point
has to be off-centered downwards by 8mm. The overall FR-4 board dimensions are 100mm by 100mm.
The ESPAR antenna design is presented in Figures 1 and 2. During HARP demo, two RRHs are used.
Each one will be connected to an ESPAR antenna. Every RRH will switch between three predefined
beams. In the demo, the cooperating RRHs will each choose one of the predefined beams so that the
chosen beam combination maximizes the sum rate of the links to their intended receivers, based on
SINR feedback. Furthermore, it is possible to precode over the chosen beams, based on the fed-back
channels seen at the receivers through the beams.
Figure1: ESPAR Antenna Design
Figure 2: ESPAR antenna design (tilted view) where via hole positions are shown.
Exploitation: Incorporating ESPAR antennas in the RRH will allow transmitters (Tx) to focus more easily
on their intended receivers (Rx) while using only a single RF chain compared to conventional uniform
linear arrays ULAs. The antenna design procedure will assist to the faster integration of RRH
techniques in future C-RAN based cellular systems. The ESPAR techniques can offer, under certain
circumstances, a large improvement in spectral efficiency and boost the overall system-level capacity.
ESPARs make a very strong candidate for such systems due to their beam-shaping abilities, compact
size, and unique RF chain need.
Market: Parasitic arrays could bring in a significant improvement at the performance of future
communication systems. They could contribute to the acceleration of miniaturization of base stations
down to RRHs and, at the same time, they could equip consumer devices such as laptops or tablets
(and potentially smartphones) providing improved performance but maintaining a low manufacturing
cost. Moreover, they could be the enabling technology of future communication systems with massive
antenna arrays at the transmitter side where they would significantly reduce the total number of RF
(Radio Frequency) chains and would provide advanced beam-forming capabilities. Both aspects are
promising candidates for future cellular systems and have been highly studied over the last few years.
In that sense, there is an important potential in the telecommunications market since major
manufacturers have already shown interest in parasitic prototypes.
Further development: Future steps for the ESPAR design for RRH involve, firstly, the investigation of
its beamforming capabilities. For these cases, different topologies and/or larger number of elements
may be needed. A second topic for further research remains the miniaturization problem. Although
ESPARs are already very compact they may not fit in future, handheld devices.
Expected impact: The constantly increasing demand for data over mobile networks and, at the same
time, the exploding growth of connected devices (with smartphones being the leading edge) has
created the urgent need for higher system capacities. The combination of parasitic array antennas
with RRHs could form an attractive solution for the next generation cellular systems that will address
the aforementioned problem.
IPR: No patents submitted. However, the designs exist at the projects public documents while early
prototypes have been fabricated and used for lab experimentation.
Contact details:
Constantinos Papadias (cpap@ait.gr)
2.2.3 Path towards standardization

The findings of HARP have given an improved understanding of the ESPAR based RRHs techniques that
can bring significant gains in the spectrum efficiency of cellular networks. These insights will become
particularly relevant to consider when evolving cellular communication standards to meet future
needs for capacity and user experience. It is foreseen that the continued densification and traffic
growth will spark new initiatives to C-RAN deployment in the cellular networks. For this purpose,
relevant key HARP findings will be recognized and used as established facts [by Alcatel-Lucent] when
defining future study items in 3GPP related to coming releases of the LTE standard.
If the use of C-RAN becomes a dedicated topic for the 3GPP standardization, the wealth of results
generated by HARP will be proposed as a starting point for the 3GPP work. It is expected that the final
technical report and selected publications will be brought to the attention of 3GPP and used towards
the development of the inclusion of ESPAR based RRHs in the LTE standard.
Furthermore, the framework and methodology developed in HARP will be considered for adoption
in the ways of working in the relevant standardization activities in 3GPP. It is believed that some of
the HARP techniques will be applicable in a much wider range of studies than has been considered in
the project since they are of a more fundamental nature and not necessarily C-RAN specific.
Finally, it is believed that the industry-academia contacts that have been fostered within HARP will
allow not only a one-directional flow of knowledge from the project towards standardization, but that
it will be possible to feed back further challenges and scenarios identified by 3GPP towards the
academia and to bring the application of the HARP results in 3GPP to the attention of the respective
originators. Thereby it will be possible to strengthen the connections between academic research and
standardization to the benefit of both parties.
3. Report on societal implications

Replies to the following questions will assist the Commission to obtain statistics and
indicators on societal and socio-economic issues addressed by projects. The questions
are arranged in a number of key themes. As well as producing certain statistics, the
replies will also help identify those projects that have shown a real engagement with
wider societal issues, and thereby identify interesting approaches to these issues and best
practices. The replies for individual projects will not be made public.
General Information (completed automatically when Grant Agreement number is

entered.
Grant Agreement Number:

Title of Project:
Name and Title of Coordinator:
318489
High capacity network Architecture with Remote radio heads & Parasitic Antenna Arrays
Parasitic antenna arrays
Prof. Tharmalingam Ratnarajah
Ethics
1. Did your project undergo an Ethics Review (and/or Screening)?
If Yes: have you described the progress of compliance with the relevant Ethics
Review/Screening Requirements in the frame of the periodic/final project reports?
No
Special Reminder: the progress of compliance with the Ethics Review/Screening Requirements should be
described in the Period/Final Project Reports under the Section 3.2.2 'Work Progress and Achievements'
2.
Please indicate whether your project involved any of the following issues (tick box) :
NO
RESEARCH ON HUMANS
Did the project involve children?
NO
Did the project involve patients?
NO
Did the project involve persons not able to give consent?
NO
Did the project involve adult healthy volunteers?
NO
Did the project involve Human genetic material?
NO
Did the project involve Human biological samples?
NO
Did the project involve Human data collection?
NO
RESEARCH ON HUMAN EMBRYO/FOETUS
Did the project involve Human Embryos?
NO
Did the project involve Human Foetal Tissue / Cells?
NO
Did the project involve Human Embryonic Stem Cells (hESCs)?
NO
Did the project on human Embryonic Stem Cells involve cells in culture?
NO
Did the project on human Embryonic Stem Cells involve the derivation of cells from Embryos?
NO
PRIVACY
NO
Did the project involve processing of genetic information or personal data (eg. health, sexual
lifestyle, ethnicity, political opinion, religious or philosophical conviction)?
Did the project involve tracking the location or observation of people?
NO
RESEARCH ON ANIMALS
Did the project involve research on animals?
NO
Were those animals transgenic small laboratory animals?
NO
Were those animals transgenic farm animals?
NO
Were those animals cloned farm animals?
NO
Were those animals non-human primates?
NO
RESEARCH INVOLVING DEVELOPING COUNTRIES
Did the project involve the use of local resources (genetic, animal, plant etc)?
NO
Was the project of benefit to local community (capacity building, access to healthcare, education
etc)?
NO
DUAL USE
Research having direct military use
NO
Research having the potential for terrorist abuse
NO
Workforce Statistics
3.
Workforce statistics for the project: Please indicate in the table below the number of
people who worked on the project (on a headcount basis).
Type of Position
Number of Women
Number of Men
Scientific Coordinator
Work package leaders

Experienced researchers (i.e. PhD holders)
10
PhD Students
Other
4.
How many additional researchers (in companies and universities) were

recruited specifically for this project?
Of which, indicate the number of men:
20
D Gender Aspects
5.
6.
Yes
No
Which of the following actions did you carry out and how effective were they?
Not at all
effective

7.

Did you carry out specific Gender Equality Actions under the project?
Design and implement an equal opportunity policy

Set targets to achieve a gender balance in the workforce
Organise conferences and workshops on gender
Actions to improve work-life balance
Very
effective

Other:
Was there a gender dimension associated with the research content i.e. wherever people were
the focus of the research as, for example, consumers, users, patients or in trials, was the issue of gender
considered and addressed?
Yes- please specify
No
Synergies with Science Education
8.
Did your project involve working with students and/or school pupils (e.g. open days,
participation in science festivals and events, prizes/competitions or joint projects)?
Yes- please specify
9.
No
Did the project generate any science education material (e.g. kits, websites, explanatory
booklets, DVDs)?
Yes- please specify
www.fp7-harp.eu
No
Interdisciplinarity
10.
Which disciplines (see list below) are involved in your project?

Main discipline1:

Associated discipline1:
Associated discipline1:
Engaging with Civil society and policy makers
11a
Did your project engage with societal actors beyond the research
community? (if 'No', go to Question 14)

Yes
No
11b If yes, did you engage with citizens (citizens' panels / juries) or organised civil society
(NGOs, patients' groups etc.)?
No
Yes- in determining what research should be performed
Yes - in implementing the research
Yes, in communicating /disseminating / using the results of the project

Yes
11c In doing so, did your project involve actors whose role is mainly to

No
organise the dialogue with citizens and organised civil society (e.g.
professional mediator; communication company, science museums)?
12. Did you engage with government / public bodies or policy makers (including international
organisations)

No
Yes- in framing the research agenda
Yes - in implementing the research agenda
Yes, in communicating /disseminating / using the results of the project
13a Will the project generate outputs (expertise or scientific advice) which could be used by
policy makers?
Yes as a primary objective (please indicate areas below- multiple answers possible)
Yes as a secondary objective (please indicate areas below - multiple answer possible)
No
13b If Yes, in which fields?
Agriculture
Energy
Human rights
Audiovisual and Media
Enlargement
Information Society
Budget
Enterprise
Institutional affairs
Competition
Environment
Internal Market
Consumers
External Relations
Justice, freedom and security
Culture
External Trade
Public Health
Customs
Fisheries and Maritime Affairs
Regional Policy
Development Economic and

Monetary Affairs
Food Safety
Research and Innovation
Foreign and Security Policy
Space
Fraud
Taxation
Humanitarian aid
Transport
Education, Training, Youth

Employment and Social Affairs
Insert number from list below (Frascati Manual).
13c If Yes, at which level?

Local / regional levels
National level
European level
International level
Use and dissemination
14.
How many Articles were published/accepted for publication in

peer-reviewed journals?
23
To how many of these is open access2 provided?
23
How many of these are published in open access journals?
23
How many of these are published in open repositories?
To how many of these is open access not provided?
Please check all applicable reasons for not providing open access:
publisher's licensing agreement would not permit publishing in a repository
no suitable repository available
no suitable open access journal available
no funds available to publish in an open access journal
lack of time and resources
lack of information on open access
other3:
15.
How many new patent applications (priority filings) have been made?
("Technologically unique": multiple applications for the same invention in different

jurisdictions should be counted as just one application of grant).
16.
Indicate how many of the following Intellectual

Property Rights were applied for (give number in
each box).
Trademark
Registered design
Other
17.
How many spin-off companies were created / are planned as a direct

result of the project?
Indicate the approximate number of additional jobs in these companies:
18. Please indicate whether your project has a potential impact on employment, in comparison
with the situation before your project:

In small & medium-sized enterprises
Increase in employment, or

Safeguard
employment,
or
In large companies

None of the above / not relevant to the project
Decrease in employment,
Difficult to estimate / not possible to quantify
19. For your project partnership please estimate the employment effect resulting
directly from your participation in Full Time Equivalent (FTE = one person working
fulltime for a year) jobs:
Indicate figure:
Difficult to estimate / not possible to quantify
Media and Communication to the general public
20.
As part of the project, were any of the beneficiaries professionals in communication or

media relations?
Yes
No
21.
As part of the project, have any beneficiaries received professional media / communication
training / advice to improve communication with the general public?
Yes
No
22
Which of the following have been used to communicate information about your project to
the general public, or have resulted from your project?

23
Press Release
Media briefing
TV coverage / report
Radio coverage / report
Brochures /posters / flyers
DVD /Film /Multimedia

Coverage in specialist press

Coverage in general (non-specialist) press
Coverage in national press
Coverage in international press
Website for the general public / internet
Event targeting general public (festival, conference,
exhibition, science caf)
In which languages are the information products for the general public produced?

Language of the coordinator

Other language(s)
English
Question F-10: Classification of Scientific Disciplines according to the Frascati Manual 2002 (Proposed
Standard Practice for Surveys on Research and Experimental Development, OECD 2002):
FIELDS OF SCIENCE AND TECHNOLOGY

1.
NATURAL SCIENCES
1.1
Mathematics and computer sciences [mathematics and other allied fields: computer sciences and
other allied subjects (software development only; hardware development should be classified in the
engineering fields)]
1.2
Physical sciences (astronomy and space sciences, physics and other allied subjects)
1.3
Chemical sciences (chemistry, other allied subjects)
Open Access is defined as free of charge access for anyone via Internet.
For instance: classification for security project.

1.4
Earth and related environmental sciences (geology, geophysics, mineralogy, physical geography and
other geosciences, meteorology and other atmospheric sciences including climatic research,
oceanography, vulcanology, palaeoecology, other allied sciences)
1.5
Biological sciences (biology, botany, bacteriology, microbiology, zoology, entomology, genetics,

biochemistry, biophysics, other allied sciences, excluding clinical and veterinary sciences)
ENGINEERING AND TECHNOLOGY
2.1
Civil engineering (architecture engineering, building science and engineering, construction

engineering, municipal and structural engineering and other allied subjects)
2.2
Electrical engineering, electronics [electrical engineering, electronics, communication engineering and

systems, computer engineering (hardware only) and other allied subjects]
2.3.
Other engineering sciences (such as chemical, aeronautical and space, mechanical, metallurgical and
materials engineering, and their specialised subdivisions; forest products; applied sciences such as
geodesy, industrial chemistry, etc.; the science and technology of food production; specialised
technologies of interdisciplinary fields, e.g. systems analysis, metallurgy, mining, textile technology
and other applied subjects)
3.
MEDICAL SCIENCES
3.1
Basic medicine (anatomy, cytology, physiology, genetics, pharmacy, pharmacology, toxicology,

immunology and immunohaematology, clinical chemistry, clinical microbiology, pathology)
3.2
Clinical medicine (anaesthesiology, paediatrics, obstetrics and gynaecology, internal medicine,

surgery, dentistry, neurology, psychiatry, radiology, therapeutics, otorhinolaryngology,
ophthalmology)
3.3
Health sciences (public health services, social medicine, hygiene, nursing, epidemiology)
4.
AGRICULTURAL SCIENCES
4.1
Agriculture, forestry, fisheries and allied sciences (agronomy, animal husbandry, fisheries, forestry,
horticulture, other allied subjects)
4.2
Veterinary medicine
5.
SOCIAL SCIENCES
5.1
Psychology
5.2
Economics
5.3
Educational sciences (education and training and other allied subjects)
5.4
Other social sciences [anthropology (social and cultural) and ethnology, demography, geography
(human, economic and social), town and country planning, management, law, linguistics, political
sciences, sociology, organisation and methods, miscellaneous social sciences and interdisciplinary ,

methodological and historical S1T activities relating to subjects in this group. Physical anthropology,
physical geography and psychophysiology should normally be classified with the natural sciences].
6.
HUMANITIES
6.1
History (history, prehistory and history, together with auxiliary historical disciplines such as
archaeology, numismatics, palaeography, genealogy, etc.)
6.2
Languages and literature (ancient and modern)
6.3
Other humanities [philosophy (including the history of science and technology) arts, history of art, art
criticism, painting, sculpture, musicology, dramatic art excluding artistic "research" of any kind,
religion, theology, other fields and subjects pertaining to the humanities, methodological, historical
and other S1T activities relating to the subjects in this group]

d1 7 Final Project Report

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

d1 7 Final Project Report

Hochgeladen von

Copyright:

Verfügbare Formate

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

SEVENTH FRAMEWORK PROGRAMME (FP7)

FINAL PROJECT REPORT D1.7

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

Author Contact Information

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

1. Final Publishable Summary report

To explore the fundamental limits (bandwidth / power efficiency) of distributed multi-antenna

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

To extend the above-mentioned transceiver and channel feedback techniques to

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

Figure 1: HARPs ESPAR-based RRH wireless network architecture.

University of Edinburgh (UEDIN) (UK, Co-ordinator)

Athens Information Technology (AIT) (Greece, Technical Leader)

Eurecom (EUR) (France)

Imperial College London (IMPERIAL) (UK)

Norwegian University of Science and Technology (NTNU) (Norway)

Technical University Denmark (DTU) (Denmark)

MTI RadioComp (RADIOCOMP) (Denmark)

Alcatel-Bell-Lucent France (ALBLF) (France)

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

Orange Labs (ORANGE) (France)

1.2 Work Packages

WP1: Management of the Project (Leader UEDIN)

WP2: Dissemination Activities and Exploitation Plan (Leader MTI RADIOCOMP)

WP3: Network requirements and Fundamental Limits (Leader ORANGE)

WP4: Single-point ESPAR-enabled RRH Access (Leader AIT)

WP5: Multi-point ESPAR-enabled RRH Access (Leader IMPERIAL)

WP6: Aggregation Network (Leader DTU)

WP7: Experimentation and Demonstrations (Leader ALBLF)

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

1.3 Executive Summary

1.3.1 Technical Highlights

we devised a novel beamforming-based approach that enables us to perform arbitrary

we studied the application of this solution in multi-cell Cloud-RAN-based ESPAR-enabled

user-centric interference nulling in multi-antenna heterogeneous networks was investigated.

T5.3 Cloud-RAN Implementation of Cooperative Techniques

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

we investigated the affect of specific aggregation network dimensioning approaches to

T6.3 Protocol Extensions Design and Implementation

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

1.3.2 Key findings

Mapping strategies to construct PMH signals have been proposed.

A novel beamforming-based technique for performing arbitrary channel-dependent

A radio transmission protocol that exploits the aforementioned precoding technique to

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

Several variants (including reduced feedback overhead methods) of transmission schemes

A hierarchical RS scheme that exploits a two-tier precoding structure to mitigate multiuser

User-selective joint detection for the uplink has been studied.

The effect of MFH dimensioning at application-level satisfaction has been quantified.

ESPAR-enabled precoding-based joint Zero-Forcing Beamforming has proven to outperform

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

1.3.3 Collaboration in the HARP Project

1.3.4 Summary of Year III Conclusions

A radio communication protocol for performing multi-cell cooperative precoding transmission

Several variants of multi-cell coordinated transmission schemes (including reduced feedback

The optimum RS precoder design has been presented and evaluated.

User-selective joint detection schemes have been implemented in a Cloud-RAN prototype.

The effect of dimensioning on application-level latency penalties has been studied.

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT

ESPAR-enabled precoding-based joint Zero-Forcing Beamforming has been shown to

FP7 PROJECT 318489 (HARP): PROJECT FINAL REPORT