2. 1646
Y. Banday et al.
1 3
the probability of an outage or successful decoding of a transmission, SINR metric is used.
Evaluating the performance of a network using SINR is difficult as it requires the esti-
mation of transmission strategies like; transmission power, Media Access Control (MAC)
protocol and antenna patterns. In case of networks using mmWave backhaul the quanti-
fying SINR becomes more difficult and complex because of high penetration loss, first
order reflection and use of MIMO antenna pattern which introduces elements of random-
ness [1–3]. Furthermore the different components of SINR viz.; Network Topology, Wire-
less Channel and Transmission strategy need to be mathematically modeled for estimating
interference in a link.
In 5G network architecture, implementation of heterogeneous networks create multi-
ple network tiers of different sizes with different transmission powers and large number
of smart and heterogeneous devices connected to each other. Transmissions from these
devices cause interference that needs to be suppressed. In these networks with increase in
deployment of cells RF noise floor increases and hence Co-Channel interference becomes
a major obstacle in improvement of network throughput. In recent research it has been
mentioned that combined effect of interference management on User Equipment (UE) side
and its network-side counterpart has greater effect than sum of their separate effects [4, 5].
Moreover in Release 12, a study has been started to support UE-side interference manage-
ment. Keeping this in consideration it is believed that in 5G networks UE-side interference
management should be given equal weightage as the network-side interference manage-
ment [6]. This paper is focussed on interference management in 5G Networks on network-
side under dense urban network environment.
The rest of paper is organized in four sections as follows. In Sect. 2 we describe the
system model, network topology and brief overview of interference models. Appropriate
interference model with respect to network topology is chosen for estimation of link outage
events under different network scenarios. Configuration and cell parameters according to
report ITU-R.M [7] are used to create dense urban eMBB transmitter sites. In this work we
have chosen NIT Entrance Road as central location of cellular network. This is followed by
mapping of SINR using UPA in Sect. 3. The paper is concluded in Sect. 4.
2 System Model
The Fig. 1 gives a network scenario with a pair of transmitter and its intended receiver
link denoted by i. The transmitters Tx1 and Tx2 act as potential interferers to transmis-
sions between (Txi)0 and its intended receiver (Rxi)0. All the set of interferer transmitters
excluding the intended transmitter are denoted by 𝔗 . The distance between the transmitter
(Txi)0 and the reference receiver (Rxi)0 is denoted by
di.
The SINR at receiver i is given by the equation
S = signal power, ik = interference from the interferer k
Both signal and interference power from interferer k are random variables with their
distributions depending on the various factors like transmit power, interference manage-
ment, antenna pattern and beam-forming techniques used. The above equation can be fur-
ther written as:
(1)
𝛾i =
S
∑
k∈𝔗
ik + η
3. 1647
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
where pi
=
power transmitted by (TXi)0, gi
Ch
=
Channel gain between TXiand reference
receiver (RX)0, gi
Tx
=
Antenna gain at TXi towards the reference receiver, gi
Rx
= Antenna
gain at reference receiver towards TXi, η = white Gaussian noise power
This equation shows dependence of SINR on the different components as mentioned
earlier. According to Eq. 2 directional antennas will provide high SINR because of high
directional gain. In this paper SINR is evaluated and compared using different antennas
with different radiation pattern and gain.
Furthermore, if β denotes the SINR threshold, an outage will occur if γ
β. A sta-
tistical model that provides an estimate of an outage is called a statistical distribution
describing [8]. To analyse the performance of mmWave systems, accurate interference
models are needed. However, accurate models of interference are mathematically very
complex. Some interference models attempt to approximate the probability of an outage
by considering certain interference components and ignoring others depending on a par-
ticular network scenario. Thus the interference model in which the simplified interfer-
ence methods achieve a good balance between accuracy and simplicity should be chosen
to evaluate network performance [9].
We consider a hexagonal layout of a network with three cells per site that can be
increased according to the requirement. We have chosen this configuration according to
guidelines given in Report ITU-R M [7] for high-speed urban macro-cell environments.
Basic hexagonal layout with three cells per site is shown in Fig. 2. Wrap-around con-
figuration of 19 sites with total of 57 cells is simulated. User equipments are uniformly
distributed over whole network. The distance between the centers of adjacent cell sites
is called inter-site distance (ISD). In dense urban-microcell and macro-cell eMBB, ISD
(2)
𝛾i =
PigTx
i
gRx
i
gCh
i
∑
k∈𝔗 pkgTx
k
gRx
k
gCh
k
+ 𝜂
Fig. 1 SINR at intended receiver i
4. 1648
Y. Banday et al.
1 3
is 200 m and 500 m respectively [10]. Two layers of cells that are formed in this net-
work are macro-cell layer and micro-cell layer.
Figure 3 shows the Base stations (BSs) in macro-layer placed in a regular grid with
hexagonal layout. This figure shows that three macro sites share a common BS which
forms three Transmission Reception points (TRxPs) at a site. The micro-layer consists of 3
micro-sites within a macro transmission reception point as shown in Fig. 4. The placement
Fig. 2 eMBB Urban cell hexago-
nal layout
Fig. 3 Macro-layer BSs placed in a regular grid
5. 1649
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
of BSs in micro layer is to be decided by the proponent. To simplify the analysis, here we
have chosen a center of site for location of the BS.
This type of topology gives rise to the interference. In most of the scenarios Co-channel
interference occurs due to reuse of same frequency in nearby cell. To model this inter-
ference caused by multiple set of interferers, different models are available with some
limitations. Now, let us consider a downlink scenario for dense urban-eMBB operating at
mmWave frequencies. The number of interferers is a random variable with uniform distri-
bution as shown in Fig. 5
Dmicro-sites
R
Macro TRxPs
MICRO-SITE
Fig. 4 Layout of a micro-layer
i1
BS0
UE 0
UE1
UE2
BS1
i2
Fig. 5 Contribution of interferers (Interference Ball Model)
6. 1650
Y. Banday et al.
1 3
To compute the effect of aggregated interference from transmitters (interferers) beyond
the interference range, the interference ball model (IBM) is used. IBM takes in considera-
tion all the near-field interferers within a certain distance. This model is adopted for the
performance evaluation and protocol design of wireless networks [3, 11, 12].
The UEs within the ball, located at certain specific distance from BS are assumed to be
in LOS condition while as the UEs located outside the ball are considered to be in non-
LOS or in an outage condition. This assumption simplifies the mathematical analysis. In
Fig. 5 BS0 is transmitting to
UE0 under LOS condition.
BS1 is also transmitting to
UE1
and UE2 under LOS condition. Without assumption of IBM there is a fair chance that the
signal from
UE1, UE2 and
BS1can reach
UE0, if
UE0 is in the interference range of
BS1.
This additional interference will considerably decrease the value of SINR at
UE0. Also the
estimation of effect of interference at
UE0 will become complex. Hence implementation of
IBM model for dense urban eMBB networks has two main reasons:
1. It simplifies the calculation of interference at
UE0by ignoring all the interferers outside
the ball.
2. In case an interferer is mobile the Interference range will also change. Interferer can be
closest or farthest from the receiver depending on its location. In such cases IBM gives
the accurate results at the cost of high complexity [13].
If uplink design is considered then Interference over Thermal (IOT) power comes into
play. In this design interference limit is an important factor to determine the total uplink
capacity of a network. The interference limit
qm
can be represented as [14]:
for some constant k ≥ 1
From this definition interference
qiat each BS i is limited and should not exceed the
thermal noise
ni. This factor kis called the Interference over Thermal (IOT) expressed in
dB.
IOT confines interference to the cell and limits the power required for new UEs to access
the network. IOT values range from 3 to 10 dB in commercial networks.
3
Configuration for Evaluation of Network
In ITU-R.M [7] different evaluation configurations are defined for each test environment.
There are three evaluation configurations for dense urban eMBB networks viz A, B C.
To evaluate spectral efficiency and mobility, configuration A and B is used and for user
experienced data rate evaluation, configuration C is used. In this paper configuration B is
used for implementation of the cellular network in dense urban environment.
In this paper we have chosen NIT Srinagar as Centre location site for transmission with
latitude 34.125194 and longitude 74.837389 as shown in Fig. 6. Nineteen cell sites are
created around this location, with each site having three microsites as shown in Fig. 2. Dis-
tances and angles for each site in a ring is defined. Angles for inner ring and outer ring of
(3)
qm
= kn
(4)
IOT = 10 log10 (k)
7. 1651
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
six sites range from 30° to 360° with step size of 60°. For middle ring site angles vary from
0° to 300° with step size of 60°.
Here we have chosen a mmWave carrier signal with frequency of 30 GHz and 25 m
height of BS antenna above rooftop. For bandwidth of 40 MHz total transmit power per
TRxP is 37 dBm. With these transmitter parameters and configuration parameters of
Table 1, cell transmitter sites are created as shown in Fig. 7.
For evaluation purposes different characteristics of antenna at BS and UE are specified.
Each BS has multiple antenna panels with different arrays. Here we have used a uniform
planar array (UPA) panel. This array comprises of
Mg × Ng antenna panels with each panel
having M × N antenna elements where
Mg is number of panels in a row,
Ng is number of
panels in a column and M is number of antenna elements having same polarization in a
column and N is number of columns [7].
Both horizontal and vertical radiation patterns of antenna need to be defined for simula-
tion. Horizontal pattern of an antenna element in dB is given by [7]
Table 1 Evaluation configuration parameters [7]
Parameters Spectral efficiency and mobility evaluation
Configuration B
Carrier frequency for evaluation 1 layer (Macro) with 30 GHz
BS antenna height 25 m
Total transmit power per TRxP 40 dBm for 80 MHz bandwidth
37 dBm for 40 MHz bandwidth
e.i.r.p. should not exceed 73 dBm
UE power class 23 dBm, e.i.r.p. should not exceed 43 dBm
Percentage of high loss and low loss building type 20% high loss, 80% low loss
Inter-site distance 200 m
Number of antenna elements per TRxP Up to 256 Tx/Rx
Number of UE antenna elements Up to 32 Tx/Rx
Device deployment 80% indoor, 20% outdoor (in–car)
Randomly and uniformly distributed over the area
under Macro layer
UE mobility model Fixed and identical speed |v| of all UEs of the same
mobility class, randomly and uniformly distrib-
uted direction.
UE speeds of interest Indoor users: 3 km/h
Outdoor users (in-car): 30 km/h
Inter-site interference modeling Explicitly modelled
BS noise figure 7 dB
UE noise figure 10 dB
BS antenna element gain 8 dBi
UE antenna element gain 5 dBi
Thermal noise level − 174 dBm/Hz
Traffic model Full buffer
Simulation bandwidth 80 MHz for TDD,
40 MHz + 40 MHz for FDD
UE density 10 UEs per TRxP
Randomly and uniformly distributed over the area
under Macro layer
8. 1652
Y. Banday et al.
1 3
where −180° ≤ φ″ ≤ 180° and, min[.] gives the minimum function, φ3dB = horizontal 3 dB
beamwidth, Am = maximum attenuation.
Radiation pattern of an antenna element in vertical direction is given by;
(5)
AE,H
(
𝜑�
)
= − min
[
12
(
𝜑��
𝜑3dB
)2
, Am
]
, 𝜑3dB = 65
◦
, Am = 30
Fig. 6 Transmitter at center of site
Fig. 7 OpenStreetMap of 19 transmitter sites at NIT Entrance Rd
9. 1653
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
where 0° ≤ θ″ ≤ 180°, φ3dB
=
vertical 3 dB beamwidth, θtilt = tilt angle, SLA = maximum
side lobe level attenuation.
When θ″ = 0° antenna pattern directs to the zenith and at θ″= 90°, antenna directs to the
horizon. Combining Eq. 5 and Eq. 6 we get the combined 3D pattern of an antenna element
in horizontal and vertical direction i.e.;
where A″(θ″, φ″) = relative antenna gain in (θ″, φ″) direction.
For evaluation of antennas at UE side with 30 GHz frequency we have assumed a direc-
tional panel. Orientation of each panel varies according to array bearing angle Ωmg,ngand
array downtilt angle Θmg,ng. Let (Ωmg,ng, Θmg,ng) represent the orientation angles of a panel,
then first panel will take(Ω00,Θ00) orientation. This is defined as orientation of UE. The
horizontal and vertical radiation pattern of an antenna at UE is given by Eqs. 8 and 9
respectively [7]
From Eqs. 5, 6, 7 we can plot a 3D radiation pattern of antenna element at 30 GHz fre-
quency as shown in Fig. 8.
To plot SINR map of single antenna element we use the free space propagation model
and receiver parameters as given in Table 1. In this paper we have plotted the SINR on
ESRI world street map using real geospatial information. The center frequency is 30 GHz
for which bandwidth specified is 80 MHz and 40 MHz with transmit power 40 dBm and
37 dBm respectively. The receiver noise power,
N0 is calculated from the equation given
below [15]
Table 2 shows the variations in UE noise power when UE noise figure is 10 dB and
13 dB with fixed values of transmission power per TRxP, BS noise figure and UE element
gain. When UE noise figure is 10 dB, the received noise power obtained is − 84.9691 dBm
at 40 dBm transmit power and 80 MHz bandwidth. Repeating for the same values of trans-
mit power and bandwidth at 13 dB UE noise figure, received noise power is −
81.9691
dBm. This is quiet higher than − 84.9691 dBm. Hence it can be observed that UE will have
better performance at 10 dB than 13 dB. Also as the bandwidth is halved, noise power of
UE decreases by approximately 3 dB.
Now SINR map of transmitter sites can be visualized as in Fig. 9. The radiation pat-
tern of each transmitter site is displayed as a colored contour. The intensity of color
shows the maximum signal strength available to the UE. The signal strength of each
(6)
AE,V
(
𝜃��
)
= −min
[
12
(
𝜃��
− 90
◦
𝜑3dB
)2
, SLAV
]
, 𝜃3dB = 65
◦
, SLAV = 30
(7)
A��
(
𝜃��
, 𝜑��
)
= −min
{
−
|
|
|
AEV
(
𝜃��
)
+ AEH
(
𝜑��
)|
|
|
Am
}
(8)
AE,H
(
𝜑��
)
= − min
[
12
(
𝜑��
𝜑3dB
)2
, Am
]
, 𝜑3dB = 90
◦
, Am = 25
(9)
AE,V
(
𝜃��
)
= −min
[
12
(
𝜃��
− 90
◦
𝜑3dB
)2
, SLAV
]
, 𝜃3dB = 90
◦
, SLAV = 25
(10)
N0 = −174 + 10 log10 (BW) + UEm noise power
10. 1654
Y. Banday et al.
1 3
transmitter varies from 3 to 5 dB. UE picks up the signal with greatest signal strength
and weak signals transmitted by other BSs act as interference.
At UE to overcome the interference of signals originating from others cells we use a
UPA. This increases directional gain and peak SINR values at each site. In UPA we are
using 8 × 8 uniform rectangular array having uniform spacing λ/2 between elements.
Figure 10 shows that directionality increases four times as compared to the direction-
ality of single antenna element shown in Fig. 9. The beamwidth of an array is narrow
which results in remarkable increase of SINR and improvement in performance. Anten-
nas can radiate and receive maximum in a specific direction. Furthermore the directiv-
ity of sidelobes ranges from −
25
dBi to 20
dBi which is very low and thus reduces
the interference. Using the free space propagation model, curvature of earth and other
effects of environment are neglected. To focus the signal towards an intended UE,
mechanical downtilt is applied to each transmitter. Here the downtilt of 15° is applied to
each antenna array [7]. Each antenna array can be assigned to a transmitter cell. Using
the real geospatial information, SINR map for transmitter sites with UPA is obtained as
shown in Fig. 11.
Clearly the SINR of each transmitter lies in red zone of color bar which shows signal
strength of an order of 15 dB to 17 dB, much higher compared to 5 dB obtained by using
single element. Implementation of UPA at each transmitting site increases directivity com-
pared to use of single antenna element. This also results in increase of signal power of
transmitters located on the perimeter of a cellular site. Hence the interference in the lobes
facing outwards on the periphery of a map is reduced.
Now instead of using a Free space path loss model we are using a Close In (CI) propaga-
tion model. Due to the accuracy, simplicity and high sensitivity performance owing to its close
Fig. 8 3D Radiation Pattern of single antenna element
11. 1655
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
Table 2 UE
Noise
power
Frequency
f
q
(GHz)
UE
element
gain
(
dB
i
)
BS
noise
figure
(dB)
Bandwidth
(MHz)
Transmission
power
per
TRxP
(dBm)
UE
received
noise
power
(dBm)
UE
noise
figure
(dB)
30
5
7
80
40
− 84.9691
10
40
37
− 87.9794
80
40
− 81.9691
13
40
37
− 84.9794
12. 1656
Y. Banday et al.
1 3
Fig. 9 SINR map of transmitter sites using single antenna element
Fig. 10 Radiation pattern of an uniform planar array
13. 1657
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
in free space reference point CI model is most suitable for outdoor networks [15]. The equa-
tion of CI propagation model is given by [16]
where d ≥ d o f is frequency in GHz,
do is close-in free space reference distance, n denotes
Path Loss Exponent(PLE) and χCI
σ
is zero mean Gaussian random variable with standard
deviation σ in dBs. From equation [12] we can see that term 10ngives path loss in terms
of distance ten times the distances starting at
doin decibels. Hence if
dois 1 m, computation
becomes easier and provides parameter stability and accuracy for outdoor Urban Micro
cell and Urban Macro cell networks [17]. 3-D separation distance between transmitter and
receiver is given by d and calculates the loss in all three coordinates of space [17–19]. Thus
we can plot SINR map using CI propagation model for 5G Urban macro-cell networks as
shown in Fig. 12.
The strength of signal is increased as indicated in colorbar, the signal power ranges from 19
to 20 dB. This reduces the value of interference as compared to free space propagation model.
From Figs. 11 and 12 we can clearly observe that the signals from different transmitters inter-
fere although high strength of signal overcomes interference in Fig. 12.
To reduce the interference within a specific cell each antenna element is replaced with
microstrip patch element. Design and analysis of network using Microstrip patch antenna is
explained in following section.
(11)
PLCI
(f, d) [dB] = FSPL (f, do) [dB] + 10n log10
(
d
do
)
+ 𝜒CI
𝜎
Fig. 11 SINR map of UPA
14. 1658
Y. Banday et al.
1 3
Fig. 12 SINR map using CI model
Fig. 13 Radiation pattern of microstrip antenna element
15. 1659
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
3.1 Evaluation of SINR Using Microstrip Patch Antenna
We have replaced the antenna element in [7] with real half-wavelength rectangular micro-
strip patch antenna model at operating frequency of 30 GHz. The dielectric substrate is air.
The radiation pattern of microstrip patch element is shown in Fig. 13.
The gain of about 9 dBi is provided with reduced value of front-to-back ratio. The maxi-
mum directivity obtained is 9.15 dBi and minimum value of directivity is approximately
−
23.9 dBi. To analyse the performance of microstrip patch in 8
×
8 antenna array, each
patch element is assigned to a cell antenna element.
Figure 14 depicts that the side lobes in microstrip patch are eliminated as compared to
directivity pattern of UPA shown in Fig. 10.
The SINR map obtained by use of a microstrip patch array shows that the transmitters
in the inner ring and middle ring transmit low power signals of an order of 11 dB. The
transmitters in outer ring show higher values of SINR ranging from 13 to 15 dB. Here
signal power decreases by 4 dB compared to 19 dB signal power obtained in Fig. 12. This
reduction of signal power is due to the higher radiation from back lobe of microstrip patch
antenna element. Moreover, the interference power obtained by using microstrip patch
antenna is lower than UPA.
To mitigate the effect of backlobe in transmitters, microstrip patch element is replaced
by cosine antenna element in an array. Cosine element has a cosine response in both azi-
muth and elevation. At backside of cosine antenna there is no response and hence it radi-
ates in a forward direction as shown in Fig. 15. The response of cosine antenna element is
given by:
Fig. 14 Directivity pattern of 8 × 8 Microstrip patch array
16. 1660
Y. Banday et al.
1 3
Fig. 15 Radiation pattern of cosine antenna element
Fig. 16 Radiation pattern of Cosine antenna array
17. 1661
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
azvec and elvec are azimuthal and elevation angles respectively. m and n are real numbers
greater than or equal to zero. Replacing each patch element in 8 × 8 antenna array by cosine
antenna element radiation pattern as shown in Fig. 16 is obtained. For millimeter wave
cellular networks cosine antenna element can improve the SINR significantly by eliminat-
ing interference from backlobes and concentrating the signal power in a particular direc-
tion. Nevertheless, the directional gain of patch array and cosine antenna array is same
but cosine array has zero backlobe radiation and minimum side lobes. To achieve high
directionality large values of exponents m and n should be used. Power response of cosine
element is given as squared value of field response.
The SINR map obtained by using cosine antenna array is shown in Fig. 17. On com-
paring with Fig. 18 it is evident that signal strength of transmitters located in outer ring
increases from 13 to 17 dB and signal strength of transmitters in inner and middle ring
increases from 11 to 13 dB.
From the directivity pattern of cosine and microstrip patch element shown in Fig. 19 it
is clearly observed that both antennas have similar pattern behaviour around 0° azimuth. To
simplify the analysis we have not considered the effect of mutual coupling in this paper, but
in realistic antenna array mutual coupling cannot be ignored. Mutual coupling and presence
of sidelobes can distort the signal considerably. Figure 19 shows that cosine element has no
backlobe while a microstrip has a significant backlobe at 80° azimuth which results in low
Front-to-back ratio. Also the directive gain of sidelobes in patch is higher than cosine element.
This suppression of sidelobes achieved in cosine element helps in further reducing interfer-
ence in a network.
(12)
f(azvec, elvec) = cosm
(azvec)cosn
(elvec)
(13)
P(azvec, elvec) =
[
cosm
(azvec)cosn
(elvec)
]2
Fig. 17 SINR map using cosine antenna array
18. 1662
Y. Banday et al.
1 3
Fig. 18 SINR map using 8 × 8 microstrip patch Array
Fig. 19 Directivity pattern of microstrip and cosine element
19. 1663
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
4 Discussion
For cellular networks in urban environments, spectrum is the scarcest resource. Hence
we have analyzed the performance of different antennas where less bandwidth is used
to achieve higher SINR values. In this paper we have extended our work to 3-D. This
allows each UE in a dense urban environment to cover a maximum number of BSs in
all directions. The directional gain achieved by using UPA is much higher than single
antenna element. The SINR map obtained using CI-propagation model shows less inter-
ference as compared to free space propagation model. Furthermore, real antenna model
using microstrip patch antenna replaces the equation based antenna element. This pro-
vides the required directional gain of 9 dBi although with lower values of front-to-back
ratios. To overcome the shortcomings of patch antenna we have used cosine antenna
element. This antenna has no back lobes since no energy is radiated in backward direc-
tion. Using cosine antenna as an array element we achieve higher values of SINR as
shown in Fig. 17.
5 Conclusion
The performance of 5G network operating at 30 GHz in dense urban scenario is evalu-
ated. Different interferers are modeled using IBM. NIT Srinagar, India is used as cen-
tre for location of transmitter site and further evaluated for downlink performance in
5G networks with frequency reuse factor 3. To view the map of a particular area, real
geospatial information is used. This allows viewing a location on Environmental Sys-
tem Research Institute (ESRI) map. SINR maps are plotted using single antenna ele-
ments as well as antenna arrays. High directive gains are achieved using arrays. Sig-
nal power in transmitters using microstrip patch array is 4 dB less than signal power
obtained using antenna element in a UPA. This is due to the higher radiation from back-
lobe of microstrip patch antenna element. To overcome this cosine antenna element is
used since it has no backlobe. This results in suppression of interference and increase
of signal strength. Comparing the values of SINR using cosine and patch element it can
be concluded that cosine element provides better directional gain, coverage and hence
increased SINR at BSs.
References
1. Akdeniz, M., Liu, Y., Samimi, M., Sun, S., Rangan, S., Rappaport, T., et al. (2014). _Millimeter wave
channel modeling and cellular capacity evaluation,_ IEEE. IEEE Journal on Selected Areas in Com-
munications, 32(6), 1164–1179.
2. Shokri-Ghadikolaei, H., Fischione, C., Fodor, G., Popovski, P., Zorzi, M. (2015). Millimeter
wave cellular networks: A MAC layer perspective. IEEE Transactions on Communications, 63(10),
3437–3458.
3. Di Renzo, M. (2015). _Stochastic geometry modeling and analysis of multi-tiermillimeter wave cel-
lular networks. IEEE Transactions on Wireless Communications, 14(9), 5038–5057.
4. Rappaport, T. S., Heath, R., Daniels, R. C., Murdock, J. N. (2014) Millimeter Wave Wireless Com-
munications. Pearson Education.
5. Jiang, X., Shokri-Ghadikolaei, H., Fischione, C., Pang, Z. (2016) A simplified interference model
for outdoor millimeter wave networks. In International wireless internet conference (pp. 101–108).
Springer, Cham.
20. 1664
Y. Banday et al.
1 3
6. Akdeniz, M. R., Rangan, S. (2013). Optimal wireless scheduling with interference cancellation.
In Proceedings of IEEE international symposim infomation theory, Istanbul, Turkey, July 2013 (pp.
246–50).
7. Draft new Report ITU-R M.[IMT-2020.EVAL]—Guidelines for evaluation of radio interface technolo-
gies for IMT-2020.
8. Abdrashitov, V., Nam, W., Bai, D. (2014). Rate and UESelection algorithms for interference-aware
receivers. Proc. IEEE VTC 2014-Spring, May 2014, Seoul, Korea.
9. Nam, W., Bai, D., Lee, J., Kang, I. (2014). Advanced interference management for 5G cellular net-
works. IEEE Communications Magazine, 52(5), 52–60.
10. Report ITU-R M.2135-1, “Guidelines for evaluation of radio interface technologies for IMT-
Advanced”, 2009. https://www.itu.int/dms_pub/itu-r/opb/rep/R-REP-M.2135-1-2009-PDF-E.pdf.
11. Le,L. B., Modiano, E., Joo, C., Shroff, N. B. (2010). Longest-queue-first scheduling under SINR
interference model. In Proceedings on ACM international symposium on mobile Ad hoc Networking
and Computing (MobiHoc) (pp. 41–50).
12. Weber, S. P., Andrews, J. G., Yang, X., De Veciana, G. (2007). Transmission capacity of wireless
ad hoc networks with successive interference cancellation. IEEE Transactions on Information Theory,
53(8), 2799–2814.
13. Shokri-Ghadikolaei, H., Fischione, C., Modiano, E. (2016). Interference models similarity index.
KTH Royal Institute of Technology, Tech.Rep., available upon request.
14. Chiang, M., Hande, P., Lan, T., Tan, C. W. (2008). Power control in wireless cellular networks.
Foundations and Trends® in Networking, 2(4), 381–533.
15. Freeman, R. L. (2015). Telecommunication system engineering (Vol. 82). Hoboken: Wiley.
16. Sun, S., et al. (2016). Investigation of prediction accuracy, sensitivity, and parameter stability of large-
scale propagation path loss models for 5G wireless communications. IEEE Transactions on Vehicular
Technology, 65(5), 2843–2860.
17. Sun, S. et al. (2016). Propagation path loss models for 5G urban micro- and macro-cellular scenarios.
In Proceedings on IEEE 83rd VTC Spring, Nanjing, China, May 2016. [Online]. Available: http://arxiv
.org/abs/1511.07311
.
18. Sun,S., Mac Cartney, Jr. G. R., Rappaport, T. S. (2016). Millimeter-wave distance-dependent
large-scale propagation measurements and path loss models for outdoor and indoor 5G systems. In
Proceedings on 10th EuCAP, Davos, Switzerland, Apr. 2016. [Online]. Available: http://arxiv.org/
abs/1511.07345
.
19. Rappaport, T. S., MacCartney, G. R., Jr., Samimi, M. K., Sun, S. (2015). Wide band millimeter-
wave propagation measurements and channel models for future wireless communication system design
(Invited Paper). IEEE Transactions on Communications, 63(9), 3029–3056.
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Yusra Banday has received B.E. in Electronics Communication
from University of Kashmir, Kashmir, India in 2011. She has done
M.Tech. in Communication Engineering from Galgotias University
Greater Noida, India in 2014. She has served Department of Electron-
ics and Communication Engineering in Islamic University of Science
and Technology, Kashmir, India as Assistant Professor for period of
one year. She has also worked for one year as contractual faculty in
National Institute of Technology, Srinagar, India. Currently she is pur-
suing Ph.D. in National Institute of Technology Srinagar. She is work-
ing as Senior Research Fellow in Advance Communication Lab at NIT
Srinagar. Her areas of interest are Millimeter Wave communication,
High speed Networks, Green Communication and Next generation
Networks. Her research aims at understanding the interference phe-
nomena in 5G networks and its management. She is student member of
IEEE and IEEE MTTS society.
21. 1665
SINR Analysis and Interference Management of Macrocell Cellular…
1 3
Professor Ghulam Mohammad Rather, Ph.D. has done Ph.D. from
IISC, Bangalore India in 1997. He is working as Professor in the
Department Of Electronics and Communication NIT Srinagar. He
received the B.E. degree in Electronics and Communications Engi-
neering from the Kashmir University, India, in 1981, the M.S. degree
(1988) in Computer Communications. He has a teaching experience of
more than 35 years. His research interests are in the field of Computer
Networks, Millimeter Wave Communication and Internet of Things.
He is a member of IET and IEEE MTTS Society India.
Dr. Gh. Rasool Begh has done Ph.D. from National Institute of Tech-
nology Srinagar (J K), India. He is working as Associate Professor in
the Department of Electronics and Communication Engineering at
NIT Srinagar. He has a teaching experience of more than 20 years. He
has guided a number of M.Tech. thesis related to OFDM, Cognitive
Radios, WLANs and Security. His areas of interest include Cognitive
Radios, OFDM, MIMO, Cooperative Communications, Error control
coding and Security. He is a member of IEEE MTTS Society.
![1646
Y. Banday et al.
1 3
the probability of an outage or successful decoding of a transmission, SINR metric is used.
Evaluating the performance of a network using SINR is difficult as it requires the esti-
mation of transmission strategies like; transmission power, Media Access Control (MAC)
protocol and antenna patterns. In case of networks using mmWave backhaul the quanti-
fying SINR becomes more difficult and complex because of high penetration loss, first
order reflection and use of MIMO antenna pattern which introduces elements of random-
ness [1–3]. Furthermore the different components of SINR viz.; Network Topology, Wire-
less Channel and Transmission strategy need to be mathematically modeled for estimating
interference in a link.
In 5G network architecture, implementation of heterogeneous networks create multi-
ple network tiers of different sizes with different transmission powers and large number
of smart and heterogeneous devices connected to each other. Transmissions from these
devices cause interference that needs to be suppressed. In these networks with increase in
deployment of cells RF noise floor increases and hence Co-Channel interference becomes
a major obstacle in improvement of network throughput. In recent research it has been
mentioned that combined effect of interference management on User Equipment (UE) side
and its network-side counterpart has greater effect than sum of their separate effects [4, 5].
Moreover in Release 12, a study has been started to support UE-side interference manage-
ment. Keeping this in consideration it is believed that in 5G networks UE-side interference
management should be given equal weightage as the network-side interference manage-
ment [6]. This paper is focussed on interference management in 5G Networks on network-
side under dense urban network environment.
The rest of paper is organized in four sections as follows. In Sect. 2 we describe the
system model, network topology and brief overview of interference models. Appropriate
interference model with respect to network topology is chosen for estimation of link outage
events under different network scenarios. Configuration and cell parameters according to
report ITU-R.M [7] are used to create dense urban eMBB transmitter sites. In this work we
have chosen NIT Entrance Road as central location of cellular network. This is followed by
mapping of SINR using UPA in Sect. 3. The paper is concluded in Sect. 4.
2 System Model
The Fig. 1 gives a network scenario with a pair of transmitter and its intended receiver
link denoted by i. The transmitters Tx1 and Tx2 act as potential interferers to transmis-
sions between (Txi)0 and its intended receiver (Rxi)0. All the set of interferer transmitters
excluding the intended transmitter are denoted by 𝔗 . The distance between the transmitter
(Txi)0 and the reference receiver (Rxi)0 is denoted by
di.
The SINR at receiver i is given by the equation
S = signal power, ik = interference from the interferer k
Both signal and interference power from interferer k are random variables with their
distributions depending on the various factors like transmit power, interference manage-
ment, antenna pattern and beam-forming techniques used. The above equation can be fur-
ther written as:
(1)
𝛾i =
S
∑
k∈𝔗
ik + η](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)