Content uploaded by Sharifah Hafizah Syed Ariffin
Author content
All content in this area was uploaded by Sharifah Hafizah Syed Ariffin on Nov 10, 2019
Content may be subject to copyright.
Analysis of Handover Performance in LTE Femtocells
Network
Nurul ‘Ain Amirrudin
1
•Sharifah Hafizah Syed Ariffin
1
•
Nik Noordini Nik Abd. Malik
1
•Nurzal Effiyana Ghazali
1
Published online: 9 October 2017
Springer Science+Business Media New York 2017
Abstract Handover management is one of the main factors representing the effectiveness
of every wireless network technology. Due to the special characteristics of a femtocell,
unnecessary handover occurs more frequently. This issue has attracted interest in devel-
oping a new handover algorithm in femtocell network. The standard handover algorithm
relies on Reference Signal Received Power or Reference Signal Received Quality (RSRQ)
level. However, this technique causes an unnecessary handover and reduces the user
throughput. Mobility prediction is one of a popular technique to be implemented in han-
dover algorithm. This paper analyzes the handover performance in femtocell network by
using two types of handover algorithm which are standard A2-A4-RSRQ handover algo-
rithm and proposed prediction handover algorithm. The analysis is performed in terms of
the number of handover, the number of unnecessary handover, and the user throughput.
The root cause of user throughput degradation is also analyzed. The results show that the
prediction handover algorithm provides better performance than the A2-A4-RSRQ han-
dover algorithm in terms of the number of handover and user throughput.
Keywords Femtocell Long Term Evolution (LTE) Handover RSRQ Prediction
&Nurul ‘Ain Amirrudin
ain@mahsa.edu.my
Sharifah Hafizah Syed Ariffin
sharifah@fke.utm.my
Nik Noordini Nik Abd. Malik
noordini@fke.utm.my
Nurzal Effiyana Ghazali
effiyana@fke.utm.my
1
Faculty of Engineering, MAHSA University, Selangor, Malaysia
123
Wireless Pers Commun (2017) 97:1929–1946
DOI 10.1007/s11277-017-4222-3
1 Introduction
Long Term Evolution (LTE) has been introduced by the Third Generation Partnership
Project (3GPP) due to high demand for higher data rate and quality of service (QoS). LTE
is a latest standard in the mobile network technology tree that previously realized the GSM/
EDGE and UMTS/HSxPA network technologies [1]. The objective of LTE development is
to build a system that meets demand for a high data rate, low latency, and optimization for
packet switched traffic [2]. The LTE is expected to provide peak data rates of at least
100 Mbps in downlink and 50 Mbps in uplink. End-user latency is reduced to 10 ms in
round trip times while control plane latency (i.e. transition time from idle state to active
state) is less than 100 ms [3,4]. A major difference of LTE to other technologies is the
radio interface where in LTE, Orthogonal Frequency Division Multiplexing (OFDM) and
Single Carrier Frequency Domain Multiple Access (SC-FDMA) are used as radio access
schemes for downlink and uplink respectively [5].
In a wireless network, voice calls and data traffic are dominated by broadband services
indoor [6,7]. Therefore, it is extremely important for mobile operators to provide a better
coverage in an indoor environment, not only for voice, but also for video and high speed
data services, which are becoming increasingly important. Hence, femtocells have been
introduced to enhance indoor coverage, deliver high bandwidths and new services to end-
customers, as well as can off-load traffic from the existing macro-cellular networks.
Moreover, from the mobile operator point of view, when the indoor coverage is good, this
may indirectly increase Average Revenue per User (ARPU) and enhance customer loyalty
[8].
By definition, a femtocell is a low-power wireless access points with small cell cov-
erage. It operates on a spectrum licensed to connect standard mobile devices to a mobile
operator’s network using residential digital subscriber line (DSL) or cable broadband
connections. Besides than offering broadband services indoor (i.e. homes and offices),
femtocells also have been developed to provide a service at outdoor scenarios with a very
limited geographical coverage [9]. Since the femtocell mainly focuses on the indoor
environment, it is known as a home based station. In 3GPP LTE, the femtocell is more
commonly known as home evolved Node B (HeNB). Femtocells require low power,
ranging between 13 and 20 dBm with coverage from 15 to 50 m [10].
In terms of mobility management in femtocell deployment, there are several chal-
lenges. A femtocell is a small base station that deployed in an unplanned manner
because it is installed by the end-user. Therefore, there is a possibility that a large
number of femtocell may install in single macrocell. This scenario will create a large
number of neighbour cell list and interference problems [11]. When the neighbour cell
list is high, the network takes a long time to scan the best femtocell to attach to the user.
Moreover, the femtocell has small coverage, which is around 50 m. This characteristic
may cause the user pass-by the femtocell within a short time, and then attach to other
femtocell or previous femtocell. This unnecessary handover and ping-pong handover will
reduce system capacity and reduce user throughput. Therefore, it is necessary to design
an appropriate handover decision, in order to have a seamless handover and high user
throughput.
This paper discusses handover in femtocell network and analyzes handover perfor-
mance in terms of number of handovers, number of unnecessary handovers, and the user
throughput. Two types of handover algorithm are use, namely the standard A2-A4-RSRQ
handover algorithm and proposed prediction handover algorithm. The rest of the paper is
1930 N. ‘A. Amirrudin et al.
123
organized as follows: In Sect. 2, the system architecture of femtocell network, the access
control and the handover scenario in femtocell are discussed. Then, Sect. 3discusses
related work of handover decision on femtocell network. The handover call flow in
femtocell network describes in Sect. 4. In Sect. 5, two types of handover algorithms are
presented. The handover performance in femtocell network using both handover algo-
rithms are analyzed in Sect. 6, before concluding the analysis in Sect. 7.
2 LTE Femtocell Network
2.1 LTE Femtocell System Architecture
LTE has been defined as an IP-based and flat core network architecture. The architecture is
a part of System Architecture Evolution (SAE) that consists of Evolved Packet Core (EPC)
and Evolved Universal Terrestrial Radio Access Network (E-UTRAN). The EPC contains
of Mobility Management Entity (MME), Serving Gateway (S-GW), and Packet Data
Network Gateway (PDN GW). The MME is in charge of all the control plane (C-plane)
functions related to subscriber and session management. The SGW serves as a mobility
anchor for user plane (U-plane) during handovers, and also responsible to route and for-
ward user data packets. The PDN GW provides access from the User Equipment (UE) to
external packet data networks and allocates IP addresses for the UE and QoS enforcement.
The E-UTRAN consists of evolved Node B (eNB) and Home eNB (HeNB), referred to as
macro base station and femto base station in the LTE, respectively. The architecture may
deploy a HeNB Gateway (HeNB GW) in order to manage a large number of HeNBs in a
scalable manner.
The eNBs and HeNBs are interconnected with each other by means of X2 interface.
The S1 interface is a reference point between the E-UTRAN and the EPC. More
specifically, the E-UTRAN is interconnected to the MME with S1-MME interface, and
interconnected to the SGW with S1-U interface. The HeNB GW serves as a concentrator
for the C-plane, and is therefore connected with HeNBs and MME through an S1-MME
interface. The S5 interface is a reference point between HeNBs and SGW. The S5
interface is available if the HeNBs support the (Local IP Access) LIPA function. The
overall E-UTRAN architecture with deployed HeNB GW is shown in Fig. 1. The HeNB
shall acts as the eNB in terms of function supported and the procedures run between the
HeNB and the EPC [12].
2.2 Access Control of Femtocell
In LTE, the UE is designed to be aware of the femtocell, where the UE should be able to
recognize a cell, whether it is a macrocell or a femtocell. If the UE is not interested in the
femtocell, the searching of femtocell is avoided. Otherwise, if the UE is interested, it
quickly completes the searching procedure. There are three types of access control models
in femtocell, which are open, closed, and hybrid access mode.
For the open access mode, the femtocell acts as the macrocell where all users can access
to the femtocell. For example, an operator deploys this femtocell to provide a good cov-
erage in an area where there is a coverage hole. For the closed access mode, the femtocell
can be accessed by the users under closed subscriber group (CSG) only. For the hybrid
access mode, the femtocell is similar to the closed access mode, but also open to the non
Analysis of Handover Performance in LTE Femtocells Network 1931
123
CSG if the bandwidth is available. To differentiate the femtocell access modes, CSG
identity (ID) and CSG indicator have been introduced. CSG ID is a unique numeric
identifier that broadcasts in system information (SI) by the CSG and the hybrid cell. CSG
indicators are presented with a value of TRUE for the CSG cell and absent for the hybrid
and the open cell [13–15].
2.3 Handover Scenario in LTE Femtocell
The latest E-UTRAN architecture as per discussed in [12] defines a direct interface
between eNB and HeNB. Therefore, it is possible to execute an X2-handover between eNB
and HeNB. If there is no X2 interface between them, the S1-handover must be performed.
It may have three handover scenarios in femtocell network:
1. Inbound handover; this represents the handover of the UE from the eNB to the HeNB.
The handover scenario is the most challenges since there are large numbers of possible
targets HeNBs and also the restriction of access control of the HeNB.
2. Outbound handover; this represents the handover of the UE from the HeNB to the
eNB. The handover scenario is not as complicated as inbound handover, since there is
only one candidate of eNB and it is open to all users.
3. Inter-HeNB handover; represents the handover between the HeNBs. The handover
procedure is similar to inbound handover because in this scenario, there are large
numbers of possible target HeNBs and also needs to consider on the access control of
the HeNB.
Fig. 1 Overall E-UTRAN architecture with deployed HeNB GW
1932 N. ‘A. Amirrudin et al.
123
3 Related Work
Mobility management is a set of tasks for controlling mobile station in a wireless networks
to maintain their connections while moving [16]. In LTE system, fast and seamless han-
dover is one of the main goals. Due to its orthogonal frequency division multiple access
(OFDMA) technology, only hard handover is available [17]. The hard handover also called
as a ‘‘break before make’’ connection. Unlike soft handover, in the hard handover, the link
to the current base station (BS) is terminated before the new link is connected to the new
BS [18]. In other words, only one link is available at one time. This scenario may create an
interruption time in the user plane, and reduce handover performance in terms of success
rate and handover delay [5].
Several mobility techniques have been proposed to enhance handover performance,
especially to reduce unnecessary handovers. By reducing the unnecessary handover, the
ping-pong effect may decrease. The ping-pong effect occurs when a call is handed over to
a new cell and handed back to the previous cell in less than a critical time [19]. This ping-
pong effect will reduce user throughput and system capacity as well. The most usual
technique to mitigate the number of handover and unnecessary handover is by optimizing
the handover parameter which is handover margin (HM) and time-to-trigger (TTT). In
[19], the authors proposed a handover optimization algorithm by tuning the handover
parameter (i.e. hysteresis and TTT) based on handover performance indicator (handover
failure ratio, ping-pong handover, and call dropping ratio). The results show that the
handover failure ratio and ping-pong handover ratio are driven to zero after 500 s simu-
lation time when the optimization is adopted. [20] proposed a cell-type adaptive handover
margin. The algorithm assigns different HM based on target cells and user’s speed. The
results show that the ping-pong rate and radio link failure (RLF) rate are decreased when
compare to the constant HM value.
Since the coverage of the femtocell is small, user with a high velocity will pass-by the
femtocell in a short time. Therefore, by considering the user’s quality of service (QoS), it is
unnecessary for high speed user to execute handover, especially for non-real-time service.
Several researchers have proposed a handover algorithm by considering the user velocity.
In [11], the authors have proposed a new handover algorithm based on the user’s speed and
user’s application (i.e. real time service and non-real time service). The UE’s speed is
divided into three categories which are low, medium, and high speed. The handover
algorithm decides to not execute handover for high speed user and medium speed with
non-real-time service, and execute handover for other criteria. Besides user velocity, [10]
considered another parameter, namely interference. The interference level is considered for
non-CSG user that need to handover to hybrid femtocell. In order to reduce the interfer-
ence, the handover is needed for low speed user. Other than that, [21] have added several
parameters for the handover decision procedure. The mobility prediction of the user is
considered in handover decision procedure. Based on user current position and user
velocity, it can estimate where the user is moving, thus the next BS that user need to
handover can be predicted. Moreover, proactive and reactive handover is proposed, where
proactive handover focuses to minimize packet loss while reactive handover is aim to
reduce the number of handover. All proposed handover algorithms are aimed to improve
the handover performance by reducing the number of handover only without considering
on the user throughput. However, from user’s point of view, it does not matter how many
handover is occurred, as long as they get the higher throughput and do not lose a
connection.
Analysis of Handover Performance in LTE Femtocells Network 1933
123
During handover executions in which the UE receives a handover command from the
source base station (i.e. eNB or HeNB), the UE cannot sends and receives any packet until
a new connection is established. Therefore, the handover latency needs to be reduced as
much as possible in order to achieve a seamless handover. One of the main factors of
handover latency is resource allocation that takes a lot of time. One of the most effective
approaches to reduce delay in resource allocation is to predict the next location of the user
[22]. The function of mobility prediction is to detect the identity of the future cell so that a
resource reservation can be performed prior to the actual handover. This technique has
attracted attention from researchers to enhance the handover performance. In [23], the
authors proposed a prediction based on user’s mobility history. The proposed algorithm
needs the network to recognize the user and record the movement information. Moreover,
the algorithm requires considering the signal strength of the user and then setting as a
candidate handover once the signal strength is higher than a certain threshold. The results
show that the more regularity of user movement, the better performance gain has been
gotten in terms of number of handover and ping-pong rate. During the normal handover
procedure, the handover decision is performed based on measurement reports sent by the
UE. The measurement report is sent frequently so that the network can aware the channel
status of the UE. However, this situation may decrease the control channel capacity for
downlink. Thus, in [24], a handover decision without relying on measurement report was
proposed. Based on past channel state information, the source base station can predict the
channel quality to perform the handover decision. The proposed scheme has reduced the
outage probability, but does not consider the number of handovers.
4 Handover Call Flow
In an LTE network, there are two types of handover, namely X2-based handover and S1-
based handover. By default, the X2-based handover is implemented unless there is no X2
interface between source eNB/HeNB and target eNB/HeNB or if there is a setting to use
S1-based handover. The difference between those handover types are the involving of the
MME. For an X2-based handover, the message from source eNB/HeNB is directly sent to
the target eNB/HeNB. However, for S1-based handover, the MME is required as a medium
to send the message. In this paper, the X2-based handover is used to analyze the handover
performance in the LTE femtocell network.
Handover can be divided into three phases; preparation (initiation), execution and
completion [25]. The message sequence diagram for X2-based handover is depicted in
Fig. 2. The first phase is handover preparation where involve the source HeNB, the target
HeNB and the UE. The main messages for this phase are described as follows:
•The source HeNB sends the measurement control to configure the UE measurement
procedure.
•The UE sends the measurement report after it meets the measurement report criteria
that set in measurement control.
•The source HeNB makes a handover decision based on the measurement report.
•The target HeNB performs admission control and checks for resource availability, then
reserves it.
•The source HeNB sends the handover command to the UE.
For handover execution, the procedures are described as follows:
1934 N. ‘A. Amirrudin et al.
123
UE Source HeNB Target HeNB SGW/PGWMME
Measurement Control
Measurement Reports
Handover
Decision
Handover Request
Admission
Control
Handover Request Ack
Handover Command
SN Status Transfer
Detach from old
cell, synchronize
to new cell
Handover Confirm
Path Switch Request
Modify Bearer Request
Packet Data Packet Data
Deliver buffered
and transit packets
to target HeNB
Buffer packets
from source
HeNB
Data Forwarding
Synchronisation
Packet Data
Switch DL
path
End Marker
End Marker Packet Data
Modify Bearer
Response
Path Switch Request
Ack
UE Context Release
Release
Resources
Packet Data
Handover Preparation
Handover Execution
Handover Completion
Fig. 2 X2-based handover call flow between HeNBs
Analysis of Handover Performance in LTE Femtocells Network 1935
123
•The UE detach from the source HeNB and synchronize to the target HeNB.
•The source HeNB sends the Sequence Number (SN) Status Transfer to the target HeNB
to convey the Packet Data Convergence Protocol (PDCP) and Hyper Frame Number
(HFN) status of the E-UTRAN Radio Access Bearers (E-RABs). At this stage, the
source HeNB freezes its transmitter/receiver status, and no data can be sent or received
[26].
In the last phase, which is handover completion, the processes are described as follows:
•Once the UE has synchronized with the target HeNB, it sends the handover confirm.
•The target HeNB sends a path switch request to inform that the UE has changed cell.
•The SGW switches the path of downlink data to the target HeNB.
•The source HeNB releases radio and control of related resources once it receives the
UE context release message.
From the handover call flow, the handover latency can be measured, which is duration
between the UE sends the measurement report and when the UE sends the handover
confirm message. It is involving the handover decision and admission control. The more it
takes on handover decision, the higher the handover latency is. Moreover, if there is no
bandwidth available in the target HeNB, the handover process may take longer time to
complete. Therefore, it is important to have a simple handover decision so that it takes less
time to complete.
5 Handover Algorithm
Handover algorithm or handover decision is a most challenging part in the handover call
flow. The decision is made based on measurement reports provided by the UE. The
common system metrics include in the measurement report for handover decision are
Signal-to-Interference-and-Noise-Ratio (SINR), Received Signal Strength Indicator
(RSSI), Reference Signal Received Power (RSRP), and Reference Signal Received Quality
(RSRQ). These metrics are used to select the possible handover candidate. Besides, the
control parameters are tuned by the handover algorithm to increase the handover perfor-
mance of the network [19]. Normally, the parameters are hysteresis and time-to-trigger
(TTT). Other than that, user’s speed and user’s application (i.e. real-time application and
non-real-time application) may also be considered. The handover is triggered if the con-
dition set in the control parameter is fulfilled.
As stated, the measurement reports are sent by the UE if the measurement report criteria
are met. The measurement report criteria can be either event triggered or periodic. Periodic
reporting is typically used for measuring an automatic neighbour cell search. The event
triggered measurements based on E-UTRA measurements are listed in Table 1[4,27].
5.1 A2-A4-RSRQ Handover Algorithm
In an LTE network, the UE has to report two parameters on reference signal which are
RSRP and RSRQ every 200 ms [28]. RSRP is the absolute signal strength of the LTE
reference signal related in dBm while RSRQ is the DL signal-to-interference ratio in dB
measured on the LTE reference signals [29]. Both RSRP and RSRQ are used to determine
the best cell for the user. However, the RSRQ provides additional information to determine
interference level at the location. Therefore, RSRQ is more appropriate than RSRP for the
1936 N. ‘A. Amirrudin et al.
123
femtocells network, which consists of large number of femtocells where the interference
level is high.
Figure 3shows a simple A2-A4-RSRQ handover algorithm. The algorithm has been
discussed and implemented in Network Simulator 3.18 [30] based on standardization of
3GPP [28]. When the measurement report criteria are met, the UE sends the measurement
report consist of RSRP and RSRQ value of the source HeNB and all neighbour cells. The
source HeNB checks the RSRQ level either lower than the threshold level or not. If the
RSRQ level is lower than the threshold, look the neighbour cell with highest RSRQ level
and the difference is determined. The handover is triggered if the difference is higher or
equal to the neighbour cell offset. The value of the source threshold and the neighbour cell
offset is set by the network. The neighbour cell offset is act as a handover margin where it
can reduce the number of handover. The higher the neighbour cell offset value, the later the
handover is triggered which may increase the handover delay.
From the algorithm, it shows that there are two parameters to control the handover
decision, which are source cell threshold and neighbour cell offset. These parameters are
set by the network to control the handover decision. The source cell threshold is a mini-
mum value of the RSRQ before handover is triggered. For the second parameter which is
neighbour cell offset, the RSRQ value of the neighbour HeNB must be higher than the
source HeNB. If the criterion is not met, no handover is imposed. In other words, there is
no suitable HeNB to serve the user with acceptable RSRQ value. Thus, the user will
experience a bad QoS or even may lose the connection.
5.2 Prediction Handover Algorithm
A handover algorithm which relies on LTE measurements such as RSRP and RSRQ may
cause a ping-pong effect. This is because the RSRP and RSRQ values are fluctuated cause
by interference or other issues. The hysteresis and TTT are introduced to reduce the ping-
pong effect, however, it may delay handover for a while and cause a user throughput
degradation. Therefore, it is necessary to handover at correct time. Prediction in handover
algorithm is made to predict the next location of the user and predict the best time to trigger
handover to the new base station. By predicting the next location, the resource reservation
can made earlier and may reduce the handover latency. When predicting the best time for
handover trigger, the handover may trigger at the optimal time to mitigate the handover
delay. In a femtocell network where the coverage area is small, handover delay should be
reduced as much as possible. This is because when the handover is late to trigger, this will
result in packet loss. Moreover, delaying too long may make it impossible for the user to
meet its QoS objectives.
Table 1 Event triggered reports for E-UTRA [4,27]
Event triggered Criteria
Event A1 Source cell becomes better than an absolute threshold
Event A2 Source cell becomes worse than an absolute threshold
Event A3 Neighbour cell becomes an amount of offset better than source cell
Event A4 Neighbour cell becomes better than an absolute threshold
Event A5 Source cell becomes worse than an absolute threshold 1 and neighbour
cell becomes better than an another absolute threshold 2
Analysis of Handover Performance in LTE Femtocells Network 1937
123
Figure 4shows the optimal point to trigger handover from source HeNB to target
HeNB. As the user moves towards the target HeNB with a certain velocity, the handover
shall occur once the reported RSRQ value of the target HeNB is greater than the source
HeNB. Theoretically, the handover shall be performed at the optimal handover point as
illustrated in Fig. 4. Later or earlier from this point may affect the handover performance in
terms of user throughput or may lose a connection. One of the ways to achieve an optimal
handover point is by mobility prediction.
Start
Source HeNB receives
measurement reports from
UE (Event A2 and A4)
Source HeNB RSRQ
<= Source Cell
Threshold?
Look for neighbour cell
with the highest RSRQ
(best neighbour RSRQ -
source HeNB RSRQ)
>= Neighbour Cell Offset?
No
No
Yes
Yes
Trigger handover procedure
for this UE to the best
neighbour
End
Fig. 3 A2-A4-RSRQ handover algorithm
1938 N. ‘A. Amirrudin et al.
123
In this paper, mobility prediction based on the user’s mobility history as discussed in
[31] is analyzed. The prediction technique is intended to predict the best target HeNB,
while the best time to trigger handover is relies on the User velocity and the user’s
direction. It assumes that the user sends their location to the network constantly. Thus, the
network can determine the user’s direction and the user velocity. From user velocity, the
optimal handover point can be determined. This paper analyzes the effect of the handover
performance when the handover is triggered at optimal handover point based on prediction
and compare it with the A2-A4-RSRQ handover algorithm.
6 Performance Evaluation
6.1 Scenario
In this paper, a scenario of 22 HeNBs is used as illustrated in Fig. 5. Distance between the
HeNBs is 100 m, and the coverage area is overlaps with each other to avoid any inter-
ruption. The initial state of the user is set between the HeNB1 and HeNB12, and it is
attached to the HeNB1 at start point. The user is moving from an initial state towards the
HeNB11 and HeNB22 with constant velocity. User velocity is varied from 1 to 10 m/s. All
HeNBs are using an open access mode, thus the user can attach to all HeNBs without any
restriction. Both handover algorithm types, namely the A2-A4-RSRQ handover algorithm
and prediction handover algorithm, are used to evaluate the handover performance. Details
of simulation parameters are listed in Table 2.
Fig. 4 Optimal handover point
Analysis of Handover Performance in LTE Femtocells Network 1939
123
6.2 Handover Performance Indicators
The handover performance indicator (HPI) is measured in terms of the number of han-
dovers, the number of unnecessary handovers, and the user throughput. The number of
handover is defined as a number of successful handover, while the number of unnecessary
handover is defined as a scenario where the handover is triggered to a new cell and within a
Fig. 5 The scenario of the evaluation
Table 2 Simulation parameters Metrics Value
Simulation time 110–1100 (s) [depends on user velocity]
Simulation area 1100 (m) 9200 (m)
Number of HeNBs 22
HeNB transmit power 20 (dBm)
Distance between HeNBs 100 (m)
Number of user 1
User velocity 1–10 (m/s)
Source cell threshold 30
Neighbour cell offset 1
Critical time 3 (s)
020 40 60 80 100 120 140 160 180 200 220
HeNB5
HeNB10
HeNB15
HeNB20
Simulation Time [s]
HeNB ID
A2-A4-RSRQ ho algorithm
Prediction ho algorithm
Fig. 6 Handover scenario for user’s speed 5 m/s
1940 N. ‘A. Amirrudin et al.
123
critical time the call is handover back to the previous cell or to another cell. For the last
HPI, user throughput is an important metric to be measured to analyze the handover
performance. Throughput is defined as a ratio of total number of packets received over the
total simulation time. Mathematically, it can be defined as follows:
Throughput ¼Total number of packets received
Total simulation time ð1Þ
6.3 Results
The experiments are evaluated by using network simulator 3 (NS3.18). Two types of
handover algorithms, A2-A4-RSRQ handover algorithm and prediction handover
1 2 3 4 5 6 7 8 9 10
0
10
20
30
40
50
UE velocity [m/s]
No. of handover
No. of handover A2-A4-RSRQ
No. of unnecessary handover A2-A4-RSRQ
No. of handover prediction
No. of unnecessary handover prediction
Fig. 7 No of handover and no of unnecessary handover for both handover algorithms
1 2 3 4 5 6 7 8 9 10
53
53.5
54
54.5
55
UE velocity [m/s]
User throughput [packet/s]
A2-A4-RSRQ ho algorithm
Prediction ho algorithm
Fig. 8 User throughput for both handover algorithm
Analysis of Handover Performance in LTE Femtocells Network 1941
123
algorithm are used to analyze the handover performance in femtocells network. Firstly, the
handover scenario is evaluated to show a scenario of unnecessary handover in the fem-
tocell network. Figure 6shows the handover scenario for both handover algorithms. The
graph shows that there is a handover occur from HeNB1 to HeNB12 at time 18.8 s for the
A2-A4-RSRQ handover algorithm. The call is handed back to the HeNB1 after 1.4 s the
user attached to the HeNB12. This scenario called as unnecessary handover. It also shows
that there is no unnecessary handover occurred if the prediction handover algorithm is
implemented. Since the RSRQ value is fluctuated, such unnecessary handovers may fre-
quently occur. One of the ways to mitigate the unnecessary handover is by increasing the
hysteresis. However, the handover is intended to delay for a while and may cause user
throughput degradation.
(a)
(b)
020 40 60 80 100 120 140 160 180 200 220
-11
-10.5
-10
-9.5
-9
-8.5
-8
-7.5
-7
Simulation Time [s]
RSRQ value [dB]
020 40 60 80 100 120 140 160 180 200 220
0
20
40
60
80
100
120
Simulation Time [s]
User throughput [packet/s]
Fig. 9 The effect of RSRQ level to user throughput. aThe RSRO value of source HeNB. bThe user
throughput
1942 N. ‘A. Amirrudin et al.
123
Then, the number of handovers and the number of unnecessary handovers between both
handover algorithm types have been analyzed. Figure 7shows the number of handover
when the User velocity is varied. The number of handovers is high for the A2-A4-RSRQ
algorithm due to unnecessary handovers. On average, the number of handover is 37 for the
A2-A4-RSRQ algorithm where most of it is denoted from the unnecessary handover. Since
no unnecessary handover occurred for the prediction algorithm, the number of handovers
was less at 10 handovers. Unnecessary handovers for the A2-A4-RSRQ algorithm are
fewer for low-velocity users, because the user spends more time in femtocell coverage than
a high-velocity user.
Third, the user throughput which is the most important HPI is analyzed. Figure 8shows
the user throughput for both handover algorithms when the user velocity is varied. The
graph shows that the user throughput is high for prediction algorithm with percentage of
1%. The difference of user throughput between the handover algorithms is denoted from
the unnecessary handover scenario. However, the overall user throughputs for both han-
dover algorithm types are quite low, as only 54% of all packets sent were received by the
user. The User velocity does not influence so much for the user throughput.
Lastly, the cause of low user throughput has been investigated. The sample of data with
user velocity of 5 m/s with A2-A4-RSRQ handover algorithm has been taken. A rela-
tionship between the user throughput and the RSRQ value is shown in Fig. 9. The graph
shows that user throughput is decreased when the RSRQ value is decreased. The drop of
RSRQ value has occurred at the overlap of the femtocell coverage area. For example, at
time 20 s, the coverage area of the femtocell overlaps among HeNB1, HeNB2, HeNB12,
and HeNB13. This kind of overlapping scenario causes a high interference and finally
decreases the user throughput. As known, RSRQ value can determine the existence of
interference. If the RSRP value remains stable or becomes even better while the RSRQ
value is declining, this is an unambiguous symptom of rising interference [29]. Therefore, a
technique must to be implemented to reduce the interference level in femtocell network,
which is beyond the scope of this paper.
7 Conclusion
The femtocell has been introduced in Long Term Evolution (LTE) technology to provide
better coverage, especially in indoor environments and outdoor scenarios with limited
geographical coverage. As in other wireless technologies, mobility management is the
most important part to be catered to show the effectiveness of the technology. This paper
has analyzed the handover performance in femtocell network by using two types of han-
dover algorithm which are standard A2-A4-RSRQ handover algorithm and proposed
prediction handover algorithm. The experiments were used analyzed the handover per-
formance in terms of the number of handover and the user throughput. The results show
that by predicting the best target cell and the best time for handover cause a better
performance compare if only rely on RSRQ value. The experiments also analyze the root
cause of user throughput degradation. Further work on mobility management in femtocell
network shall examine the interference level as well.
Acknowledgements The authors would like to thank all who contributed toward making this research
successful. The authors wish to express their gratitude to Ministry of Higher Education (MOHE), Research
Management Center (RMC) for the sponsorship, and Telematic Research Group (TRG), Universiti
Analysis of Handover Performance in LTE Femtocells Network 1943
123
Teknologi Malaysia for the financial support and advice for this project. (Vot Number
Q.J130000.2509.05H58 and PY/2013/01168).
References
1. Motorola. (2007). Long Term Evolution (LTE): A technical overview. Technical White Paper.
2. Singhal, D., Kunapareddy, M., Chetlapalli, V., James, V. B., & Akhtar, N. (2011). LTE-advanced:
Handover interruption time analysis for IMT-A evaluation. In 2011 International conference on signal
processing, communication, computing and networking technologies (ICSCCN) (pp. 81–85). IEEE.
3. Motorola. (2007). Long term evolution (LTE). Technical White Paper.
4. Vehanen, J. (2011). Handover between LTE and 3G Radio Access Technologies: Test measurement
challenges and field environment test planning.
5. Dimou, K., Wang, M., Yang, Y., Kazmi, M., Larmo, A., Pettersson, J., et al. (2009). Handover within
3GPP LTE: Design principles and performance. In 2009 IEEE 70th vehicular technology conference
fall (VTC 2009-Fall) (pp. 1–5). IEEE.
6. Yang, G., Wang, X., & Chen, X. (2011). Handover control for LTE femtocell networks. In 2011
International conference on electronics, communications and control (ICECC) (pp. 2670–2673). IEEE.
7. Chandrasekhar, V., Andrews, J. G., & Gatherer, A. (2008). Femtocell networks: A survey. Commu-
nications Magazine, IEEE, 46(9), 59–67.
8. Wang, L., Zhang, Y., & Wei, Z. (2009). Mobility management schemes at radio network layer for LTE
femtocells. In IEEE 69th vehicular technology conference, 2009. VTC Spring 2009 (pp. 1–5). IEEE.
9. Capozzi, F., Piro, G., Grieco, L. A., Boggia, G., & Camarda, P. (2012). On accurate simulations of LTE
femtocells using an open source simulator. EURASIP Journal on Wireless Communications and Net-
working, 2012(1), 1–13.
10. Wu, S. J. (2011). A new handover strategy between femtocell and macrocell for LTE-based network. In
2011 4th International conference on Ubi-media computing (U-Media) (pp. 203–208). IEEE.
11. Zhang, H., Wen, X., Wang, B., Zheng, W., & Sun, Y. (2010). A novel handover mechanism between
femtocell and macrocell for LTE based networks. In Second international conference on communica-
tion software and networks, 2010. ICCSN’10 (pp. 228–231). IEEE.
12. 3GPP. (2014). TS36.300 V12.0.0: Technical specification group Radio Access Network; Evolved
Universal Terrestrial Radio Access (E-UTRA) and Evolved Universal Terrestrial Radio Access Net-
work (E-UTRAN); Overall description; Stage 2 (Release 12). Technical Specification.
13. 3GPP. (2011). TR25.367 V10.0.0: Technical specification group Radio Access Network; Mobility
procedures for Home Node B (HNB); Overall Description; Stage 2 (Release 10). Technical
Specification.
14. Golaup, A., Mustapha, M., & Patanapongpibul, L. B. (2009). Femtocell access control strategy in
UMTS and LTE. Communications Magazine, IEEE, 47(9), 117–123.
15. Lin, P., Zhang, J., Chen, Y., & Zhang, Q. (2011). Macro-femto heterogeneous network deployment and
management: From business models to technical solutions. Wireless Communications, IEEE, 18(3),
64–70.
16. Qutqut, M., & Hassanein, H. (2012). Mobility management in wireless broadband femtocells. Technical
Report 2012-590, Queen’s University.
17. Kim, T. H., Yang, Q., Lee, J. H., Park, S. G., & Shin, Y. S. (2007). A mobility management technique
with simple handover prediction for 3G LTE systems. In 2007 IEEE 66th vehicular technology con-
ference, 2007. VTC-2007 Fall (pp. 259–263). IEEE.
18. Akhila, S., & Lakshminarayana, M. (2008). Averaging mechanisms to decision making for handover in
GSM. 32rd World Academy of Science, Engineering and Technology.
19. Jansen, T., Balan, I., Turk, J., Moerman, I., & Kurner, T. (2010). Handover parameter optimization in
LTE self-organizing networks. In 2010 IEEE 72nd vehicular technology conference fall (VTC
2010-Fall) (pp. 1–5). IEEE.
20. Bae, H. D., Ryu, B., & Park, N. H. (2011). Analysis of handover failures in LTE femtocell systems. In
Australasian telecommunication networks and applications conference (ATNAC), 2011 (pp. 1–5). IEEE.
21. Ulvan, A., Bestak, R., & Ulvan, M. (2010). Handover scenario and procedure in LTE-based femtocell
networks. In UBICOMM 2010, the fourth international conference on mobile ubiquitous computing,
systems, services and technologies (pp. 213–218).
22. Duong, T. V. T., & Tran, D. Q. (2012). An effective approach for mobility prediction in wireless
network based on temporal weighted mobility rule. International Journal of Computer Science and
Telecommunications, 3(2), 29–36.
1944 N. ‘A. Amirrudin et al.
123
23. Ge, H., Wen, X., Zheng, W., Lu, Z., & Wang, B. (2009). A history-based handover prediction for LTE
systems. In International symposium on computer network and multimedia technology, 2009. CNMT
2009 (pp. 1–4). IEEE.
24. Tu, H. M., Lin, J. S., Chang, T. S., & Feng, K. T. (2012). Prediction-based handover schemes for relay-
enhanced LTE-A systems. In 2012 IEEE wireless communications and networking conference (WCNC)
(pp. 2879–2884). IEEE.
25. Hussein, Y. S., Ali, B. M., Varahram, P., & Sali, A. (2011). Enhanced handover mechanism in long term
evolution (LTE) networks. Scientific Research and Essays, 6(24), 5138–5152.
26. Sesia, S., Toufik, I., & Baker, M. (2009). LTE—The UMTS Long Term Evolution. Chichester: Wiley.
27. Holma, H., & Toskala, A. (2009). LTE for UMTS-OFDMA and SC-FDMA based radio access.
Chichester: Wiley.
28. 3GPP. (2011). TS36.331 V10.4.0: Technical specification group Radio Access Network; Evolved
Universal Terrestrial Radio Access (E-UTRA); Radio Resource Control (RRC); Protocol Specification
(Release 10). Technical Specification.
29. Kreher, R., & Gaenger, K. (2010). LTE signaling: Troubleshooting and optimization. Chichester: Wiley.
30. NS-3 project. (2013). ns-3 Model library (Release ns-3.18). NS-3 Network Simulator.
31. Amirrudin, N. A., Ariffin, S. H., Malik, N. N. N., & Ghazali, N. E. (2013). User’s mobility history-based
mobility prediction in LTE femtocells network. In 2013 IEEE international on RF and microwave
conference (RFM) (pp. 105–110). IEEE.
Nurul ‘Ain Amirrudin received her Ph.D degree in Electrical Engi-
neering (2016), and her B.E. in Electrical Engineering - Telecommu-
nication (2008) from Universiti Teknologi Malaysia, Malaysia. She
joined Telekom Malaysia Berhad as an assistant manager of Product
Development and Management starting July 2008 until end of 2011
before pursuing her Ph.D. She is currently a lecturer of Faculty of
Engineering in MAHSA University, Malaysia. Her research interests
include mobility management, mobility prediction and their applica-
tions in Long Term Evolution.
Sharifah Hafizah Syed Ariffin received her Ph.D. degree in 2006
from Queen Mary University of London, London, received her Master
degree in Mobility Management in Wireless Telecommunication
(2001) from Universiti Teknologi Malaysia and her B.E. (Hons) in
Electronic and Communication Engineering from University of North
London, London, England in 1997. She is currently an associate
proffessor in the Faculty of Electrical Engineering, Universiti Tekno-
logi Malaysia, Malaysia. Her research interests include Wireless
Sensor Network, IPv6 network and mobile computing system, handoff
management in WiMax, low rate transmission protocol using IPv6-
6loWPAN, network modelling and performance, priority scheduling in
packet network.
Analysis of Handover Performance in LTE Femtocells Network 1945
123
Nik Noordini Nik Abd. Malik graduated with B.E. (Electrical-
Telecommunication); M. E. (Radio Frequency and Microwave Com-
munications) and Ph.D. (Electrical Engineering) from Universiti
Teknologi Malaysia (UTM), Malaysia; University of Queensland,
Australia and Universiti Teknologi Malaysia (UTM), Malaysia in
2003, 2005 and 2013, respectively. She was a research and develop-
ment electrical engineer in Motorola Solutions Penang, Malaysia in
2004. Then, she has been a lecturer with the Faculty of Electrical
Engineering, UTM since 2005.
Nurzal Effiyana Ghazali received her B.E. (Hons) Electrical
(Telecommunications) from Universiti Teknologi Malaysia in 2007.
Master of Engineering in Electrical and Computer Science from Shi-
baura Institute of Technology, Japan in 2010 and Master of Engi-
neering in Electrical (Electronics and Telecommunications) from
Universiti Teknologi Malaysia in 2011. Currently, she is a Ph.D.
candidate and her research interests are Long Term Evolution (LTE),
WiMAX, WiFi, Network Mobility, IPv6 network, Handover Man-
agement in Proxy Mobile IPv6, Mobility Management and Mobile
Computing.
1946 N. ‘A. Amirrudin et al.
123
A preview of this full-text is provided by Springer Nature.
Content available from Wireless Personal Communications
This content is subject to copyright. Terms and conditions apply.
