Interfered Users Protection Algorithm for Self-Organizing
Networks
Chih-Hsiang Ho1, Li-Sheng Chen
1, Wei-Ho Chung
2, and Sy-Yen Kuo
1
1 Department of Electrical Engineering, National Taiwan University, Taipei, 10617, Taiwan
2 Research Center for Information Technology Innovation, Academia Sinica, Taipei, 11529, Taiwan
Email: {d97921025, d99921032, sykuo}@ntu.edu.tw; [email protected]
Abstract Small cells are generally used by telecom operators
to effectively share the loading of conventional macrocells and
backbone networks, because they cost considerably less to
deploy and are also capable of improving communication
quality and capacity. However, radio frequency spectrum
resources are limited and expensive, so using the same
frequency band to deploy small cells in densely populated
metropolitan environments could lead to overlapping signal
coverage between neighboring cells. If a user is situated in such
areas, this can result in severe radio signal interference and
result in a significant decline in network transmission
performance. This study therefore proposes combining cell
grouping and base station power adjustment, as well as radio
resource allocation algorithms, into an interference mitigation
method for small cells. According to simulation results, the
proposed Interfered Users Protection (IUP) algorithm produces
a significantly higher signal-to-interference-plus-noise ratio
(SINR) under varying numbers of base stations deployed, and
the effect is stronger under densely deployed base stations with
severe interference. Approximately 56% of the interfered users
maintained a data rate of 250 Kbps and above. In addition, the
average SINR of regular users were not affected; 67% of the
regular users were still capable of maintaining this transmission
rate. The IUP algorithm is capable of effectively improving
interfered users’ signals and transmission capacity, while
accounting for the transmission performance of the overall
network. Index Terms—Interfered user, regular user, cell grouping,
power adjustment, radio resource allocation
I. INTRODUCTION
Fourth Generation (4G) mobile communication
technology provides a higher data transfer rate and wider
signal coverage compared to its second and third
generation (2G and 3G) counterparts. However, several
studies and field investigations have discovered that over
90% of network services and two-thirds of the phone
calls in the cellular network occur in indoor environments
[1]. Simply deploying more macrocells is no longer
sufficient for improving communication quality and
capacity at a lower cost with high efficiency, as well as
mitigating the ever increasing volume of network traffic.
Therefore, the use of small cells in 4G technology and
standards (such as Long Term Evolution, LTE) has been
Manuscript received January 12, 2017; revised April 24, 2017.
doi:10.12720/jcm.12.4.234-239
proposed as the technological solution to improve
communication quality and offload macrocell and
backbone network traffic. Because small cells may be
deployed closely together in densely populated
metropolitan areas, this often results in overlapping
coverage between cells, as well as between small cells
and macrocells. The radio signals transmitted by Base
Stations (BS) with overlapping coverage can interfere
with each other and severely diminish network
performance. Automatic interference detection and
management have always been an essential research topic
in the Self-Organizing Networks (SON) field.
Interference management technology is a vital function of
self-optimization, which includes interference detection
and interference mitigation. As shown by the network
architecture of interference management in Fig. 1,
positioned at the back end of the Evolved Packet Core
(EPC) is a control center called the SON server, which is
responsible for the centralized collection and control of
BS and network parameters, as well as the execution of
interference detection and interference mitigation
algorithms [2], [3].
Fig. 1. Network architecture of interference management
Currently, analyses and improvements of network
interference problems are being actively explored and
researched. The problem of intercell interference was
noted in [4], and adaptive interference cancellation
Journal of Communications Vol. 12, No. 4, April 2017
234©2017 Journal of Communications
—
technology has been proposed to mitigate its effects.
Several studies have also employed approaches such as
radio resource management or power adjustment to solve
or reduce the interference [5]. A general description and
investigation of interference management and radio
resource allocation was offered in [6], whereas further
allocation of radio resources to improve use efficiency in
overlapping coverage is explored in [7]. A femtocell-
based management system that calculates and allocates
subcarriers to each femtocell was proposed in [8]. The
conceptions in [9] and [10] were constructed based on
graph coloring theory to develop an interference
reduction mechanism, whereas [11] and [12] used game
theory to consider the interference factors to construct the
optimal strategy for BS actual behavior prediction and
power adjustment. A distributed power adjustment
method was proposed in [13] to effectively improve the
signal quality for users at the cell edge of the base
stations. Power adjustment approaches were proposed
based on the Quality of Service (QoS) in [14] and [15] to
improve signal quality in environments with interference.
In addition, several studies on interference reduction used
the cognitive technology of wireless channels to avoid
interference on the basis of cognitive network technology,
as seen in [16] and [17], whereby cognitive radio was
used to allocate radio resources and transmit power.
Despite the many advantages of small cells, the severity
of their impact on communication quality due to
interference problems determines to a large extent their
actual transmission efficiency and application. In order to
overcome the interference problem, this study proposed
the interfered user protection (IUP) algorithm to improve
signal and connection quality to users in the interference
area. This study proposes using cell grouping, power
adjustment, and dynamic radio resource allocation
scheme with interference mitigation technology to inhibit
interference and improve overall network transmission
efficiency.
This paper is organized as follows. The system model
is described in Section II, IUP algorithm is proposed in
Section III, and performance evaluations are described in
Section IV. Concluding remarks are presented in Section
V.
II. SYSTEM AND PROBLEM MODELING
Under the dense and co-channel deployment of small
cells, the coverage of small cells overlaps with
neighboring small cells and macrocells. While user
moves into the overlapping area, the physical-layer signal
experiences severe interference, and the communication
quality and capacity are severely degraded. This section
elaborates and defines the system. Considering that the
small cell experiences interference from neighboring
small cells and macrocells, the SINR of the user j
connected to small cell i is expressed as:
𝑆𝐼𝑁𝑅 𝑖,𝑗
𝑘=
𝑎𝑖,𝑗𝑘 𝑝𝑖,𝑗
𝑘 𝑔𝑖,𝑗𝑘 𝑃𝐿𝑖,𝑗
∑ 𝑝𝑖′, 𝑗𝑘 𝑔
𝑖′𝑘
𝑖≠𝑖∗ 𝑃𝐿𝑖′,𝑗
+ 𝑁0 (1)
where 𝑎𝑖,𝑗𝑘 represents whether the physical resource block
(PRB) k at base station i has been allocated to user j ,
where 𝑝𝑖,𝑗𝑘 represents the power of the PRB k at the small
cell i received by user j, 𝑔𝑖,𝑗𝑘 represents the antenna gain
of PRB k at small cell i, 𝑃𝐿𝑖,𝑗 represents the path loss
between base station i and user j, 𝑝𝑖′, 𝑗𝑘 represents the
power of PRB k at another small cell 𝑖′ received by user j,
𝑔𝑖′ ,𝑗𝑘 represents the antenna gain for PRB k at another
base station 𝑖′, 𝑃𝐿𝑖′,𝑗 represents the path loss between the
small cell 𝑖′ and user j, and 𝑁0 represents white Gaussian
noise. The data rate DRi,j for user j at small cell i can be
expressed as:
𝐷𝑅𝑖,𝑗 = ∑ 𝑎𝑖,𝑗 𝑘 . 𝐵 . 𝑙𝑜𝑔2 (1 + 𝑆𝐼𝑁𝑅
𝑖,𝑗
𝑘)
|𝐾|𝑘 (2)
where 𝑎𝑖,𝑗𝑘 represents whether the PRB k at small cell i
has been allocated to user j; B represents the bandwidth
of each PRB; and 𝑆𝐼𝑁𝑅 𝑖,𝑗
𝑘 represents the SINR for user j,
who is connected to small cell i and who has been
allocated PRB k.
The objective function of the proposed algorithm can
be expressed as follows:
𝑀𝑎x 𝛼 . ∑ ∑ 𝐷𝑅𝑖,𝑗
| 𝑈𝑅𝐶𝑈𝑖 |
𝑗=1 + 𝛽 . 𝑅𝑅𝐶𝑈
|𝑆|𝑖=1 (3)
Subject to:
𝐷𝑅𝑖,𝑗 ≥ 𝐷𝑖,𝑗_𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑 , 𝑗 ∈ 𝑈𝑅𝑈𝑖 (4)
∑ 𝑎𝑖,𝑗𝑘 =
|𝑈𝑖|
𝑗=1 1 (5)
𝛼 + 𝛽 = 1 (6)
where S represents the set of small cells, 𝑈𝑅𝐶𝑈𝑖 represents
the set of users that can successfully connect to base
station i after recovery, 𝑅𝑅𝐶𝑈 represents the proportion of
users in the interference area whose connections are
restored after effective recovery (i.e., recovered users,
RCUs), 𝐷𝑅𝑖,𝑗 represents the data rate for user j at cell i,
𝐷𝑅𝑖,𝑗_𝑡ℎ𝑟𝑒𝑠ℎ𝑜𝑙𝑑 represents the threshold data rate of
regular user, 𝑈𝑅𝑈𝑖 represents the set of regular users at
small cell i, 𝑎𝑖,𝑗𝑘 represents whether PRB k at small cell i
has been allocated to user j, and 𝛼 and 𝛽 are weight
values.
The IUP algorithm is proposed to improve the signal
and transmission quality of the interfered users through
BS power adjustment and radio resource allocation
without reducing the connection quality of other regular
users, thereby enhancing the overall network capacity.
Journal of Communications Vol. 12, No. 4, April 2017
235©2017 Journal of Communications
III. I PROTECTION ALGORITHM
Before the BS can perform dynamic power
adjustments, the SINR of the downlink from the BS to
UE must be obtained, as well as the reference signal
received power (RSRP) from the other interfering BS. If
the SINR of a user is smaller than the SINR threshold,
and the interfering RSRP is larger than the interfering
RSRP threshold, we define this user as the interfered user.
The pseudocode for our proposed IUP algorithm is shown
in Fig. 2.
After classifying the interfered user, the SON server
calculates the number of interfered user in each BS and,
based on this number of interfered user, categorizes the
base stations. The base stations with higher numbers of
interfered user are placed in the high-priority group.
Power adjustments are then performed. Generally, a
higher transmitting power of a BS results in a higher
receiving power for the users associated with it, and a
higher interference for users associated with other base
stations. However, the lower transmitting power of a BS
results in lower receiving power for its associated users,
and lower interference for users associated with other
base stations. The base stations in the higher priority
group have more interfered users, and their power
adjustments are expected to induce a higher influence on
the overall SINR. Therefore, power adjustments are first
performed on the base stations in the higher priority
group. The power is initially increased by 1 unit from its
initial power level, and then decreased by 1 unit from the
initial power level; the average power levels
corresponding to the two adjustment directions are
checked, and the power level corresponding to the higher
average SINR is maintained.
Subsequently the power adjustment continues in the
direction of increasing the average SINR until one of the
following conditions is met: (1) the average SINR of
interfered user ceases to increase; (2) the upper bound or
lower bound of power levels is reached; (3) the power
adjustment leads to the downgrade of SINR level in a
user; (4) any of the regular users’ status is changed to
interfered user. The base stations without interfered user
continue to decrease transmission power until one of the
above four conditions are met. Finally, the power levels
of each BS can be obtained and set.
IUP algorithm
Notation:
𝑔: The groups with number of interfered user of BS
G: The set of base station group
BS𝑔: The BS is belong to the group 𝑔
𝑝𝑜𝑤𝑒𝑟𝑔: The initial tx power of BS of group 𝑔
𝑝𝑜𝑤𝑒𝑟𝑔∗: The adjusted tx power of BS of group 𝑔
SINR𝑟ui.j: The SINR of regular user j of cell i
SINR𝑟𝑢𝑖,𝑗
lower_bound: The SINR threshold of regular user j of cell i
𝜏: The constant guard interval of SINR for power adjustment
process
∆P: The minimum adjustment step of tx power forBS
𝑂𝑏𝑗+: The value of objective function after increasing power of
BS𝑔by ∆P 𝑂𝑏𝑗−: The value of objective function after decreasing
power of BS𝑔by ∆P
𝑂𝑏𝑗∗: The current value of objective function
K: The set of PRBs
U: The set of users
𝑢𝑖,𝑗 : The user 𝑗 of the cell i
𝜔𝑖,𝑗 : The weighting factor of user 𝑢𝑖,𝑗
𝜑𝑢𝑖,𝑗: The number of PRBs allocated to 𝑢𝑖,𝑗
σui,j: The number of PRBs requested by 𝑢𝑖,𝑗
𝑆𝐼𝑁𝑅𝑖,𝑗𝑘 : The SINR at PRB k of 𝑢𝑖,𝑗
𝑎𝑗∗ 𝑘𝑖 : The PRB k at small cell i has been allocated to user 𝑗∗
𝑆𝑗∗𝑖 : Non-interference relationship set of 𝑢𝑖,𝑗∗
𝐵𝑛𝑘 : The updating user set for resource allocation
BEGIN
Grouping base stations with the number of interfered users into
groups
FOR 𝑔 = |𝐺| to 1
Frist try to increasing and decreasing power of BS𝑔by ∆P
IF 𝑂𝑏𝑗+ > 𝑂𝑏𝑗− && 𝑂𝑏𝑗+ > 𝑂𝑏𝑗∗ then
𝐑𝐄𝐏𝐄𝐀𝐓
𝑂𝑏𝑗∗ = 𝑂𝑏𝑗+
𝑝𝑜𝑤𝑒𝑟𝑔= 𝑝𝑜𝑤𝑒𝑟𝑔 + ∆𝑃
calculate 𝑂𝑏𝑗+
𝑝𝑜𝑤𝑒𝑟𝑔∗ = 𝑝𝑜𝑤𝑒𝑟𝑔
UNTIL 𝑂𝑏𝑗+ < 𝑂𝑏𝑗∗ || 𝑝𝑜𝑤𝑒𝑟𝑔∗ is maximum tx power
of BS || SINR𝑟ui.j+ 𝜏 ≤ SINR𝑟𝑢𝑖,𝑗
lower_bound ||
regular user become the interfered user
RETURN 𝑝𝑜𝑤𝑒𝑟𝑔∗ and 𝑂𝑏𝑗∗
ELSEIF 𝑂𝑏𝑗− > 𝑂𝑏𝑗+ && 𝑂𝑏𝑗− > 𝑂𝑏𝑗∗ then
𝐑𝐄𝐏𝐄𝐀𝐓
𝑂𝑏𝑗∗= 𝑂𝑏𝑗−
𝑝𝑜𝑤𝑒𝑟𝑔 = 𝑝𝑜𝑤𝑒𝑟𝑔 − ∆𝑃
calculate 𝑂𝑏𝑗−
𝑝𝑜𝑤𝑒𝑟𝑔∗ = 𝑝𝑜𝑤𝑒𝑟𝑔
UNTIL 𝑂𝑏𝑗− < 𝑂𝑏𝑗∗ || 𝑝𝑜𝑤𝑒𝑟𝑔∗ is minimum tx power
of BS || SINR𝑟ui.j+ 𝜏 ≤ SINR𝑟𝑢𝑖,𝑗
lower_bound ||
regular user become the interfered user
RETURN 𝑝𝑜𝑤𝑒𝑟𝑔∗ and 𝑂𝑏𝑗∗
ELSE
without changing the current power
ENDIF
ENDFOR
FOR 𝑔 = 0
𝐑𝐄𝐏𝐄𝐀𝐓
𝑂𝑏𝑗∗ = 𝑂𝑏𝑗−
𝑝𝑜𝑤𝑒𝑟𝑔 = 𝑝𝑜𝑤𝑒𝑟𝑔 − ∆𝑃
calculate 𝑂𝑏𝑗−
𝑝𝑜𝑤𝑒𝑟𝑔∗ = 𝑝𝑜𝑤𝑒𝑟𝑔
UNTIL 𝑂𝑏𝑗− < 𝑂𝑏𝑗∗ || 𝑝𝑜𝑤𝑒𝑟𝑔∗ is minimum tx power of BS
Journal of Communications Vol. 12, No. 4, April 2017
236©2017 Journal of Communications
NTERFERED USER
|| SINR𝑟ui.j+ 𝜏 ≤ SINR𝑟𝑢𝑖,𝑗
lower_bound || regular user
become the interfered user
RETURN 𝑝𝑜𝑤𝑒𝑟𝑔∗ and 𝑂𝑏𝑗∗
ENDFOR
Execute the dynamic radio resource allocation after the power
adjustment
FOR 𝑘 = 1 to |𝐾|
𝑛 = 1 ;
𝐵𝑛𝑘 = U ;
DO
𝑢𝑖,𝑗∗ = argmax𝑗∈ 𝐵𝑛
𝑘 𝜔𝑖,𝑗
𝜑𝑢𝑖,𝑗/𝜎𝑢𝑖,𝑗
. 𝑆𝐼𝑁𝑅𝑖,𝑗𝑘
𝑎𝑗∗ 𝑘𝑖 = 1
𝑆𝑗∗𝑖 is set to empty set
Find the non-interference relationship set 𝑆𝑗∗𝑖 of user
𝑢𝑖,𝑗∗
n = n + 1;
𝐵𝑛𝑘 = 𝐵𝑛−1
𝑘 ∩ 𝑆𝑗∗𝑖
WHILE 𝑘 ≤ |𝐾| 𝑎𝑛𝑑 𝐵𝑛𝑘 ≠ ∅
ENDFOR
END
Fig. 2. Pseudocode for IUP algorithm
Because the interfered users differ from regular users
who possess more stable radio resources allocated from
the BS, the interfered users should have a higher priority
than the regular users in obtaining radio resources during
the process of BS power adjustment and radio resource
allocation. Therefore, the dynamic radio resource
allocation mechanism was added, where 𝜔𝑖,𝑗 represents
the weight of the j-th user in the i-th BS. If j belongs to
the interfered user set, then 𝜔𝑖,𝑗 will be higher; 𝜑𝑢𝑖,𝑗
represents the number of PRBs allocated to 𝑢𝑖,𝑗, and 𝜎𝑢𝑖,𝑗
represents the number of PRBs requested by 𝑢𝑖,𝑗. For the
proposed PRB dynamic allocation mechanism, each PRB
is first individually allocated to the 𝑢𝑖,𝑗∗ that yields the
highest 𝜔𝑖,𝑗
𝜑𝑢𝑖,𝑗/𝜎𝑢𝑖,𝑗
. 𝑆𝐼𝑁𝑅𝑖,𝑗 . The PRB can also be
allocated to users in other cells who do not have an
interfering relationship with 𝑢𝑖,𝑗∗ [18]. The priority
allocation of higher quality PRBs to interfered users was
achieved through 𝜔𝑖,𝑗 to restore the stability of the
interfered users’ signal and attain an adequate
transmission quality; the fairness of allocation was
simultaneously improved through 𝜑𝑢𝑖,𝑗/𝜎𝑢𝑖,𝑗
to maintain
the transmission quality of other regular users.
IV. PERFORMANCE ASSESSMENT
The proposed IUP algorithm was compared to the QoS
priority-based dynamic frequency allocation algorithm
(QPDFA) [13] and the opportunistic spectrum access and
channel dynamics algorithm (OSACD) [14] to assess its
performance. In the simulation environment, the distance
between base stations was defined as 50~200 m, the
antenna downtilt was defined as 15°, the maximum base
station output power was defined as 30 dBm, the
maximum UE output power was defined as 20 dBm, and
the antennas were defined as 3GPP models [19]. Table I
lists these and other relevant simulation parameters.
Algorithm performance was assessed by four indicators:
(1) the average SINR of interfered users; (2) the average
SINR of regular users; (3) the cumulative distribution
function (CDF) of the data rate among interfered users;
(4) the CDF of data rate among regular users.
TABLE I: SIMULATION PARAMETERS
Number of BS 10~50
Inter-site distance 50~200 m
Antenna downtilt 15°
System bandwidth 10 MHz
Maximum Base Station
output power 30 dBm
Maximum UE output
power 20 dBm
Path Loss 128.1 + 37.6 log10
r, with r in km
Antenna model 3GPP model [19]
Noise level -199 dBW/Hz in DL, -195
dBW/Hz in UL
Service
Generic elastic data service
with a requested throughput of
0.5 ~ 1 Mbps (DL)
Fig. 3 shows the average SINR of the interfered users
regulated by the three interference mitigation algorithms
under differing numbers of base stations deployed (i.e.,
the average SINR of the interfered users was significantly
higher regardless of the number of base station deployed).
The IUP algorithm displayed a superior performance
especially in the case of a large number of BSs (dense BS
deployment) with more severe interference.
Fig. 3. Average SINR of interfered users under differing
Fig. 4 shows the average SINR of regular users by IUP
algorithm was also maintained above a specific level,
Journal of Communications Vol. 12, No. 4, April 2017
237©2017 Journal of Communications
numbers
of base stations
meaning that the transmission quality of the other regular
users were not sacrificed to enhance the average SINR of
the interfered users. In terms of the SINR of the regular
user, despite the inclination of the IUP algorithm to
increase the SINR of the interfered user, it was also
careful not to diminish the SINR of other regular user to
values lower than the minimum of the current SINR level
during the power adjustment process, and improved the
allocation fairness through 𝜔𝑖,𝑗
𝜑𝑢𝑖,𝑗/𝜎𝑢𝑖,𝑗
. Therefore, the
transmission quality for regular user was not reduced
while protecting the interfered user.
Fig. 4. Average SINR of regular users under differing number of
small cells
Fig. 5. Data rate CDF of interfered users
Increasing the SINR and allocating more radio
resources to the interfered users resulted in a more stable
signal quality for interfered users and a higher data rate.
As shown in Fig. 5, 56% of the interfered users’ data
rates were increased by the IUP algorithm to 250 Kbps or
above, whereas the other two algorithms achieved only
40%. Nor did the IUP algorithm sacrifice the
transmission quality of the other regular users while
maintaining the interfered users’ data rates. As shown in
Fig. 6, 67% of regular users maintained a data rate of 250
Kbps and above, a result similar to those of the other two
algorithms. The reason was because before the resource
allocation prioritization mechanism was initiated, the
proposed method did not reduce the SINR of regular
users to a level lower than that of the current regulation,
meaning it did not sacrifice the transmission quality of
the other regular users to enhance the transmission
efficiency of the interfered users.
Fig. 6. Data rate CDF of regular users
V. CONCLUSION
To address the interference problem of small cells, this
study proposed an IUP algorithm, using interfered user
determination, cell grouping to dynamically adjust the
transmission power of base stations and radio resource
allocation for improved performance. The effects of
interference on users can be mitigated to improve signal
quality without affecting the connection quality for other
regular users. The simulation results demonstrated that
compared with the other methods, the proposed method
resulted in both a higher SINR and data rate for interfered
users. The IUP algorithm effectively improved the signal
quality of the interfered users to facilitate a stable
transmission, and simultaneously maintained the fairness
and transmission efficiency of other regular users.
REFERENCES
[1] J. Zhang and G. D. L. Roche, Femtocells: Technologies
and Deployment, United Kingdom: John Wiley & Sons Ltd,
2010.
[2] 3GPP TS 32.500, “Telecommunication management; Self-
Organizing Networks (SON); Concepts and requirements
(Release 11),” 2011.
[3] 3GPP TS 28.627, “Telecommunication management; Self-
Organizing Networks (SON) Policy Network Resource
Model (NRM) Integration Reference Point (IRP);
Requirements, (Release 13),” 2015.
[4] Z. Jun and J. G. Andrews, “Adaptive spatial intercell
interference cancellation in multicell wireless networks,”
IEEE Journal on Selected Areas in Communications, vol.
28, pp. 1455-1468, Dec. 2010.
[5] H. C. Lee, D. C. Oh, and Y. H. Lee, “Mitigation of inter-
femtocell interference with adaptive fractional frequency
reuse,” in Proc. IEEE International Conference on
Communication, May 2010, pp. 1-5.
[6] M. Peng, C. Wang, J. Li, H. Xiang, and V. Lau, “Recent
advances in underlay heterogeneous networks: Interference
control resource allocation and self-organization,” IEEE
Commun. Surveys Tuts., vol. 17, no. 2, pp. 700-729, 2015.
[7] K. I. Pedersen, “Enhanced inter-cell interference
coordination in co-channel multilayer LTE-Advanced
Journal of Communications Vol. 12, No. 4, April 2017
238©2017 Journal of Communications
Networks,” IEEE Wireless Commun. Mag., vol. 20, no. 3,
pp. 120-27, June 2013.
[8] H. C. Lee, D. C. Oh, and Y. H. Lee, “Mitigation of inter-
femtocell interference with adaptive fractional frequency
reuse," in Proc. IEEE International Conference on
Communication, May 2010, pp. 1-5.
[9] T. Li, Z. Feng, W. Li, Z. Jing, and T. A. Gulliver, “Graph
coloring based spectrum allocation for femtocell downlink
interference mitigation,” in Proc. IEEE Wireless
Communication Networking Conference, Mar. 2011, pp.
1248-1252.
[10] S. Uygungelen, G. Auer, and Z. Bharucha, “Graph-based
dynamic frequency reuse in femtocell networks,” in Proc.
IEEE Vehicular Technology Conference, May 2011, pp. 1-
6.
[11] L. Zou, B. Liu, and C. Chen, “Bayesian game based power
adjustment scheme for inter-WBAN interference
mitigation,” in Proc. IEEE Global Communications
Conference, 2014, pp. 240–245.
[12] Y. Zhang and Q. Zhao, “Channel selection with dynamic
users: A game-theoretic approach,” in Proc. WCSP IEEE,
2014.
[13] V. Chandrasekhar, J. G. Andrews, Z. Shen, T.
Muharemovic, and A. Gatherer, “Distributed power
adjustment in femtocell-underlay cellular networks,” in
Proc. IEEE GLOBECOM, 2009, pp. 1-6.
[14] I. Ahmad, Z. Kaleem, and K. H. Chang, “QoS priority
based femtocell user power adjustment for interference
mitigation in 3GPP LTE-A HetNet,” J-KICS, vol. 39A, no.
2, pp. 61-74, Feb. 2014.
[15] Z. Kaleem, B. Hui, and K. H. Chang, “QoS priority-based
dynamic frequency band allocation algorithm for load
balancing and interference avoidance in 3GPP LTE
HetNet,” EURASIP J. Wireless Commun. Netw., vol. 2014,
pp. 1-18, Nov. 2014.
[16] Y. Zhang, Y. Xu, Q. Wu, et al., “Optimal opportunistic
spectrum access with heterogeneous and unknown channel
dynamics in cognitive radio network,” KSII Transactions
on Internet and Information Systems, vol. 8, no. 6, pp.
2675-2690, 2014.
[17] H. Kpojime and G. Safdar, “Interference mitigation in
cognitive radio based femtocells,” IEEE Commun. Surveys
Tuts., vol. 17, no. 3, pp. 1511-1534, 2015.
[18] L. Narayanan, “Channel assignment and graph
multicoloring,” in Handbook of Wireless Networks and
Mobile Computing, Wiley, 2002, pp. 71–94.
[19] 3GPP TR 36.814, “Further advancements for E-UTRA–
Physical layer aspects,” v1.0.1, 2009.
Chih-Hsiang Ho received his B.S. and
M.S. degrees in aeronautical engineering
and applied mechanics from the Tamkang
University and National Taiwan
University, Taiwan, in 1996 and 1998,
respectively. Since 2000, he has been
with Institute for Information Institute
(III), Taiwan. Currently, he is a section
manager of III and working towards his Ph.D. degree in the
Department of Electrical Engineering at National Taiwan
University. His research interests include self-organized
network algorithms in HetNets, hybrid home networking and
broadband wireless communication such as LTE, WiMAX.
Li-Sheng Chen received the M.S. degree
in Electronic Engineering from National
I-Lan University, I-Lan, Taiwan. He is a
software engineer of Institute for
Information Institute (III) and working
toward his Ph.D. degree in the
Department of Electrical Engineering at
National Taiwan University. His research
interests include the self-organized network, wireless
communication and computer networking.
Wei-Ho Chung received received the
B.S. and M.S. degrees in elctrical
engineering from National Taiwan
University, Taipei, Taiwan, in 2000 and
2002, respectively, and the Ph.D. degree
in electrical engineering from the
University of California, Los Angeles, in
2009. From 2002 to 2005, he was with
ChungHwa Telecommunications
Company. In 2008, he was involved on CDMA systems with
Qualcomm, Inc., San Diego, CA. Since 2010, he has been an
Assistant Research Fellow in Academia Sinica, Taiwan, and
was promoted to the rank of an Associate Research Fellow in
2014. He holds an adjunct associate professor positions at the
Department of Electrical Engineering, National Taiwan
University, and the Department of Communication Engineering,
National Taipei University. He leads the Wireless
Communications Laboratory, Research Center for Information
Technology Innovation, Academia Sinica. His research interests
include communications, signal processing, and networks.
Sy-Yen Kuo received the B.S. degree in
electrical engineering from National
Taiwan University, Taipei, Taiwan, in
1979, the M.S. degree in electrical and
computer engineering from the
University of California, Santa Barbara,
CA, USA, in 1982, and the Ph.D. degree
in computer science from theUniversity
of Illinois at Urbana-Champaign, Champaign, IL, USA, in 1987.
From 1982 to 1984, he was an Engineer with Fairchild
Semiconductor, Sunnyvale, CA, USA, and Silvar-Lisco, Menlo
Park, CA, USA. From 1988 to 1991, he was a Faculty Member
with the Department of Electrical and Computer Engineering,
University of Arizona. He was the Chairman in the Department
of Electrical Engineering, NTU, from 2001 to 2004. He took a
leave from NTU and served as a Chair Professor and Dean of
the College of Electrical Engineering and Computer Science,
National Taiwan University of Science and Technology, from
2006 to 2009. He was the Dean of the College of Electrical
Engineering and Computer Science, NTU, from 2012 to 2015.
He is a Distinguished Professor with the Department of
Electrical Engineering, NTU. His current research interests
include dependable systems and networks, mobile computing,
cloud computing, and quantum computing and communications.
He also received the Best Paper Award in the International
Symposium on Software Reliability Engineering in 1996, and
the IEEE/ACM Design Automation Scholarship in 1990 and
1991, respectively.
Journal of Communications Vol. 12, No. 4, April 2017
239©2017 Journal of Communications