cell configuration strategy for ofdma based hierarchical cell...
TRANSCRIPT
Cell Configuration Strategy for OFDMA Based
Hierarchical Cell Structure
Eunsung Oh
Department of Electronics Engineering, Hanseo University, Korea, 360-706
(e-mail: [email protected])
Abstract. This paper presents cell configuration strategies for orthogonal
frequency division multiple access (OFDMA) based hierarchical cell structure
(HCS) downlink systems. It is considered that HCS systems are used to extend
the data rate coverage. The cell configuration strategy problem is formulated
with spectral reuse planning and channel allocation constraints, and is proposed
the solution. The numerical results show that the proposed algorithm can
enhance the system performance of HCS systems.
Keywords: Cell configuration, hierarchical cell structure, orthogonal frequency
division multiple access, resource allocation.
1 Introduction
Wireless communication systems recently provide multimedia services such as voice,
video streaming and high-speed Internet access, etc. These services require high data
rate wireless communications. To provide various high data rates services in the same
location, it is preferable that some macro-cells overlay other microcells [1]. This
cellular structure is called the overlaid cell structure or hierarchical cell structure
(HCS) [2].
In HCS systems, UEs with different characteristics (i.e., mobility, required traffic,
etc.) are initialized at different cells. Therefore, to enhance the system performance,
the efficient cell configuration strategy is required in HCS systems [3].
It has been researched about the cell configuration algorithms [4] and the
frequency planning [5] in HCS systems. Prior works have shown that the HCS system
can enhance the system capacity. However, proposed strategies are based on the
heuristic approach. In addition, most of the schemes are proposed based on the
coordinated cell configuration across cells with several practical challenges.
In this paper, the cell configuration strategy is established for the total transmitted
data rate maximization in HCS downlink systems. Firstly, the optimization problem
for maximizing the total transmitted data rate is formulated. Using some assumptions,
the optimal cell configuration strategy is proposed for the total transmitted data rate
maximization. The control strategy is researched with partial channel state
information (CSI).
Advanced Science and Technology Letters Vol.108 (Mechanical Engineering 2015), pp.80-85
http://dx.doi.org/10.14257/astl.2015.108.18
ISSN: 2287-1233 ASTL Copyright © 2015 SERSC
2 System Model and Problem Formulation
2.1 System model
It is considered a system that its service area is divided into hexagonal macro-cells
of equal size with the base station (BS) at the center of each macro-cell, and micro-
cells are overlaid on each macro-cell service area. Macro/micro-cells use same
frequency assignment (FA). The subcarrier space is divided into successive groups. A
subchannel has one element from each group allocated. Subchannels are randomly
allocated to each UE. It is assumed partial CSI at the BS and perfect CSI at the UE.
The best serving cell selection is considered that each UE is served by the BS which
has the maximum signal strength. Let the sets of macro/micro BSs be B1 and B2
,
and the UEs served by macro-cells and micro-cells is presented as U1 and U2
,
respectively.
2.1 Problem formulation
Assuming the homogeneous system, the objective function for the total transmitted
data rate is calculated as
f Ru( ) = RuuÎU1
å + RuuÎU2
å (1)
where Ru [bits/sec] is the transmitted data rat of a UE u. The transmitted data rate is
modeled by the Shannon capacity as,
Ru
=BW
Nnulog
21+g
u
Rx( ) (2)
where nu is the number of subchannel allocated to a UE u, BW is the system
bandwidth, and gu is the transmitted SIR of a UE u, which is calculated as
gu
Tx =Gubpi
ai1
Gub
1
p1
b1ÎB
1
å +ai2
Gub
2
p2
b2ÎB
2
å, for uÎ U1,U2{ } and iÎ 1,2{ } (3)
where Gub is the channel power from BS b to UE u, p1 and p2
are the transmitted
power at macro/micro BS, and aij is the channel orthogonal factor between cell j
and cell i [6].
The optimization problem for maximizing the total transmitted data rate is
formulated as
Advanced Science and Technology Letters Vol.108 (Mechanical Engineering 2015)
Copyright © 2015 SERSC 81
maxr ,n
Ru
uÎU1
å + Ru
uÎU2
åìíï
îï
üýï
þï
s.t. Ru
³hu, "uÎ U
1,U
2{ }
0 £ ri£1, i Î 1,2{ }
nu
uÎUi
å / N £ ri, i Î 1,2{ }
(4)
3 Cell Configuration Strategy
One of the important usage models of HCS is the coverage extension. In this case, it
is assumed that UEs are uniformly distributed. Therefore, we assume that the spectral
reuse planning for macro-cell and micro-cell are equally controlled.
Based on the above assumption, the transmitted SIRs of (3) are calculated as
gu
1
Rx =Gubp
1
r1r
1Gub
1
p1
b1ÎB
1
å + r1r
2Gub
2
p2
b2ÎB
2
å=
1
r1
2gu
1
Tx for uÎU1
gu
2
Rx =Gubp
2
r2r
1Gub
1
p1
b1ÎB
1
å + r2r
2Gub
2
p2
b2ÎB
2
å=
1
r2
2gu
2
Tx for uÎU2
(5)
Using this model, the optimization problem of (4) is reformulated as
maxr ,n
R̂u
uÎU1
å + R̂u
uÎU2
åìíï
îï
üýï
þï
s.t. R̂u
³hu, "uÎ U
1,U
2{ }
0 £ x £1,
x £ ri, i Î 1,2{ }
nu
uÎUi
å / N £ ri, i Î 1,2{ }
(6)
where R̂u
=BW
Nnulog
21+g
u
Tx / ri
2( ) . The object function with R̂u is monotonic
decreasing in ri . Thus, the second constraint for the spectral reuse planning in (4) is
modified as the second and third constraints in (6). The optimization problem of (6) is
the convex optimization problem because the constraints are satisfied the convexity
and the object function is concave [7].
To solve the problem, the second and third constraints are relaxed using the
Lagrangian relaxation,
Advanced Science and Technology Letters Vol.108 (Mechanical Engineering 2015)
82 Copyright © 2015 SERSC
L n,r,x,l,m( ) = Ru
uÎUi
å + l 1- x( ) +m1
r1- x( ) +m
2r
2- x( )
(7)
From (7), the problem is decomposed into two parts. The first part is the problem
of the global spectral reuse planning, x,
maxx
- l +m1+m
2( ) x+ l (8)
Considering the slackness condition and the dual values, the solution of (8) is
determined as
x = min 1,r1,r
2( ) (9)
It is said that the system is not limited by CCI when x =1. However, if r1 =1 or
r2 =1 then the system is restricted by CCI effected to the macro- or the micro-cell.
The problem of the second part is constructed by the residual part of (7) with the
first and fourth constraints. It is decomposed to the independent problem per each cell.
Similar to solve the first part, the first and fourth constraints in (7) is relaxed using
the Lagrangian relaxation,
L n,r,m,n ,w( ) = Ru
uÎU1
å +m1r
1+ n
uRu-h
u( )uÎU
1
å +w1Nr
1- n
u
uÎU1
åæ
è
çç
ö
ø
÷÷ (10)
From KKT, slackness condition and the characteristic of dual variables, nu,w
1> 0,
it is expressed as
nu
=w1
N
BW
log2
log 1+gu
Tx
r1
2
æ
èçç
ö
ø÷÷
-1³ 0 w
1³BW
Nlog
21+
gu
Tx
r1
2
æ
èçç
ö
ø÷÷, "u ÎU
1
(11)
Considering the dual problem, w1 is calculated as
w1=BW
Nlog
21+
gu*
Tx
r1
2
æ
è
çç
ö
ø
÷÷ (12)
where u* = arg maxuÎU
1
gu
Tx( ).
The channel for each UE except UE u* is allocated as
2 2
1
lo g 1
u
u T x
u
n
N
B W
(13)
It means that the minimum channel satisfied the minimum required data rate is
allocated to UEs without UE, which has the best CSI. For the best CSI UE the
channel is allocated to satisfying the follow condition,
nu
gu
Tx / r1
2
1+gu
Tx / r1
2
1
log 1+gu
Tx / r1
2( )-
1
2
æ
è
çç
ö
ø
÷÷
uÎU1
å = 0 (14)
Advanced Science and Technology Letters Vol.108 (Mechanical Engineering 2015)
Copyright © 2015 SERSC 83
4 Simulation results
In our simulation, the system considered subchannelized OFDMA with 1024
subcarriers. The number of subchannel sets is 30 subchannels. The cellular system
being simulated consists of 19 two-tier hexagonal cells. In order to avoid the
boundary effect, the results from the center hexagonal cell are used. The simulation
values were chosen for urban environments [8]. To fairly compare, it is calculated the
spectral efficiency per a macro-cell area, i.e, one macro-cell plus microcells. In
figures, d means the distance from a macro-cell BS to a micro-cell BS.
2 6 10 14 18 22 26 301
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
Number of user equipment
Spe
ctr
al effic
iency [bits/s
ec/H
z]
without microcells
with 5 microcells at d=200m
with 5 microcells at d=600m
with 3 microcells at d=200m
with 3 microcells at d=600m
d=600m
d=200m
Fig. 1. Spectral efficiency versus number of micro-cell and UE with hu = 256[kbps / sec]
Spectral efficiencies verse number of UEs is shown in Fig. 2. Spectral efficiency
without micro-cells is increased in low traffic region, but that is decreased in medium
and high traffic region. That is why in low traffic case the system can be obtained the
multi-user diversity when the number of UEs are increased, but CCI constraints the
spectral efficiency in medium and high traffic region [12]. However, spectral
efficiencies with micro-cells are maintained in medium and high traffic region. It
means that HCS systems can be enhanced the system stability in high traffic region.
And, the spectral efficiency is improved when the number of micro-cell is increasing
and the distance between the macro-cell BS and the micro-cell BS becomes far off. It
is shown that HCS systems are obtained the performance enhancement by spreading
the traffic. Without loss of generality, when the micro-cell is posited at the macro-cell
edge, the enhancement is maximized in medium and high traffic region. In low traffic
region, the switching-off BS without serving UE is occurred. Therefore, the spectral
efficiency has the maximum value when the micro-cell is posited near the macro-cell
BS.
Advanced Science and Technology Letters Vol.108 (Mechanical Engineering 2015)
84 Copyright © 2015 SERSC
4 Conclusions
In this paper, it is proposed a cell configuration strategies for OFDMA based HCS
downlink systems. First, the optimization problem is formulated with spectral reuse
planning and channel allocation constraints, and obtained the optimum solution using
the Lagrangian relaxation because the optimization problem is the convex problem.
The proposed solutions are only required partial CSI and are able to be distributed cell
configuration. Results shown that HCS systems can enhance the system stability and
improve the system performance.
Acknowledgment. This research was supported by Basic Science Research
Program through the National Research Foundation of Korea (NRF) funded by the
Ministry of Science, ICT & Future Planning (No. 2013R1A1A1009526).
References
1. H. Claussen, L. T. W. Ho, and L. G. Samuel, “Financial analysis of a pico-cellular home
network deployment,” in Proc. IEEE Int. Conf. Communications, Jun. 24-28, 2007, pp.
5604–5609.
2. K. Son, E. Oh, and B. Krishnamachari. "Energy-efficient design of heterogeneous cellular
networks from deployment to operation." Com. Networks vol.78, pp. 95-106, 2015.
3. L. Ortigoza-Guerrero and A. H. Aghvami, Resource Allocation in Hierarchical Cellular
Systems. London, UK: Artech House, 1999.
4. D. G. Jeong and W. S. Jeon, “Load sharing in hierarchical cell structure for high speed
downlink packet transmission,” in Proc. IEEE Global Telecommunications Conf., vol. 2,
Nov. 17-21, 2002, pp. 1815–1819.
5. S. Kim, D. Hong, and J. Cho, “Hierarchical cell deployment for high speed data cdma
systems,” in Proc. IEsEE Wireless Communications and Networking Conf., Mar. 17-21,
2002, pp. 7–10.
6. E. Oh, M.-G. Cho, S. Han, C. Woo, and D. Hong, “Performance analysis of reuse-
partitioning-based subchannelized OFDMA uplink systems in multicell environments,”
IEEE Trans. Veh. Technol., vol. 57, no. 5, pp. 2617–2621, Jul. 2008.
7. G. Li and H. Liu, “On the optimality of the OFDMA network,” IEEE Commun. Lett., vol. 9,
no. 5, pp. 438–440, May 2005.
8. IEEE 802.16m Evaluation Methodology Document (EMD), IEEE Std. 802.16m, 2009.
9. E. Oh, S. Han, C. Woo, and D. Hong, “Call admission control strategy for system throughput
maximization considering both call- and packet-level QoSs,” IEEE Trans. Commun., vol.
56, no. 10, pp. 1591–1595, Oct. 2008.
Advanced Science and Technology Letters Vol.108 (Mechanical Engineering 2015)
Copyright © 2015 SERSC 85