1 Introduction
1.1 Previous work
1.2 Contribution of this work
2 Formulation of design criteria for placing femtocells
2.1 Minimum average pathloss
2.2 Maximum worst connection quality
2.3 Maximum average connection quality
2.4 Maximum cell overlapping
3 Preliminary analysis of femtocell location strategies
3.1 Analysis setup
Scenario
|
Path loss
|
σ
SF
| ||
---|---|---|---|---|
Ind2ind | A1 | LOS |
A=18.7, B=46.8, C=20 | 6 |
NLOS |
A=20B=46.4C=20 | |||
X=5n
w
(wall att.) | ||||
FL=17+4(n
f
−1) (floor att.) | ||||
Ind2out | A2 | PL=PL
b
(d
in+d
out)+ | 6 | |
PLtw+PLin(d
in) | ||||
PL
b
→A=22.7, B=41, C=20 | ||||
PL
tw
=14+15(1−cos(θ)) | ||||
PLin=PL
A1
| ||||
Out2out | C2 | LOS |
d<d
BP→A=26B=39C=20 | 4 |
d>d
BP→PL=40 log(d)+13.47+ | 6 | |||
+14 log(h
BS−1)−14 log(h
MS−1)+ | ||||
+6 log(f
c
/5) | ||||
NLOS | PL=(44−6.55 log(h
BS))log(d)+ | 8 | ||
+34.46+5.83 log(h
BS)+23 log(f
c
/5) | ||||
Out2ind | C4 | PL=PL
C2(d
in+d
out)+17.4+ | 8 | |
0.5d
in−0.8h
MS
|
3.2 Analysis results
3.2.1 Coverage layout (CL)
3.2.2 Worst-case SINR layout (WSL)
3.2.3 Average SINR layout (ASL)
3.2.4 Overlapping layout (OvL)
3.2.5 Overall performance comparison
Scenario
|
\(\boldsymbol {\overline {\text {PL}}\:[\!\text {dB}]}\)
|
L
10%, DL SINR
[dB]
|
\(\boldsymbol {\overline {\text {SINR}}_{\text {DL}}\:[\!\text {dB}]}\)
|
LDR [%]
|
---|---|---|---|---|
OrL | 60.50 | 0.92 | 7.98 | 14.41 |
CL |
59.29
| 0.83 | 8.99 | 11.95 |
WSL | 62.52 |
2.70
| 8.99 | 6.48 |
ASL | 59.39 | 0.74 |
9.01
| 13.03 |
OvL | 60.84 | -1.25 | 7.27 |
21.23
|
4 Traffic sharing techniques
-
Margin traffic sharing (MTS): HO margins are modified on a per-adjacency basis (i.e., MarginHO(i,j)) with the aim of balancing CBR between adjacent cells i and j. Changes of the same amplitude and opposite signs are performed in margins of both directions of the adjacency to maintain cell overlapping, i.e.:$$ \text{MarginHO}\left(i,j\right) + \text{MarginHO}\left(j,i\right) = \text{Hyst} \, $$(5)where Hy s t is a constant defining the hysteresis value. In this work, H y s t=6 dB and the default value of MarginHO(i,j) is 3 dB. Margins are restricted to a limited interval between -7 and 13 dB to avoid connection quality problems [14].
-
Power traffic sharing (PTS): Femtocell transmit power is modified on a per-cell basis (i.e., P TX (i)) with the aim of balancing CBR of cell i compared to the average CBR of its neighbors. A femtocell decreases (increases) power if its CBR is larger (smaller) than that of its neighbors. It is assumed that both data and pilot power are jointly tuned. Thus, traffic steering is effective not only for connected users but also for idle users, since it has an impact on both cell reselection and HO processes.
-
Combined traffic sharing (CTS): CTS modifies both P TX and MarginHO parameters. By alternating both changes, the limits inherent to individual approaches (i.e., MTS and PTS) are overcome. First, MarginHO settings are modified while P TX settings remain unchanged. Only when all MarginHO(i,j) values in cell i have reached their limits, P TX(i) is modified. As a result, traffic sharing in cell i is achieved with minimal deviation of transmit power from default values (and, hence, minimal impact on network coverage) [14].
5 Analysis of traffic sharing schemes for different femtocell location plans
5.1 Simulation setup
Simulation parameters and features
| ||
---|---|---|
Time resolution | 100 ms | |
Propagation models | Indoor-indoor | Winner II A1 |
Indoor-outdoor | Winner II A2 | |
Outdoor-outdoor | Winner II C2 | |
Outdoor-indoor | Winner II C4 | |
BS model | EIRP |
B
S
femto=13 dBm |
B
S
macro=43 dBm | ||
Directivity | femto: omnidirectional | |
macro: tri-sectorial | ||
Access | macro/femto: open access | |
MS model | Noise figure | 9 dB |
Noise spectral density | -174 dBm/Hz | |
Traffic model | Arrival process | Poisson |
(avg. 0.42 | ||
calls / (user ×hour)) | ||
Call duration | exponential (avg. 180 s) | |
Mobility model | Outdoor | 3 km/h, random direction and |
wrap-around | ||
Indoor | Random waypoint | |
Service model | Voice over IP | 16 kbps |
RRM model | 6 PRBs (1.4 MHz) | |
Cell reselection | C1-C2 | |
Access control | Directed retry | |
Handover | Power budget | |
Scheduler: | Time: round-robin | |
Frequency: best channel |
5.2 Simulation results
Scenario
|
CBR [%]
|
OR [%]
|
UUR [%]
|
---|---|---|---|
OrL | 0.08 | 0.41 | 0.49 |
CL (minimum pathloss) | 0.02 | 0.49 | 0.51 |
WSL (maximum 10th percentile of SINR) | 0.13 | 0.67 | 0.79 |
ASL (maximum SINR average) | 0.04 | 0.68 | 0.72 |
OvL (maximum overlapping) | 0.44 | 2.32 | 2.74 |
Scenario
|
UUR
(1)
[
%
]
|
UUR
(25)
[
%
]
|
UUR
impr
[
%
]
|
---|---|---|---|
OrL | 17.20 | 12.30 | 28.48 |
CL (minimum pathloss) | 15.55 | 10.55 | 32.14 |
WSL (maximum 10th-percentile | 18.18 | 12.86 | 29.27 |
of SINR) | |||
ASL (maximum SINR average) | 17.68 | 12.45 | 29.54 |
OvL (maximum overlapping) | 14.00 | 5.67 | 59.49 |