HCS in UMTS/HSPA Systems

Tiago A. S. Rebelo, Antnio Rodrigues

Instituto Superior Tcnico / Instituto de Telecomunicaes
Technical University of Lisbon
Lisbon, Portugal
e-mail: [email protected], [email protected]

Abstract New applications put out by the service providers II. SYSTEM MODELS
require high data rates in order to enjoy the best performance
of its functionalities, hence to have a good Quality of Service Four distinct HCS models were adopted to simulate a
(QoS). To provide the system with a solution to enable a good network environment. The deployed scenarios are presented in
service to user equipments one can adopt the HCS Figure 1.
(Hierarchical Cell Structure) model. The model is considered As shown in the figure, three of the strategies use cell
to improve the systems capacity, coverage and quality of partition with co-located base stations, in Figure 1-(a), (b) and
service. In this paper, several HCS scenarios are implemented (c). The other scenario uses a micro-cell located at the macro-
over the UMTS, HSDPA and HSUPA technologies. Four cell edge, as shown in Figure 1-(d).
distinct profile users are considered and a snapshot method
was used to simulate the mobile users. The scenarios are
individually analyzed and submitted to performance
comparisons. The results show that the deployed scenario with
higher number of co-located cell partitions is the one that
provides a higher system capacity, i.e. better cell throughput,
and also higher service rates (near 95%). The model with cell-
edge topology allowed the increase of the inter-cell area
throughput.
(a) (b)
Keywords UMTS, HSDPA, HSUPA, Cell Partitioning,
Hierarchical Cell Structure

I. INTRODUCTION

To provide high data rates it is essential to increase the

networks capacity and also the quality of service. Recently
introduced technology like HSDPA, in 3GPPs Release 5, and
HSUPA, in Release 6, can provide theoretically data rates up to
14.4 Mbps and 4 Mbps respectively [1]. (c) (d)
It is expected an increase of the systems capacity with the
deployment of the HCS model in the UMTS system. Several Figure 1. Cell deployment strategies for the HCS model
models are proposed and evaluated in the following sections.
The systems analysis is performed concerning the
downlink and uplink connections. This paper evaluates the In Figure 1 the colors represent distinct frequency carriers
hierarchical system performance of HSDPA and HSUPA in and also distinct emitted base station power.
WCDMA networks underlying the development of cell The macro-cell is used to provide the system with a better
deployment strategies. Some of the scenarios have already coverage area while the micro-cells have the utility of serving
been used in previous works [2] and [3] the most demanding users in terms of bit rate, i.e., a better
Section II presents the system model used in the simulation quality of service, functioning as a hot-spot.
process. The simulators main functions and implementation The macro cell base stations are spaced by 600 meters. Tri-
are presented in Section III. Simulations results are presented sectored, with 65 degrees antennas, and omni-directional
and analyzed in Section IV. Finally, conclusions are drawn in antenna pattern are used with different power levels for
Section V. different frequency carriers (a step of 5 dB was selected as used
in [4]).

1
 III. SIMULATOR ASPECTS losses are estimated in order to be possible to confer them the
best base station signal situation. After that, the SINR and the
The simulator developed for this work is based on the Ec/N0 parameters are computed in order to be able to define the
Nokias NPSW (Network Planning System for WCDMA), users state, i.e., connected, delayed or outage.
which can be found in [5]. This simulator is employed to The SINR is then used to compute the maximum available
perform radio network planning. As this simulator version did throughput, using the expressions in [6]. This parameter plays
not present the needed requirements for this work, i.e. the an important role in the simulations, since it is the downlink
HSDPA and HSUPA technologies, these have been developed, sensibility parameter. The same can be said in relation to the
as well as the HCS models and specifications. uplink parameter, Ec/N0.
The simulator can be denominated as a static simulator, In what concerns the possibility of an excess of service
since it does not apply any mobility model to the mobile requests to a given base station a reduction strategy is
stations (MS). Instead it produces consecutive snapshots to the implemented. The strategy adopted for this simulator has the
network varying the mobile stations position. main goal to decrease the throughput available for every user
The simulation environment consists in a 4000X4000 of the base station on which the strategy will be applied. The
meters area, where the base stations are positioned and the amount of throughput reduction can be defined by the
mobile stations are randomly spread throughout the area with a simulator user in the main settings window. The default value,
pre-defined type of service and frequency carrier. which was also the one considered for this work was a 10%
The COST 231 Walfisch-Ikegami was the propagation reduction of users throughput. This strategy is implemented
model used in a dense urban scenario. in both HSDPA and HSUPA.
In what concerns the base stations, its configuration is a Concerning the post processing, it is executed when the
fundamental step in the definition of the simulation scenario iteration process ends, providing several information data such
playing a fundamental role in the network planning process. as the number of served users by each used carrier, the users
The information that characterizes the base stations is average throughput or the service rates for each profile
stored in a text file, which is used in the main window of the mentioned above.
simulator. The information can be summed in four important
characteristics: location coordinates; antenna specifications (tri- IV. RESULTS
sectored or omnidirectional); frequency and transmitted power.
The mobile stations have several characteristics: The results attained with the simulator are intended to study
geographic position; height; transmitted power; frequency each scenario maximum throughput, aimed capacity and
carrier used; type of service used; maximum and minimum coverage, in terms of number of served users for both HSDPA
service throughput in DL and UL and QoS priority. and HSUPA technologies. Some parameters for HSDPA and
In order to provide a non-uniform mobile station HSUPA were defined and are presented in Table II. These
distribution in the scenario, the simulator performs a MS parameters are essential for the link budget estimation.
generation, allocating users containing the previously referred
specificities and with a random geographic position, TABLE II. VALUES USED IN HSDPA AND HSUPA SIMULATIONS
concerning the scenario limits. Afterwards the MSs are
distributed by the frequency carriers available in the scenario. Parameter HSDPA HSUPA
In addition to the mobile stations characteristics Number of HS-PDSCH codes 10 and 15 -
mentioned before, four users profiles were employed in order Mobile terminal transmission
to simulate the HCS models applicability for different end - 24
power [dBm]
users conditions (pedestrian, vehicular, indoor low loss and Mobile terminal antenna gain
0 0
indoor high loss). Details for these can be found in [7]. [dBi]
After allocating the users with a type of service and with Mobile terminal height [m] 1.8 1.8
one of the profiles mentioned above, they are distributed by the Node B antenna gain [dBi] 17 17
frequency carriers available in each scenario. In Table I the User losses [dB] 1 1
distribution of users per number of carriers is specified. Cable losses [dB] 3 3
Noise Figure [dB] 9 5
TABLE I. DISTRIBUTION OF USERS PERCENTAGE BY AVAILABLE Interference Margin [dB] 6 6
CARRIERS
Diversity Gain [dB] - 2
Percentage of Total users [%]
Scenario Carrier Carrier Carrier Carrier SHO Gain [dB] - 2
1 2 3 4 Signalling and power control R99: 25
-
percentage power [%] HSDPA: 10
1 65 35
Throughput Throughput
2 50 30 20 Reduction Strategy
Reduction Reduction
3 45 30 15 10
4 80 20 Simulations were performed for the four defined scenarios,
considering 2500 and 4000 initial users, in order to observe the
Following the users distribution some primary calculations systems capabilities of supporting distinct amounts of users. In
are performed including link losses, SINR and Ec/N0. The link Figures 4-(a) - (e) the maximum available throughput in DL are

2
presented, in this case for 15 HS-PDSCH codes, for scenario 1,
2, 3, 4 with omnidirectional antennas and 4 with tri-sectored
antennas, respectively.

(a) (b)

(a) (b)

(c) (d)

(c) (d)

(e)
Figure 3. Maximum available throughput in uplink

TABLE III. RESULTS ATTAINED FOR HSDPA AND HSUPA SIMULATIONS

(e)
Figure 2. Maximum available throughput in downlink HSDPA HSUPA
Total
Average Average
Number Served Total Served
From the graphs presented in Figure 4-(a) - (c) one can Scen Antenna Users Userss
of Initial Users Users Rate
ario Type Rb Rb
Users Rate [%]
conclude that the best throughput situation is achieved in areas [%]
[Mbps] [Mbps]
near the center of the cell, i.e. near the Node B. In Figures 4-(d) 1
2500 98,6 0,467 74 0,328
- (e) the available throughput is more spread given the 4000 90,5 0,491 73,2 0,310
introduction of micro-cells at the macro-cells edge, not 2
2500 94,8 0,399 78 0,306
Tri
allowing a sharp drop of the macro-cells signal when a user is 4000 92,6 0,457 77,9 0,287
sectored
2500 93,68 0,494 81 0,279
moving away from the center of that cell. 3
4000 94 0,502 83 0,256
In what concerns the HSUPA results, in Figure 5 one 2500 89,24 0,207 73,4 0,165
presents the maximum available throughput in uplink, for the 4 4000 87,48 0,232 70,3 0,152
same scenarios. It can be seen that the best throughput is Omni 2500 85,68 0,186 71,04 0,162
directional
achieved in locations near the base station. An increase of the 4000 84,8 0,197 70,1 0,149

available throughput is verified when the number of carriers

also increases, as presented in Figure 5-(a) - (c). From Figures with 98,6%. As for scenario 4, with cell-edge topology, the
5-(d) - (e) it is possible to observe that the introduction of base results have different dimensions than those attained with the
stations at the macro-cells edge avoided the throughput drop previous three scenarios, with a rate variation from 84,79% to
in locations far from the center of the macro-cells. 89,23%. Once again, a slight reduction of the served users
In Table III, the HSDPA and HSUPA most relevant results rate is detected when 4000 initial users are considered.
are presented. For HSDPA, in the first three scenarios, which Concerning the antenna type, it is possible to conclude that a
have a similar structure, the maximum served users rates micro-cell with a tri-sectored antenna achieves better results in
achieved are very close to each other. One can observe a small terms of served users rate, with rates up to 89% compared
decrease in the percentages when comparing the scenarios with with omnidirectional antennas with rates around 85% .
2500 and 4000 initial users, with the exception of scenario 3. The highest percentages of served users, presented in Table
The highest percentage is achieved in scenario 1, for 15 codes, III, were all achieved with resource to 15 HS-PDSCH codes.

3
The co-located cell topology, used in the scenarios 1-3 is the which can be explained by the increase in the number of users
one that presents better results, not only in the number of spread across the scenario that can afford a better Node B
served users but also with the users average throughput, in signal condition, since a higher number of initial users implies
comparison with the cell-edge topology adopted in scenario 4. an augment of the probability of being closer to the base
Comparing scenarios 1 and 4, given the fact that both employ stations. In what regards HSUPA, smaller service rates are
two distinct carriers in the system, one can conclude that attained, which can in some way be explained not only by the
scenario 1 achieves better results than scenario 4, not only in values achieved for Ec/N0 and the maximum available
terms of number of served users but also in average throughput are shown but also by the lower theoretical data
throughput per user. The increase in the available throughput rate value, which is 1,22 Mbps. From Figure 5, one can
is most noticed in the cell-edge topology than in the co-located observe that the best signal conditions are restricted to a tight
one, since in the former the signal provided by the micro-cell limited perimeter surrounding the base stations. This can
at the macro-cell edge is less influenced and overlaid than in a explain the reason of not achieving service rates as high as the
co-located cell topology, and given the fact that the macro-cell HSDPA. The most interesting service rates results are attained
transmits its power at a higher level compared to the micro- for scenario number 3, reaching up to 83% of served users.
cell, allowing a larger covered area. One can notice a constant increase in the service rate when the
Concerning HSUPA results, one can easily observe that number of carriers is also increasing, from scenarios 1 to 3.
there is a constant increase in the served users rate from However, there is also a lower offered throughput connection
scenarios 1 to 3, due to the carrier addition in each scenario between the MT and the Node B, as one can behold in Table
allowing the system to serve more users. In what regards the III.
uplink average data rate, this is mostly related to the MTs Concerning scenario number 4, with cell-edge topology,
locations. The systems load of number of users is another the served users rates achieved vaguely lower than the
constraint factor, so if a system is required a higher number of previous referred scenarios, with rates around 70% to 73% of
service and it is reaching its limits, than a reduction in the the total initial users. From Figure 5, it is possible to conclude
throughput is a primary action, causing an overall decrease of that the scenario number 4 provides a more efficient data rate
the systems average throughput. distribution, serving areas that the macro-cell may not support
As can be observed in Table III, the average throughput is the introduction of new co-located base stations with an
decreasing with the raise in the systems number of carriers. emitted power not high enough leading to difficulties of
This can be viewed as a trade-off between having more users communication between the Node Bs and the MTs, if the
with smaller data rate or having a less number of users but distance between them increases.
achieving higher data rate. As before mentioned the maximum
theoretical data rate in HSUPA is 1,22 Mbps, but given the REFERENCES
services throughput requirements, an average throughput of [1] Holma,H. and Toskala,A., HSDPA/HSUPA for UMTS
near 0,245 Mbps is required. Therefore, the data rate results High Speed Radio Access for Mobile Communications,
shown in Table III are quite acceptable. John Wiley and Sons, Chichester, UK, 2006.
V. CONCLUSIONS [2] Kim,S., Hong,D., Cho,J., Hierarchical Cell Deployment
for High Speed Data CDMA Systems, IEEE Wireless
The main objective of this work was to analyze the Communications and Networking Conference, 2002,
behavior of 4 distinct network topologies in UMTS/HSPA. Volume 1, pp 7-10. ISBN: 0-7803-7376-6 .
HSDPA results show that increasing in the number of carriers
in the system provides service rates from 90% up to 98%, for [3] Kim,C.-Y., The Evaluation of HCS Scenarios included in
the co-located carriers scenarios, being the highest value HSDPA, IEEE Vehicular Technology Conference, 2004,
Volume 7, pp 4772-4776. ISBN: 0-7803-8521-7 .
achieved with a two co-located carrier scenario.
A divergent conclusion is taken when the average delivered [4] Ribeiro,D., Increasing the CDMA-based Cell Capacity
throughput is considered. The overall throughput values for Urban Area with 3 and 4 Cell Partition topologies,
achieved are below 500 kbps, which is quite lower than the MSc Dissertation, Instituto Superior Tcnico, Lisboa,
theoretical 14 Mbps. This corroborates the idea of a necessary Portugal, June 2007.
closeness between the HSDPA base station and the end user to
[5] Laiho,J., Wacker,A., Novosad,T., Radio Network
afford data rates higher than 1 Mbps. Planning and Optimisation for UMTS, John Wiley &
With the exception of the third scenario, in every scenario a Sons, Sussex, England, 2002.
small decrease of the served users rate is verified, when 4000
users are initially allocated instead of 2500, however an [6] Lopes,J., Performance of UMTS/HSDPA/HSUPA at the
increase of average throughput is evidenced in each scenario. Cellular Level, MSc Dissertation, Instituto Superior
Tcnico, Lisboa, Portugal, February 2008
For scenario 4 with omnidirectional and tri-sectored
antennas an increase of the average throughput is verified,
[7] Rebelo,T., Hierarchical Cell Structure in UMTS/HSPA
Systemsl, MSc Dissertation, Instituto Superior Tcnico,
Lisboa, Portugal, February 2008