PERFORMANCE OF ENHANCED-UMTS HSDPA USING

TRANSMIT DIVERSITY AND POWER CONTROL SCHEMES

João Carlos Silva, Nuno Souto, António Rodrigues

Instituto Superior Técnico/IT, Torre Norte 11-11, Av. Rovisco Pais 1, 1049-001 Lisboa, Portugal,

Américo Correia, Francisco Cercas

ADETTI/IT, Torre Norte 11-08, Av. Rovisco Pais 1, 1049-001 Lisboa, Portugal,

Nuno Cota

Instituto Superior de Engenharia de Lisboa,Rua Conselheiro Emídio Navarro-1, 1950-062 Portugal

Keywords: UMTS, HSDPA, Power Control, STD, STTD, Simulation Results

Abstract: This paper addresses the performance of the downlink High Speed Data Packet Access (HSDPA) in QPSK

mode. Transmit Diversity (TD) enhancement schemes such as Space Time Transmit Diversity (STTD) and

Selective Transmit Diversity (STD), alongside a Power Control (PC) scheme, are covered to improve the

system capacity. To evaluate the performance and the advantages of all the schemes under different

conditions, several combinations of these were simulated in the AWGN, Indoor A and Pedestrian A

channels. For the best combination, a gain of 11.5dB can be achieved, for a BLER of 1%.

1 INTRODUCTION

The 3G cellular system, known as UMTS, is in the

standard phase since 1999 and has the maximum

capacity of 2 Mbps. Although UMTS is being

licensed throughout Europe and implementation

involves huge investments, it is recognized that, at

the eve of implementation and deployment of the

system, still many aspects need to be investigated.

This is mainly related to the fact that there is no

previous experience on mobile cellular systems

operating in multi-service and multi-rate.

Furthermore, some applications require much

higher bit rates than those currently offered in

UMTS, like real-time audio/video on demand, web

browsing, etc.

This paper was elaborated as a result of the

participation in the B-BONE (Broadcasting and

Multicasting over Enhanced UMTS Mobile

Broadband Networks) project. The B-BONE project

aims to provide a fundamental contribution to the

evaluation and development of Enhanced UMTS

networks.

Future Enhanced-UMTS Systems are mostly

based on the current UMTS Systems but also add

new ‘Space Time Processing’ techniques, ‘namely

Adaptive Antennas, MIMO Systems, Space-Time

Coding and spectrally efficient coding/

modulation/ spreading schemes, to increase the

capacity of the access network up to 10 Mbps.

Recently, the HSDPA mode was introduced in

the standards, and the maximum downlink bit-rate

was increased significantly. Two possible

modulations were defined: QPSK and 16 QAM. The

latter mode delivers higher bit rate than the former

and is intended for good channel conditions. This

work addresses solely the usage of the QPSK

reference mode, though the main conclusions are

valid for both modes. The main parameters and

physical channel segmentation for the HSDPA-

QPSK mode are summarized in Figures 1 and 2,

respectively. Note that the turbo code has a rate of

Carlos Silva J., Souto N., Rodrigues A., Correia A., Cercas F. and Cota N. (2004).

PERFORMANCE OF ENHANCED-UMTS HSDPA USING TRANSMIT DIVERSITY AND POWER CONTROL SCHEMES.

In Proceedings of the First International Conference on E-Business and Telecommunication Networks, pages 46-51

DOI: 10.5220/0001383400460051

 SciTePress

1/3, and 50% puncturing is applied to its output.

Five physical channels are used, which are spreaded

with a Spreading Factor (SF) of 16. Sub-frames of 2

ms are used, corresponding to 3 time slots of the

normal UMTS mode. A link level simulator was

built, from the transport block level, turbo coding,

transmission of the encoded chips, emulation of the

channel, RAKE receiver, turbo decoding, and all

necessary operations up to the received transport

block. All modules specified in the standards were

built (3GPP, 25.211-213), with the addition of the

STD, STTD (3GPP 25.211) and PC (3GPP 25.101)

modules. Although HSDPA is based on Adaptive

Coding and Modulation (AMC), PC was considered

in order to further improve the performance of the

system, although the AMC itself is not considered.

Results were drawn with and without the

enhancement schemes, in order to evaluate the

performance gains.

This paper is structured as follows: in section 2,

the STD, STTD and PC are summarized, and section

3 analyzes the obtained results. Conclusions are

drawn in section 4.

2 ENHANCEMENT SCHEMES

Both STD and STTD schemes are based on the

usage of 2 antennas, though more antennas could be

employed. The STTD scheme consists on having the

message bits transmitted by two antennas (Figure 3),

encoded as shown in Figure 4. The hand-set receives

both transmissions from the antennas, and performs

the STTD decoding, as explained in Figure 6 and

Equations (1) and (2). From Figure 6, the signal after

the RAKE receiver is:

()

***

1001100

22 2

**'

001100

***

2001101

222

** '

0011 10

***

1100110

22 2

**'

100101

()

(2 )

rk T S c h S c h n c h

Sc h Schh n

rk T S c h S c h n c h

Schh Sc h n

rk T S c h S c h n c h

Sc h S c hh n

rk T S c

=⋅⋅−⋅⋅+×⋅=

=⋅ ⋅ −⋅⋅⋅+

=⋅⋅−⋅⋅+×⋅=

=⋅ ⋅⋅−⋅ ⋅ +

=⋅⋅+⋅⋅+×⋅=

=⋅ ⋅ +⋅ ⋅⋅+

=⋅⋅

()

***

00 11 1

222

** '

1010 11

hSchn ch

Sc hh S c h n

+⋅⋅+ ×⋅=

=⋅ ⋅⋅+ ⋅ ⋅ +

(1)

The final STTD-decoding combination yields:

()

222

*''

12 0 01

222

*''

12101

() (2)

(2 ) ( )

SrkTrkT Sc h h n

=+ =⋅⋅++

=−=⋅⋅++

(2)

The STD algorithm is quite simple; the mobile

chooses the best antenna for transmission, and uses

such antenna until the other antenna starts yielding

better estimated values for the channel state. The

equivalent received power for the STD scheme is

thus

(

)

max ,

STD

Phh=

instead of

(

)

STTD

. Note that, in the STTD

scheme, both antennas transmit the original

message, though the second antenna uses a modified

version of the message. There is no feedback for

these kind of transmit diversity. In the STD scheme,

only one antenna transmits the message, whereas the

other antenna transmits pilot bits to assess the

channel quality, with the mobile transmitting back

control bits in order to choose the best antenna

(closed loop scheme – Figure 5). The antenna that

transmits the message may vary, according to the

channel state associated to each antenna.

At the receiver side, a soft decision Rake receiver

was built, and a turbo decoder using the Maximum

A Posteriori (MAP) algorithm (Bahl, 1974), was

employed. A maximum of 8 iterations were used,

with iteration stop criteria linked to the Cyclic

Redundancy Check (CRC) result. The simulated PC

mechanism was the standardized downlink type 1,

with a step size of 1dB (which is the value

mandatory for the UTRAN to support, though other

values can also be used) and a delay of 1 Time Slot

(TS), due to being the PC scheme that yields best

results (Silva, 2003). A dynamic range of ±15dB

was employed, to simulate saturation. Figure 7

shows a simple diagram illustrating the downlink

Power Control scheme. At a time resolution equal to

1 TS (0,667ms), the power control algorithm

instructs the hand-set to increase/ decrease its power

level by 1dB. There is a delay of 1TS for the

simulated case, since the corrected power level for

the current TS takes into account the power level of

the previous TS.

The Tapped Delay Line (TDL) model specified

for UMTS (ETSI, 1998), (3GPP 25.943) was

simulated, and a speed of 3 km/h was used for the

Indoor A channel, whereas a velocity of 10 km/h

was used for the Pedestrian A channel. Only fast

fading was considered; slow fading/ shadowing was

assumed to be completely compensated. Perfect

channel estimation was assumed. For the STD

simulations, a delay equivalent to 1 TS was used, in

order to model a realistic scenario. When the STD

was combined with PC, a smart mechanism was

activated which reduced the STD delay to 0, thus

simulating a STD capable of predicting what the

power control is about to do for the next TS, and to

anticipate what the next channel value will be.

Though the latter assumption is not very realistic,

PERFORMANCE OF ENHANCED-UMTS HSDPA USING TRANSMIT DIVERSITY AND POWER CONTROL

SCHEMES

the practical deviation of results that arises from it is

negligible for channels with low speed values, such

as those simulated.

3 NUMERICAL RESULTS

Bit Error Rates (BER) and Block Error Rates

(BLER) were obtained via Monte Carlo simulation

of the different scenarios with the link level

simulator using a total number of iterations

sufficient enough to ensure convergence.

The simulated results are depicted in Figures 8-

19. After analyzing the results, the performance gain

by applying transmit diversity can be quantified. We

note that STD has always a better performance than

STTD, as would be expected, due to always

choosing the best antenna for transmission.

Results were obtained for using PC Type 1 with

1 TS resolution simulated alone and also in

combination with STTD and STD. From the results,

it can be seen that the usage of PC combined with

either STD or STTD is clearly the solution with

greatest performance increase over the normal

QPSK simulations.

It can also be seen that both Indoor A and

Pedestrian A channels are similar in performance,

though the Indoor A channel is slightly better

without power control, and the Pedestrian A channel

is better with schemes involving PC. This result is

simply explained by the fact that PC schemes

usually work better in channels that are close to a

single Rayleigh path, such as the Pedestrian A

channel. This is due to the minimization of

intersymbolic interference, effectively aiding the PC

algorithm, since all the power adjustments are

practically for the main path of the message.

Figures 20 and 21 show the Eb/N0 values for the

different schemes, operating under considered

normal operation points (BLER of 0.1 for non-real

time applications and of 0.01 for real-time

applications).

4 CONCLUSIONS

By direct inspection of the results, it was shown that

the HSDPA mode alone, without the enhancement

schemes such as transmit diversity and power

control, performs very poorly. Thus, in order to

achieve high bit rates, and use the UMTS to its full

potential, enhancement schemes, such as STD,

STTD and PC need to be employed, in conjunction

with each other.

In order to further close the gap to the AWGN

results, more complex schemes can be employed.

This can be done by using Adaptive Antennas,

MIMO Systems, higher order Space-Time Coding

techniques (the work in this paper considered an

order 2 system, using 2 antennas for the STD and

STTD scheme), Power Control with less delay and

spectrally efficient coding/ modulation/ spreading

schemes, namely the use of Code-Spread (CS)

CDMA schemes (Souto, 2004). The usage of AMC

effectively combined with the proposed schemes

could also further improve the system’s

performance.

REFERENCES

3GPP, 25.212-370, ‘Multiplexing and Channel Coding

(FDD)’.

3GPP, 25.213-360, ‘Spreading and Modulation (FDD)’.

L.R. Bahl, J. Cocke, F. Jelinek and J. Raviv, “Optimal

decoding of linear codes for minimizing symbol error

rate” IEEE Trans. Inform. Theory, vol. IT-20, pp.284-

287, Mar. 1974.

3GPP, 25.211-370, ‘Physical Channels and mapping of

Transport Channels onto Physical Channels’.

3GPP TS 25.101 V5.4.0, ‘Power Control’

J.C. Silva, N. Souto, A. Correia, F. Cercas, A. Rodrigues,

“Characterization of Power Control for UMTS”,

WPMC 2003, Yokosuka, Japan

ETSI, Selection procedures for the choice of radio

transmission technologies of UMTS, TR 101 112

v3.2.0, ETSI, Sophia Antipolis, France, 1998

3GPP, Deployment aspects, 3GPP TR 25.943 v5.1.0,

Sophia Antipolis, France, 2002.

N. Souto, J.C. Silva, F. Cercas ‘Usage of Turbo Super-

Orthogonal Codes for an enhanced UMTS CS-CDMA

Uplink transmission’, to be presented in VTC’04

Spring, Milano, Italy, 2004.

ICETE 2004 - WIRELESS COMMUNICATION SYSTEMS AND NETWORKS

BER vs Eb/N0 - Indoor A

1,E-05

1,E-04

1,E-03

1,E-02

1,E-01

1,E+00

0 2 4 6 8 10 12 14 16

Eb/N0 (dB)

BER

QPSK-STTD

QPSK PC

QPSK STTD+PC

BER vs Eb/N0 - Indoor A

1,E-05

1,E-04

1,E-03

1,E-02

1,E-01

1,E+00

024681012141618

Eb/ N0 ( dB)

BER

QPSK-STTD

QPSK STD

QPSK STTD+PC

QPSK STD+PC

Figure 1: HSDPA main parameters Figure 2: Physical channel segmentation for HSDPA

Figure 3: Schematics of STTD operation Figure 4: Schematics of STTD encoder

Figure 5 – Schematics of STD algorithm

Figure 6: STTD reception scheme Figure 7: Schematics of downlink Power Control algorithm

Figure 8: HSDPA BER results for Indoor A Figure 9: HSDPA BER results for Indoor A

(PC, STTD and PC+STTD) (Comparison of STD and STTD, with and without PC)

PERFORMANCE OF ENHANCED-UMTS HSDPA USING TRANSMIT DIVERSITY AND POWER CONTROL

SCHEMES

BER vs Eb/N0 - Indoor A

1, E - 0 6

1, E - 0 5

1, E - 0 4

1, E - 0 3

1, E - 0 2

1, E - 0 1

1, E + 0 0

0 5 10 15 2 0 2 5

Eb/ N0 (dB)

BER

QPSK

QPSK AWGN

QPSK STD

QPSK STD+PC

BER vs Eb/N0 - Pedestrian A

1,E-06

1,E-05

1,E-04

1,E-03

1,E-02

1,E-01

1,E+00

024681012141618

Eb/ N0 ( dB)

BER

QPSK PC+STTD

QPSK PC

QPSK STTD

BLER vs Eb/N0 - Indoor A

1, E - 0 4

1, E - 0 3

1, E - 0 2

1, E - 0 1

1, E + 0 0

0 5 10 15 2 0 2 5

Eb/ N0 (dB)

BLER

QPSK

QPSK AWGN

QPSK STD

QPSK STD+PC

BLER vs Eb/N0 - Indoor A

1, E - 0 4

1, E - 0 3

1, E - 0 2

1, E - 0 1

1, E + 0 0

0 5 10 15 2 0

Eb/ N0 (dB)

BLER

QPSK-STTD

QPSK STD

QPSK STTD+PC

QPSK STD+PC

Figure 10: HSDPA BER results for Indoor A Figure 11: HSDPA BER results for PedestrianA

(Normal, STD, PC+STD and AWGN) (PC, STTD and PC+STTD)

BER vs Eb/N0 - Pedestrian A

1,E-06

1,E-05

1,E-04

1,E-03

1,E-02

1,E-01

1,E+00

0 5 10 15 20 25

Eb/N0 (dB)

BER

QPSK

QPSK AWGN

QPSK STD

QPSK STD+PC

Figure 12: HSDPA BER results for PedestrianA Figure 13: HSDPA BLER results for IndoorA

(Normal, STD, PC+STD and AWGN) (Normal, STD, PC+STD and AWGN)

BER vs Eb/N0 - Pedestrian A

1,E-06

1,E-05

1,E-04

1,E-03

1,E-02

1,E-01

1,E+00

0 2 4 6 8 10 12 14 16 18

Eb/N0 (dB)

BER

QPSK STTD

QPSK STD

QPSK PC+STTD

QPSK STD+PC

Figure 14: HSDPA BER results for PedestrianA Figure 15: HSDPA BLER results for IndoorA

(Comparison of STD and STTD, with and without PC) (Comparison of STD and STTD, with and without PC)

ICETE 2004 - WIRELESS COMMUNICATION SYSTEMS AND NETWORKS

BLER vs Eb/N0 - Pedestrian A

1, E - 0 4

1, E - 0 3

1, E - 0 2

1, E - 0 1

1, E + 0 0

0 5 10 15 2 0

Eb/ N0 (dB)

BLER

QPSK-PC

QPSK-PC+STTD

QPSK-STTD

BLER vs Eb/N0 - Indoor A

1, E-04

1, E-03

1, E-02

1, E-01

1, E+00

02 46810121416

Eb/ N0 ( dB)

BLE R

QPSK-STTD

QPSK PC

QPSK STTD+PC

Figure 16: HSDPA BLER results for IndoorA Figure 17: HSDPA BLER results for PedestrianA

(PC, STTD and PC+STTD) (PC, STTD and PC+STTD)

BLER vs Eb/N0 - Pedestrian A

1,E-04

1,E-03

1,E-02

1,E-01

1,E+00

0 5 10 15 20 25

Eb/N0 (dB)

BLER

QPSK

QPSK AWGN

QPSK STD

QPSK-PC+STD

BLER vs Eb/N0 - Pedestrian A

1,E-04

1,E-03

1,E-02

1,E-01

1,E+00

0 2 4 6 8 1012141618

Eb/N0 (dB)

BLER

QPSK-STTD

QPSK STD

QPSK-PC+STTD

QPSK-PC+STD

Figure 18: HSDPA BLER results for PedestrianA Figure 19: HSDPA BLER results for PedestrianA

(Normal, STD, PC+STD and AWGN) (STD and STTD, with and without PC)

Figure 20: Results for BLER of 0,1 Figure 21: Results for BLER of 0,01

PERFORMANCE OF ENHANCED-UMTS HSDPA USING TRANSMIT DIVERSITY AND POWER CONTROL

SCHEMES