RESEARCH Open Access An energy efficient semi-static power

RESEARCH Open Access

An energy efficient semi-static power control andlink adaptation scheme in UMTS HSDPAYi Huang, Jie Xu and Ling Qiu*

Abstract

High speed downlink packet access (HSDPA) has been successfully applied in commercial systems and improvesuser experience significantly. However, it incurs substantial energy consumption. In this article, we address thisissue by proposing a novel energy efficient semi-static power control and link adaptation scheme in HSDPA.Through estimating the EE under different modulation and coding schemes (MCSs) and corresponding transmitpower, the proposed scheme can determine the most energy efficient MCS level and transmit power at the NodeB. And then the Node B configures the optimal MCS level and transmit power. In order to decrease the signalingoverhead caused by the configuration, a dual trigger mechanism is employed. After that, we extend the proposedscheme to the multiple input multiple output (MIMO) scenarios. Simulation results confirm the significant EEimprovement of our proposed scheme. Finally, we give a discussion on the potential EE gain and challenge of theenergy efficient mode switching between single input multiple output (SIMO) and MIMO configuration in HSDPA.

Keywords: HSDPA, energy efficiency, power control, link adaptation

1 IntroductionHow to acquire higher throughput with lower powerconsumption has become an important challenge for thefuture wireless communication systems [1]. “Moore’sLaw” renders the use of ever more powerful informationand communications technology (ICT) systems for themass market. In order to transport this exponentiallyrising amount of available data to the user in an accep-table time, the transmission rate in cellular networkrises at the speed of nearly 10 times every 5 years,meanwhile the energy consumption doubles every5 years, as illustrated in [2].High speed downlink packet access (HSDPA) has been

successfully applied commercially, which brings highspectral efficiency (SE) and enhances user experience.According to [3], HSDPA has introduced a new down-link physical channel called high speed physical down-link shared channel (HS-PDSCH), and some newfeatures such as adaptive modulation and coding scheme(AMC), hybrid automatic repeat request (HARQ), fastscheduling and multiple input multiple output (MIMO).

Thus it improves the downlink peak data rate and sys-tem throughput greatly. For the MIMO technology inHSDPA, the so-called dual stream transmit adaptiveantennas (D-TxAA) is applied, in which the Node Bwould select single stream mode or dual stream modebased on the channel conditions.To the best of the authors’ knowledge, most of the pre-

vious research studies focused on spectral efficientschemes in UMTS HSDPA and only a few literaturesfocused on the network energy savings [4,5]. In [4],Micallef et al. proposed to switch off a second carrier inDual-Cell HSDPA to save energy through exploiting thenetwork traffic variations. And in [5], Marsan et al. inves-tigated the possibility of cutting down the energy con-sumption of the wireless networks by reducing thenumber of active cells when the traffic load is low. Theseworks mainly considered energy savings from a networkpoint of view. However, there is no literature focusing onthe link level energy efficient schemes in HSDPA, whichis also an important aspect in green communicationresearch.Energy efficiency (EE) is always defined as the transmis-

sion rate divided by the total power consumption, whichrepresents the number of information bits transmittedover unit energy consumption measured in bits/Joule.

* Correspondence: [email protected] Communication Network & Spread Spectrum Laboratory (PCN&SS),University of Science and Technology of China (USTC), Hefei, Anhui 230027,China

Huang et al. EURASIP Journal on Wireless Communications and Networking 2012, 2012:87http://jwcn.eurasipjournals.com/content/2012/1/87

© 2012 Huang et al; licensee Springer. This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium,provided the original work is properly cited.

mailto:[email protected]

http://creativecommons.org/licenses/by/2.0

In the previous studies considering EE from a link levelperspective [6-10], EE maximization problems are formu-lated and solved based on Shannon capacity, in which theimpact of constant circuit power is involved. It is demon-strated that joint power control and link adaptation is aneffective method to improve the EE. However, practicalmodulation and channel coding schemes are not consid-ered in these works and the users’ quality of service (QoS)constraints are not taken into account either. Moreover, asthe fast power control is not available in HS-PDSCH dueto the functionality of AMC and HARQ, it is hard to applyjoint power control and link adaptation in the HSDPA sys-tem directly.In this article, we will discuss the potential link level

energy saving in HSDPA. First, a power model includingdynamic circuit power related with antenna number istaken into account. Based on this model, we propose apractical semi-static joint power control and link adapta-tion method to improve EE, while guaranteeing the users’transmission rate constraints. As fast power control is nolonger supported, we propose a dual trigger mechanism toperform the method semi-statically. After that, we extendthe scheme to the MIMO HSDPA systems. Simulationresults confirm the significant EE improvement of ourproposed method. Finally, we give a discussion on thepotential EE gain and challenges of the energy efficientmode switching between single input multiple output(SIMO) and MIMO configuration.The rest of the article is organized as follows. Section 2

introduces the preliminaries. Section 3 proposes theenergy efficient power control and link adaptation schemein the single input single output (SISO) HSDPA systems.The extension of the scheme to the MIMO HSDPA sys-tems is presented in Section 4. Simulation results and dis-cussion are given in Section 5, and finally Section 6concludes this article.

2 PreliminariesIn this section, preliminaries are provided. The systemmodel and power model are introduced at first. The theore-tic SE-EE tradeoff is then provided to help the description.

2.1 System modelWe consider the system with a single Node B and a sin-gle user in this article, but note that our work can beextended to the multi-user scenario easily. We assumethat the Node B has a maximum transmit power con-straint Pmax and the user has a minimum modulationand coding scheme (MCS) constraint θmin which can beviewed as the QoS requirements.2.1.1 Link adaptation scheme and D-TxAA functionalityin HSDPAThe traditional link adaptation of the HSDPA systems isillustrated as follows. First, the Node B determines the

transmit power of HS-PDSCH. Once the transmit poweris determined, it cannot be changed frequently, due tothe existence of AMC and HARQ. The user measuresthe channel quality between the Node B and itself andfeeds back a channel quality indication (CQI) to theNode B. The feedback CQI corresponds to a MCS levelwhich is always chosen to maximize the transmissionrate under a certain bit error rate (BER). Then the NodeB delivers data to the user with the MCS level. In thisway, the transmission parameters can be adjustedaccording to current channel conditions and thus highthroughput can be provided.D-TxAA is selected as the MIMO scheme for HSDPA in

3GPP specification Release 7 [11]. Two antennas at theNode B and the user are supported. Specifically, the NodeB sends buffered data through either one or two indepen-dent data streams at the physical layer. At first, the userdetermines the preferred CQI for the single stream modeand the preferred pair of CQIs for the dual stream mode.After comparing the transmission rates of the two modes,the user can choose the better mode and correspondingCQI(s) and then feed them back to the Node B. Thus, theNode B can decide the mode and corresponding MCSlevel(s).In addition to CQI feedback, the user also reports pre-

coding control indicator (PCI) index which indicates theoptimal precoding weights {w1, w2} for the primary stream,based on which precoding weights {w3, w4} for the secondstream can be calculated. The precoding weights aredefined as follows [11]:

w1 = w3 = 1/√2,

w4 = −w2,

w2 =[1 + j2

,1 − j2

,−1 + j

2,

−1 − j2

].

(1)

2.2 Power consumption modelPower consumption model here is based on [10] in orderto capture the effect of transmit antenna number. Denotethe number of active transmit antennas as Ma and trans-mit power as P. The total power consumption of Node Bis divided into three parts. The first part is the powerconversion (PC) power

PPC =P

η, (2)

accounting for the power consumption in the poweramplifier and related feeder loss, in which h is the PCefficiency. The second part is the dynamic circuit powerwhich corresponds to antenna number Ma and can begiven by:


Page 2 of 13

PDyn = MaPcir, (3)

representing circuit power consumption for radio fre-quency (RF) and signal processing. The third part is thestatic power PSta related to cooling loss, battery backupand power supply loss, which is independent of Ma andPPC. The total power consumption can be modeled as

Ptotal = PPC + PDyn + PSta. (4)

2.3 SE and EE trade-offBefore introducing our proposal, we need to have a discus-sion about the theoretical basis of the energy efficientpower control and link adaptation scheme. According tothe Shannon capacity, SE and EE of a SISO additive whiteGaussian noise (AWGN) channel can be expressed as

μ = log2

(1 +

PN0W

)(5)

and

ξ = Wlog2

(1 +

PN0W

)/Ptotal (6)

respectively, where W and N0 represent system band-width and the noise density, respectively.It is obvious from (5) that the transmit power P is expo-

nentially increasing as a function of the SE with theassumption of constant bandwidth and noise power. Inother words, higher SE incurs significant increase ofenergy consumption. In fact, EE is monotonically decreas-ing with SE if only the transmit power is considered [12].Thus in order to improve EE, Node B should reduce thetransmit power. However, the existence of practical PDynand PSta breaks the monotonic relation between SE and

EE, so balancing the PPC, PDyn and PSta is also importantto increase EE. Figure 1 shows the EE-power and SE-power relations in an AWGN channel with the theoreticalShannon capacity formula. As indicated in Figure 1, thereexists a globally optimal transmit power for EE. Moreover,based on the Shannon capacity, we can obtain the explicitclose-form solution of the globally optimal EE and optimaltransmit power, and some examples in MIMO systemscan be found in [10].However, one may argue that whether the relation

between EE and SE still satisfies in the HSDPA systemswhen practical AMC and HARQ are taken into account.Fortunately, we confirm this principle through the HSDPAlink level simulation and the result with SISO channelsbased on TABLE G is shown in Figure 2. The MIMO sys-tems with D-TxAA have the similar relations, which isshown later in this article. Although this trend is still ful-filled, the challenge in the HSDPA systems is that theexplicit close-form solution to obtain the optimal transmitpower and corresponding MCS level is no longer availablewhen practical AMC and HARQ are applied here. To meetthis challenge, we will solve this problem through a novelEE estimation mechanism in the rest of this article. Besides,the data rate constraints are considered due to the users’QoS requirements in practice. According to the con-straints, we should find the feasible transmit power regionfirst, and then determine the transmit power with con-strained optimal EE based on the feasible region. Moredetails will be given in the following section.

3 Energy efficient power control and linkadaptation scheme in SISO systemsA semi-static power control and link adaptation methodis proposed in this section to improve the EE whileguaranteeing the MCS level constraint. Different from

0 10 20 30 400

2

4

6

8

10

12

14SE calculated with shannon capacity formula

Transmit Power(dBm)

SE

(bit/

s/H

z)

0 10 20 30 400

0.5

1

1.5

2

2.5

3

3.5x 10

6 EE calculated with shannon capacity formula

Transmit Power(dBm)

EE

(bit/

Joul

e)

Figure 1 SE and EE calculated using shannon capacity formula.


Page 3 of 13

the previous energy efficient schemes which are onlyapplicable for the Shannon capacity, our proposedscheme determines the energy efficient transmit powerand MCS level according to a practical EE estimationmechanism, which is based on CQI feedback. Further-more, we propose a semi-static dual trigger to controlthe transmit power and MCS level configuration, whichis practical in the HSDPA systems.Figure 3 shows the operational flowchart of the pro-

posed power control and link adaptation procedure at theNode B. As long as CQI and acknowledgement/negativeacknowledgement(ACK/NACK) information are receivedby the Node B, Node B can estimate the EE and therequired transmit power for each MCS level based on theestimation mechanism. Then Node B can determine theMCS level and transmit power with maximum EE. Afterthat, the Node B will determine whether they need to beconfigured immediately or not, where a semi-static dualtrigger mechanism is employed. If it is triggered, thederived optimal transmit power and corresponding opti-mal MCS level will be reconfigured. In this way, thescheme is realized in a semi-static manner. There are twobenefits here. For one thing, the semi-static feature makesthe scheme practical in HSDPA which does not supportinner loop power control. For another, the cost of signal-ing can be reduced significantly through controlling thepower reconfiguration cycle length adaptively.In the following subsections, we will introduce the

scheme in details.

3.1 EE estimation and optimal transmit powerdeterminationWe propose the addition of an EE estimation mechan-ism to the traditional link adaptation operation,whereby it employs the MCS table to estimate the EEand required transmit power for different MCS levels

based on CQI feedbacks, and then determines the EEoptimal transmit power and MCS level. The MCStable here is defined as the mapping relationshipbetween HS-PDSCH received signal to interferenceand noise ratio (SINR) threshold and the correspond-ing feedback CQI index, based on the initial BER tar-get Γtar. Each CQI index corresponds to a dedicatedMCS level in HSDPA. An example of TABLE G [3] isshown in Figure 4.At first, we need to estimate the transmit power

required for different MCS levels. According to [13], theSINR of HS-PDSCH is denoted as

ρ(PHS) =SF · PHSg

(1 − α)Ior + Ioc +N0W, (7)

where SF, PHS, g, a, Ior, and Ioc denote the spreadingfactor, HS-PDSCH power, the instantaneous path gain,the channel orthogonality factor, the total received powerfrom the serving cell and the inter-cell interference,respectively. As the link level simulation has captured theeffect of the inter-code interference, according to (7),received SINR is proportional to transmit power PHS

assuming that the interference is constant. By taking thelogarithm on both sides of (7), we can find that the differ-ence between two transmit power P1 and P2 is equal tothe difference between the two SINR r(P1) and r(P2)derived from them:

P1(dBm) − P2(dBm) = ρ(P1)(dB) − ρ(P2)(dB), (8)

where transmit power is measured in dBm and SINRis measured in dB.After replacing the actual SINRs in (8) By the SINR

thresholds in the MCS table, we can utilize the equationto estimate the transmit power required for the MCSlevels. In other words, we propose to approximate the

0 10 20 30 400.5

1

1.5

2

2.5

3

3.5

4SE acquired from HSDPA link level simulation

Transmit Power(dBm)

SE

(bit/

s/H

z)

0 10 20 30 400.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8x 10

6 EE acquired from HSDPA link level simulation

Transmit Power(dBm)

EE

(bit/

Joul

e)

Figure 2 SE and EE acquired from HSDPA link level simulation.


Page 4 of 13

difference between the transmit power required for twoMCS levels as the difference between the two’s SINRthresholds. For example, assume that the current trans-mit power is P and the feedback CQI index is i. For anarbitrary CQI index denoted by j, the correspondingSINR threshold is denoted as bj and the MCS leveldenoted as θj. We can estimate the transmit power Pjrequired for MCS level θj as follows:

Pj = P + βj − βi + δ. (9)

The offset δ here is to deal with the impact of chan-nel variations which can be determined based on thefeedback ACK/NACK information from the user side.

In the simplest case, δ can be set to zero and (9) canbe rewritten as:

Pj = P + βj − βi. (10)

Note that transmit power is measured in dBm andSINR threshold is measured in dB in (9) and (10).One may argue that the adjustment would cause the

variation of BER, and then affect the average number ofthe retransmissions, which may cause the energy wast-ing. This is not the case. The same BER can be guaran-teed for the current and adjusted power level and MCSlevel, which can be explained as follows. Note that theMCS table at both the BS and the user is based on a

Figure 3 Flowchart of the proposed energy efficient power control procedure.


Page 5 of 13

fixed BER target. Therefore, it is obvious that the cur-rent power level and feedback CQI can guarantee theBER. During the adjustment, to make sure the sameBER can be guaranteed, the transmit power and theMCS level are jointly adjusted. That is to say, when thetransmit power is decreased, the corresponding MCSlevel should also be decreased. As the same BER is guar-anteed in this way, the same retransmission probabilitycan also be guaranteed, and the average number of theretransmissions will not be affected. In a word, ourscheme would work well without affecting the mechan-ism of the retransmission, which is practical in realsystems.Then the estimation of EE for the MCS level θj is

given by:

ξj =τj

ts · ( pjη+ PDyn + PSta)

, (11)

where τj represents the transport block size of theMCS level θj , and ts is equal to two milliseconds andrepresents the duration of one TTI for HSDPA. Thenwe compare the estimated EE for each MCS level, deter-mine the optimal CQI index j* by

j∗ = argmaxj

ξj. (12)

The corresponding MCS level is denoted as θj* andthe required transmit power denoted as Pj∗ .As the minimum MCS level of the user is θmin and

the maximum transmit power of the Node B is Pmax,the constrained optimal MCS level and the optimaltransmit power can be given by:

θopt = min(max(θmin, θj∗), θmax),

Popt = min(max(Pmin, Pj∗), Pmax).(13)

The same estimation mechanism above can beemployed to determine the corresponding minimumtransmit power Pmin and the corresponding maximumMCS level θmax. Correspondingly, the estimated EE forthe optimal MCS level and transmit power is denoted asξopt.In our proposed algorithm, only the feedback CQI and

ACK/NACK information are necessary for Node B to dothe EE estimation and energy efficient powerdetermination.

0 5 10 15 20 25 30−5

0

5

10

15

20

25

30

35

CQI index

SIN

R T

hres

hold

(dB

)MCS table G required IBLER=0.1

Simulated CQI Table G

Figure 4 MCS Table for User Category G.


Page 6 of 13

3.2 Semi-static power reconfiguration triggerHowever, the power configuration cannot be performedinstantaneously due to the following two reasons. Forone thing, the support for fast AMC and HARQ func-tionality in HSDPA does not allow the transmit powerchange frequently. For another, in order to guarantee theaccuracy of the CQI measurement and user demodula-tion especially for high order modulation, Node B shouldinform the user of the transmit power modificationsthrough the signalling called measurement power o set(MPO) in radio resource control (RRC) layer when thetransmit power is reconfigured. If the configuration per-forms frequently, the signaling overhead is significant.Therefore, we propose a semi-static trigger mechanismto control the procedure.Assume that the EE derived from the last transmission

is ξ , define relative EE difference D as follows:

D =ξopt − ξ

ξopt. (14)

In our proposed scheme, the minimum trigger intervalis set to be gprohibit, and the maximum trigger interval tobe gperiodic which satisfies gperiodic ≫ gprohibit. A timer isused to count the time from the last power configura-tion and the timing is denoted as t.First, if

{D ≥

t > γprohibit(15)

both are satisfied, the proposed energy efficient powerconfiguration and corresponding MCS reselection pro-cess is triggered. This event trigger can guarantee EEgain and also avoid frequent power configuration. Onthe other hand, if

t > γperiodic (16)

is satisfied, the power configuration process must betriggered regardless of the value of D. This periodical trig-ger ensures that the scheme is always active and guranteesthe EE gain. If the power configuration is triggered, thetimer must be reset to zero. The whole trigger mechanismabove is robust as its parameters can be configured adap-tively according to actual systems. It can be implementedpractically in HSDPA and signaling overhead can bereduced.

4 Extension to MIMO systemAs the MIMO technique called D-TxAA can be applied inHSDPA, we propose a modified power control and linkadaptation scheme which is applicable to MIMO HSDPAsystems in this section.

When MIMO is configured, the Node B will transmitdata to the user through either single stream or dualstreams in the physical layer. If the former is selected,the proposed scheme in the previous section still workswell and the estimated EE is also given by (11). If thelatter is selected, only the EE estimation mechanism inthe Node B need to be modified. In this situation, theNode B estimates the sum EE of the two streams insteadof a single stream. As transmit power is always sharedequally between the two streams, transmit power modi-fications of the two streams must be the same duringthe reconfiguration. According to (8), the correspondingSINR threshold difference between the reconfiguredCQI and the previous one is also the same for the twostreams. For example, denote the feedback CQI index ofthe first stream as i1 and the second stream i2. TheMCS levels they indicated are θi1 and θi2 respectively. Ifthe corresponding CQI index for the first stream isadjusted to j1 and that for the second stream is adjustedto j2 when the transmit power is reconfigured, the MCSlevels used will be changed into θj1 and θj2 respectively.Denote the corresponding SINR threshold for CQIindex i1, i2, j1, and j2 as βi1 , βi2 , βj1 , and βj2 , respec-tively, the following equation must be satisfied:

βj1 − βi1 = βj2 − βi2 . (17)

The estimation of transmit power required for thenew MCS level pair θj1 and θj2 can be given by

Pnew = P + 2 · (βj1 − βi1). (18)

The estimation of the sum EE can be given by

ξnew =τj1 + τj2

ts ·(Pnew

η+ PDyn + PSta

) . (19)

Through comparing the sum EE among all possibleMCS level pairs of the two streams, the optimal trans-mit power and the corresponding MCS level pair fordual streams is selected.As the mode switching between single stream and dual

streams is done at the user side based on maximizing SE,one may argue that the chosen mode may not be themost energy efficient one. Interestingly, as the totalpower consumption is the same for the two modeaccording to the power model given by (4), the choicemade at the user side can lead to the most energy effi-cient mode, which can be explained as follows. Compar-ing (11) with (19), we can know that the denominators ofthe expressions on the right side are the same, so thevalue of estimated EE is determined by the numerators.Thus if the sum of transport block size of the preferredMCS levels for dual stream mode is greater than that for


Page 7 of 13

single stream mode, dual stream mode is selected by theuser, and vice versa. So the energy efficient criterion formode selection between single stream and dual streamsis the same as maximizing SE criterion.

5 Simulation resultsIn this section, we evaluate the performance of the pro-posed algorithm in different scenarios and give somediscussions on mode switching between MIMO andSIMO configuration along with our proposed schemeaccording to HSDPA link level simulation results. Amulti-path rayleigh fading channel model and path lossmodel of PA3 is considered. Bandwidth is 5 MHz, andthe duration of a subframe is 2 ms. The parameters ofpower model are set as h = 0.38, Pcir = 6W , and Psta =6W . The maximum transmit power is set to be 43dBm.Figures 5, 6, and 7 depict the performance of the pro-

posed semi-static power control method. Proposedenergy efficient power control used in every subframe isviewed as a performance upper bound and the traditionalscheme as a baseline where a transmit power of 40.5dBm is configured. If the energy efficient scheme is used,the transmit power will be configured based on the EEestimation as long as user’s feedback is available. Here

parameters of the semi-static trigger are set as gprohibit =20ms, gperiodic = 200 ms and Δ = 20%. Figure 5 showsthat a considerable EE gain of our proposed semi-staticpower control scheme can be acquired over the baseline.Furthermore, the proposed scheme’s EE performance iscomparable with the upper bound. Figure 6 demonstratesthat transmit power reconfiguration frequency is reducedcompared with the upper bound algorithm, thus signal-ing overhead is significantly reduced, due to the proposeddual trigger. The event trigger which sets a threshold forthe gap and the periodical trigger also ensures EE gain.Figure 7 also evaluates the performance of the algorithmunder different user speed. We can find that the EE gainwould decline with increasing user moving speed, andthe reason is explained that when the channel fluctuationbecomes faster because of increased moving speed, EEoptimal power changes more quickly. However, our pro-posed power configuration can not track this rapidchange due to the semi-static characteristic, so the EEgain decreases, but a considerable EE gain can still beobserved at high user speed.Figures 8 and 9 show the impact of path loss and mini-

mum CQI constrains on EE gain of our proposedscheme. Each minimum CQI constraint corresponds to aminimum MCS constraint. User speed is set as 3 km/h.

0 50 100 150 2000.8

1

1.2

1.4

1.6

1.8

2

2.2

2.4x 10

6

Subframe Index

EE

(bit/

Joul

e)

EE comparison for different strategy in 200 subframes

PC everysubframesemi−static PC no PC strategy

Figure 5 EE comparison of different strategies in 200 subframes.


Page 8 of 13

0 50 100 150 2002

3

4

5

6

7

8

9

10

11

Subframe Index

tran

smit

pow

er(d

BW

)transmit power reconfiguration frequency comparison in 200 subframes


Figure 6 Transmit power reconfiguration frequency comparison of different strategies in 200 subframes.

0 10 20 30 40 50 601

2

3

4

5

6

7

8

9

10

11x 10

5

User Speed(km/h)

EE

(bits

/Jou

le)

EE comparison of different strategies under different user speed


Figure 7 EE comparison of different strategies under different user speed.


Page 9 of 13

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10

0.5

1

1.5

2

2.5

3

3.5x 10

6

User distance from Node B(km)

EE

(bit/

Joul

e)

EE comparison under different user distance from Node B

PCwithCQImin=20PCwithCQImin=23PCwithCQImin=26NO PowerControl

Figure 8 EE comparison of power control strategies under different user distance from Node B.

18 20 22 24 26 28 300.6

0.8

1

1.2

1.4

1.6

1.8

2

2.2

2.4x 10

6

minimum CQI constraint

EE

(bit/

Joul

e)

EE comparison under different CQI constraints

PCDistance=400mPCDistance=500mPCDistance=600mNPCDistance=400mNPCDistance=500mNPCDistance=600m

Figure 9 EE comparison of different strategies under different CQI constraints.


Page 10 of 13

When the minimum CQI constraints are not so tight, wecan see that the EE gain of the proposed algorithm issimilar in Figure 8. EE gain decreases when user movesaway from Node B and the reason is that the optimaltransmit power increases and gradually approachesthe transmit power configured in the baseline. FromFigure 9, we can also observe that the looser the mini-mum CQI constrain is, the larger EE gain we can acquire.Figure 10 gives EE comparison between D-TxAA and

SIMO configuration under different transmit power inHSDPA, and Figure 11 illustrates the simulation resultsfor different EE performances between HSDPA-SIMOand HSDPA-MIMO systems by employing our pro-posed power control method. From Figure 10, we cansee that there exists an EE optimal transmit power foreach mode. Another observation is that EE performanceof SIMO mode is better than MIMO when transmitpower is not large, and vice versa. The reason isexplained as follows. The total power can be dividedinto three parts: PC power, transmit antenna numberrelated power PDyn, and transmit antenna number inde-pendent power Psta. When transmit power is large, PPCdominates the total power (the denominator of the EE)and PDyn is negligible. Because the MIMO mode can

acquire higher capacity, higher EE is available for thismode in the large transmit power scenario. When trans-mit power is low, the ratio of PDyn to the total powerincreases, and leads to lower EE for MIMO comparedwith SIMO. Figure 11 provides insights on the impactof the distance on the mode switching. When the dis-tance between the user and the Node B is getting larger,MIMO is better, and vice versa. This is because in thelong distance scenario, the first part increases and dom-inates the total power, then more active antenna num-ber is preferred.From Figures 10 and 11, we can conclude that signifi-

cant energy saving can be further acquired when adaptivemode switching between SIMO and MIMO is applied.However, adaptive mode switching may be dificult due tosome practical reasons. First, when SIMO mode is con-figured, parameters like PCI and CQI for the secondstream are not available because the second antenna isswitched off to save energy. Thus, how to estimate theavailable EE for D-TxAA is a challenge. Second, thetransmit antenna number information should beinformed through the system information, so the modeswitching will impact all users in the cell and bring hugesignaling overhead. To sum up, the protocol may need to

−10 0 10 20 30 400

2

4

6

8

10

12

14

16

18x 10

5

Transmit Power(dBm)

EE

(bits

/Jou

le)

EE comparison of MIMO and SIMO systems with different Power

SIMOMIMOsingle steamMIMOdual stream

Figure 10 EE comparison of MIMO and SIMO systems with different transmit power.


Page 11 of 13

be redesigned to utilize the potential EE improvementwith mode switching. Nevertheless, the Node B candecide the active antenna number according to the loadof the systems, which should be realized in the networklevel and is beyond the scope of this article.

6 ConclusionIn this article, we investigate the impact of transmitpower and MCS level configurations on EE in HSDPAand propose an energy efficient semi-static joint powercontrol and link adaptation scheme. We extend the pro-posed scheme to the MIMO HSDPA scenario. Simula-tion results prove that the EE gain is significant and themethod is robust. Finally, we have a discussion aboutthe potential EE gain of mode switching between SIMOand MIMO configuration along with the practical chal-lenging issues.

AbbreviationsHSDPA: high speed downlink packet access; MCS: modulation and codingscheme; MIMO: multiple input multiple output; SIMO: single input multipleoutput; ICT: Information and communications technology; HS-PDSCH: highspeed physical downlink shared channel; AMC: Adaptive Modulation andCoding scheme; HARQ: hybrid Automatic Repeat request; D-TxAA: dualstream transmit adaptive antennas; QoS: Quality of Service; PC: power

conversion; BER: bit error rate; SISO: single input single output; CQI: channelquality indicator; PCI: precoding control indicator; RF: Radio Frequency; SE:spectral efficiency; EE: energy efficiency; AWGN: additive white Gaussiannoise; SINR: signal to interference and noise ratio; ACK: acknowledgement;NACK: negative acknowledgement; MPO: measurement power offset; RRC:radio resource control.

AcknowledgementsThis study was supported by Huawei Technologies, Co. Ltd., China.

Competing interestsThe authors declare that they have no competing interests.

Received: 31 August 2011 Accepted: 6 March 2012Published: 6 March 2012

References1. L Correia, D Zeller, O Blume, D Ferling, Y Jading, I Godor, G Auer, L Der

Perre, Challenges and Enabling Technologies for Energy Aware MobileRadio Network. IEEE Commun Mag. 48(11), 66–72 (2010)

2. G Fettweis, E Zimmermann, ICT energy consumption - trends andchallenge, in Proceedings of the 11th International Symposium on WirelessPersonal Multimedia Communications (WPMC 2008), vol. 1. (Laplend, Finland,2008), pp. 2006–2009

3. 3GPP Technical Specification, TS 25.308 version 10.1.0, High SpeedDownlink Packet Access (HSDPA) Overall description http://www.3gpp.org/ftp/Specs/archive/25 series/25.308/25308-a50.zip

4. G Micallef, P Mogensen, H-O Scheck, Dual-Cell HSDPA for network energysaving, in Proceedings of IEEE Vehicular Technology Conference (VTC) 2010Spring, vol. 1. (Taipei, Taiwan, 2010), pp. 1–5

5. MA Marsan, L Chiaraviglio, D Ciullo, M Meo, Optimal energy savings incellular access networks, in Proceedings of IEEE Intl Conference on

200 300 400 500 600 700 800 900 10000

1

2

3

4

5

6

7

8x 10

6EE comparison for MIMO and SIMO systems with different pathloss

distance from Node B(km)

EE

(bit/

Joul

e)

MIMO UEspeed=3km/hSIMO UEspeed=3km/h

Figure 11 EE comparison of the proposed strategy for MIMO and SIMO systems in HSDPA with different user pathloss from Node B.


Page 12 of 13

http://www.3gpp.org/ftp/Specs/archive/25 series/25.308/25308-a50.zip


Communications (ICC)’09 Workshop, vol. 1. (GreenComm, Dresden, Germany,2009), pp. 1–5

6. HS Kim, B Daneshrad, Energy-constrained link adaptation for MIMO OFDMwireless communication systems. IEEE Trans Wirel Commun. 9(9),2820–2832 (2010)

7. SQ Zhang, Y Chen, SG Xu, Improving energy efficiency through bandwidth,power, and adaptive modulation, in Proceedings of IEEE Vehicular TechnologyConference (VTC) 2010 fall, vol. 1. (Ottawa, Canada, 2010), pp. 1–5

8. H Kim, GD Veciana, Leveraging dynamic spare capacity in wireless systemto conserve mobile terminals’ energy. IEEE/ACM Trans Netw. 18(3), 802–815(2010)

9. H Kim, C-B Chae, GD Veciana, RW Heath, Energy-efficient adaptive MIMOsystems leveraging dynamic spare capacity, in Proceedings of Conference onInformation Sciences and Systems (CISS), vol. 1. (Princeton, NJ, USA, 2008),pp. 68–73

10. J Xu, L Qiu, C Yu, Improving energy eficiency through multimodetransmission in the downlink MIMO systems. EURASIP J Wirel CommunNetw. 2011(1), 200 (2011). doi:10.1186/1687-1499-2011-200

11. 3GPP Technical Specification, TS 25.214 version 7.17.0 ‘Physical layerprocedures (FDD)http://www.3gpp.org/ftp/Specs/archive/25 series/25.214/25214-a30.zip

12. Y Chen, SQ Zhang, SG Xu, GY Li, Fundamental tradeoffs on green wirelessnetworks. IEEE Commun Mag. 49(6), 30–37 (2011)

13. J-B Landre, A Saadani, F Ortolan, Realistic performance of HSDPA MIMO inmacro-cell environment, in Proceedings of IEEE Int Symp on Personal, Indoorand Mobile Radio Commun (PIMRC), vol. 1. (Tokyo, Japan, 2009), pp. 365–369

doi:10.1186/1687-1499-2012-87Cite this article as: Huang et al.: An energy efficient semi-static powercontrol and link adaptation scheme in UMTS HSDPA. EURASIP Journal onWireless Communications and Networking 2012 2012:87.

Submit your manuscript to a journal and benefi t from:

7 Convenient online submission

7 Rigorous peer review

7 Immediate publication on acceptance

7 Open access: articles freely available online

7 High visibility within the fi eld

7 Retaining the copyright to your article

Submit your next manuscript at 7 springeropen.com


Page 13 of 13



http://www.springeropen.com/

http://www.springeropen.com/

Documents

RESEARCH Open Access An energy efficient semi-static power