WO2011096862A1

WO2011096862A1 - Method and arrangement in a wireless communication system

Info

Publication number: WO2011096862A1
Application number: PCT/SE2010/050137
Authority: WO
Inventors: Martin Alm; Fredrik Athley; Erik Eriksson; Magnus Lundevall; Arne Simonsson
Original assignee: Telefonaktiebolaget LM Ericsson AB
Current assignee: Telefonaktiebolaget LM Ericsson AB
Priority date: 2010-02-05
Filing date: 2010-02-05
Publication date: 2011-08-11
Anticipated expiration: 2012-08-05
Also published as: CN102742202A; US20150049629A1; US9871619B2; EP2532108A4; US20120287813A1; US8953480B2; CN102742202B; EP2532108A1

Abstract

Method and arrangement in a network entity for supporting link adaptation in a wireless communication system. The method comprises obtaining 204 one or more predicted parameters related to the quality of a radio link. The method further comprises measuring 206 one or more actual parameters, corresponding to the one or more predicted parameters. The method further comprises deriving 208 one or more error distributions based on the difference between the predicted and actual one or more parameters, from which error distributions a link adaptation margin estimate is derived, based on a predetermined radio link quality target. The link adaptation margin estimate is then used for supporting link adaptation for the radio link.

Description

MEIHO D AND ARRANGEMENT IN A W EME3S CO MMUNIC AID N SYSTEM

Technical field

[0001] The invention relates to a method and an arrangement in a wireless communication system, in particular to the support of link adaptation

Background

[0002] A commonly used practice to reach spectral efficiency in radio

communication systems of today is the use of adaptive modulation and coding (AMC). When applying AMC in a radio communication system, the modulation and coding of data to be transmitted are selected to match the current channel quality, in order to achieve high system throughput and low delay. R>r example, AMC is used in LIE both in the uplink (UI) and the downlink (DI).

[0003] In ITE, an instantaneous downlink channel quality is estimated by a mobile terminal, based on measurements performed on reference signals transmitted from a base station E om the measurements, a channel quality report, e.g. a Channel Quality Indicator (CQD, is derived and transmitted in the uplink to the base station Then, a modulation order and a code rate to be used when addressing said mobile terminal, can be selected in the base station, at least partially based on the reported quality.

[0004] A typical scenario comprising a network node 102 and a mobile

terminal 104 is illustrated in figure 1. The network node 102 transmits signals 108 to the mobile terminal 104 in the downlink, and the mobile terminal 104 transmits signals 106 to the network node 102 in the uplink Unwanted signal energy and/ or thermal noise, i.e. interference, is illustrated as dashed arrows 110. In the uplink, a channel quality estimate can be obtained from measurements on received mobile terminal transmissions 106. Hie mobile terminal transmissions 106 may comprise regular data or so called sounding reference symbols. A sounding symbol is a special reference symbol, which is known to the receiver, and is used for channel estimation. The interference 110 experienced in a network node 102, e.g. a base station, can be estimated by measuring, in the network node 102, the total received power, and then subtracting the power of the desired signal 106. The thus obtained interference estimate applies to all mobile terminals in the cell, even mobile terminals which are not currently transmitting. The modulation and coding scheme (MCS) to be used in the uplink is then indicated to the respective mobile terminals in a "grant' message. The grant is transmitted from the base station 102 to the mobile terminal 104 and indicates assigned resources and selected MCS.

[0005] Since the MCS mustbe selected prior to the transmission, it is always based on an estimate or prediction of the actual channel quality during the transmission. More or less advanced schemes can be used in the process of channel quality estimation, but a common strategy is to filter historic channel quality values, and subtract a safety margin, or "backoff', to compensate for the inherent uncertainty of the estimate.

[0006] The MCS should generally be selected to give high spectral efficiency and low packet delay. Often, a block error rate (BIER) target is given as a guideline of a good trade-off between throughput and delay. Different algorithms based on BIER targets are commonly used in communication systems such as High Speed l¾cket Access (HSPA), where the backoff is typically based on the experienced block error performance. Algorithms based on BIERare typically applied as an outer-loop, adjusting a backoff margin, set by a relatively fast inner-loop, based on e.g. CQI An example of such an outer-loop algorithm, which is commonly used, is the so called "jump algorithm", which is further described e.g. in the patent document

US7310499.

[0007] One of the major drawbacks of previously used outer-loop solutions is that they converge slowly and require a large number of transmissions in order to obtain sufficiently good statistics. This is especially true for low BIER targets. In OEDM-like systems, where mobile terminals can be assigned parts of different size of the frequency oand, the channel quality prediction can be very poor for small parts or allocations comprising few e.g. resource blocks, but quite good for large parts or allocations comprising a comparatively large number of resource blocks, lb handle this difference with one of said previously known outer-loop algorithms would slow it down even further.

[0008] Consequently, it is a problem that the above described estimation and adjustment of a link adaptation (IA) margin is slow, which has a negative effect on the efficiency in radio link resource utilization, especially at low to medium traffic loads.

Summary

[0009] It would be desirable to obtain a faster process for adjusting a link

adaptation to the actual/ true quality of a channel or radio link It is an object of the invention to address at least some of the issues outlined above. Further, it is an object of the invention to provide a method and an arrangement for estimating a link adaptation margin.

[0010] According to one aspect, a method is provided for supporting link

adaptation in a wireless communication system. Within the method, one or more predicted parameters related to the quality of a radio link are obtained. Further, one or more actual parameters, corresponding to the one or more predicted parameters, are obtained. Then, one or more error distributions are derived, which are based on the difference between the predicted and actual one or more parameters. Then, a link adaptation margin estimate is derived from the one or more error distributions, based on a predetermined radio link quality target, which link adaptation margin estimate is used for supporting link adaptation for the radio link

[0011] According to another aspect, an arrangement is provided, which is adapted to support link adaptation in a wireless communication system. The arrangement comprises a functional unit for obtaining predictions of one or more parameters related to the quality of a radio link, and a functional unit for measuring the one or more parameter. The arrangement further comprises functional units for deriving one or more error distributions based on the difference between the predicted and actual one or more parameters, and for deriving a link adaptation margin estimate from the one or more error distributions, based on a predetermined radio link quality target, which link adaptation margin estimate is used for supporting link adaptation for the radio link

[0012] The above method and arrangement may be used to rapidly obtain a link adaptation margin estimate, and may be applied, located or integrated in a network node, such as a base station, or in a mobile terminal. Further, the method and arrangement enables a joint measurement of interference in the uplink for all user terminals, which gives a more accurate result as compared to when measurements are made per user terminal. Further, the method and arrangement enables that the received signal strength/ gain in the uplink can be measured for sounding signals and scheduled CSI transmissions, which is not the case in currently used solutions. This possibility enables a faster updating of a link adaptation margin estimate, than when using previously known solutions. [0013] The above method and arrangement may be implemented in different embodiments. In some embodiments, the predetermined quality target is related to a BIER target

[0014] In some embodiments, the predicted and measured one or more

parameters comprise Signal-to-titerference-and-Noise Ratio (S1NR), and in some other embodiments, the predicted and measured one or more parameters comprise received signal strength/ gain and interference. The dividing of estimation and measurement of the SINRinto separate estimations and measurements of received signal strength/ gain and interference may be beneficial, since received signal strength/ gain, typically, is different for each link, whereas the interference may be similar for multiple links. Further, it may be beneficial to differentiate the updating frequency of the estimates related to the different components received signal strength/ gain and interference.

[0015] In some embodiments, an error distribution of at least one correction factor e is derived. The correction factor e could be related to e.g.

characteristics of one or more network entities, such as interference suppression or cancellation. It could further be related to M O (Multiple- put Multiple-Output), such as effects of MU-M O (Multi User-M O) and/ or different MEMO transmission formats; a receiver method, or an error in modeling receiver performance. The correction factor e enables the method and arrangement to take e.g. the above mentioned items into account when estimating a link adaptation margin, thereby obtaining a more accurate estimate.

[0016] In some embodiments, different error distributions are derived for each receiver method being used, and/ or for each MMO transmission format being used. [0017] In some embodiments, values from atleasttwo error distributions are combined when deriving the link adaptation margin estimate. Pbr example, a signal strength/ gain prediction error and a corresponding interference prediction error could be added to form a link adaptation margin estimate, or, two error distributions could be combined to one error distribution before deriving an estimated link adaptation margin.

[0018] In some embodiments, measurements are performed in a frequency selective manner, i.e. for parts or subsets of the frequency band, which enables deriving different link adaptation margin estimates for different fractions of the frequency band.

Erief description of drawings

[0019] The invention will now be described in more detail by means of

exemplary embodiments and with reference to the accompanying drawings, in which:

[0020] iigure 1 is a schematic view illustrating a network node and a mobile terminal, in which embodiments of the invention could be applied.

[0021] figure 2 is a flow chart illustrating method steps according to an

embodiment

[0022] Figure 3 is a block diagram illustrating an embodiment of an

arrangement in a network entity.

[0023] iigure 4 shows the probability distributions of the signal strength/ gain- and the interference estimation error, according to one embodiment

[0024] iigure 5 shows the probability distributions of the S1NR estimation error at different mobile terminal speeds and antenna separations, according to one embodiment [0025] Figure 6 shows an estimate of a suitable link adaptation margin at different antenna separations and different mobile terminal speeds.

[0026] Figure 7 shows the BIER at different mobile terminal speeds for two different antenna separations when the link adaptation margin of figure 6 has been applied, according to one embodiment

Detailed description

[0027] Briefly described, a method and arrangement are provided for

estimation of a link adaptation margin by use of one or more error distributions, e.g. a Signal-to-lnterference-and-Noise Ratio (S1NR) estimation error distribution. An adequate link adaptation margin may be derived, in alternative ways, from the one or more error distributions, based on a desired quality target The one or more error distributions, and thereby the estimated link adaptation margin, can be made frequency selective. The link adaptation margin can be derived e.g. from a estimation error distribution of the S1NR, or alternatively, from the estimation error distributions of the received signal strength and the interference

[0028] Some definitions:

Within this document some expressions will be used when discussing the procedure of estimating a link adaptation margin, of which some will be briefly defined here.

[0029] The expressions "backoff' and "backoff margin" are used as referring to a safety margin, such as a link adaptation margin.

[0030] The terms "gain" and "signal strength", when referring to estimated or measured parameters, are used interchangeably, even though the channel gain is a normalization, in terms of transmission power, of a received signal strength The parameter, which is measured in a receiver is the received signal strength, but the parameter, which clearly indicates the effect of a radio link on a received signal is the gain. Therefore, the expression "signal strength/ gain" is used throughout this document

[0031] The term "network entity'' is used as referring to both network nodes and mobile terminals. The term "network node" is used as referring to radio transmitting network entities having a base station-like function and applying link adaptation. Such network nodes could also be referred to as e.g. eNB, base station, relay node, RB3, NodeB, etc, depending on function and on which radio access technology (BAT) that is discussed. The terms "mobile terminal" and "user terminal" are used as referring to any radio transmitting wireless mobile devices, such as e.g. user equipment (UE), personal digital assistants (IDA), laptops, or the like, for which link adaptation is applied.

[0032] The "radio link', of which the quality is discussed herein, is only

considered in one direction, i.e. the direction towards the network entity in which an embodiment of the invention is applied.

[0033] Generally, when deciding which channel coding and modulation to apply in a network entity, a SINRis estimated. Then, the SINRestimate is adjusted by an estimated link adaptation margin, which is updated by an outer-loop based on ACKZ NACK statistics from a receiver. Then, a modulation and coding scheme (MCS) is selected based on the adjusted SlNRestimate. In other words, a parameter related to the quality of a radio link is estimated, and an MCS is selected based on the estimated parameter adjusted by an LA margin.

[0034] Within embodiments of the invention, it is instead the probability

distribution of the error in the estimates of one or more parameters related to the quality of a radio link, which is used as base for deriving e.g. an LA margin and an MCS. Given a probability distribution, e.g., of a S1NR estimation error distribution, an IA margin is selected to achieve a

predetermined radio link quality target, e.g. a Block Error Rate (BIER) target, from a percentile of the S1NR estimation error distribution corresponding to the predetermined radio link quality target This will be described in more detail later below.

[0035] Further, the effective SlNRestimate can be seen as a function of the signal strength, or gain, S and the interference 1

SINR = f(S,I) . (i )

Thus, as an alternative to deriving the IA margin and MCS from the error distribution of a SlNRestimate, these parameters could be derived from the separate error distributions of an S-estimate and an lestimate. This may be done either directly, e.g. from the error distributions of S-estimates and I- estimates, respectively, as illustrated in figure 4, or from an error distribution of SlNRestimate s, which error distribution is formed by combining the error distributions of the S- and lestimate s. The dividing of estimation and measurement of the SINRinto separate estimations and measurements of S and I may be beneficial, since S, typically, is differentfor each link, whereas I may be similar for multiple links

[0036] Further, a S1NR estimation error depends on the estimation errors in the respective components signal strength/ gain and interference. These components may fluctuate in different ways. Therefore, it may be beneficial to differentiate the updating fiiequency of the estimates related to the different components, e.g. such that the estimates related to a component, which changes relatively fast may be updated more frequently than estimates related to a component, which changes comparatively slow. 37] Below, different embodiments of an IA margm estimation procedure will be described. The procedure of estimating a link adaptation margin can be applied for transmissions both in the uplink (UI) and the downlink (DI), i.e. in both network nodes and mobile terminals. The example embodiments below are written with the 3rd Generation Rirtnership RiojectOGEP) LIE standard as baseline, but are applicable, possibly with some modifications to match respective system implementations, to any system using AMC, such as e.g. Mobile WiMAXand WCDMA. 38] A step-wise description of an embodiment for UL transmissions, i.e. an embodiment in a network node, where the gain and interference are measured and estimated separately, could be made as follows:

• Derive the interference estimation error distribution per cell. This could be done frequency selective, e.g. per resource block (KB) in HE.

• Derive the signal strength, or gain, estimation error distribution per mobile terminal, based on measurements on uplink transmissions from mobile terminals. This could also be done frequency selective, or not, depending on which system that is used.

• Schedule a mobile terminal for uplink transmission in a number of resource blocks.

• Extract the error percentile corresponding to a predetermined radio link quality target, e.g. a desired maximal HARQ (Hybrid Automatic Repeat Request) BIER, from the signal strength/ gain- and interference estimation error distributions for the scheduled resource blocks. • Combine the extracted percentiles, to form an estimated S1NR error percentile.

• Use the estimated SINR error percentile as link adaptation

margin, e.g. when selecting MCS for uplink transmissions in the resource blocks in question

[0039] Alternatively, the SINR estimation error distribution could be derived, either by estimating the SINR rather than the signal strength/ gain and interference separately, or, by coinbining the interference- and signal strength/ gain estimation error distributions to a SINR estimation error distribution before extracting the error percentile corresponding to the predetermined radio link quality target

[0040] In general terms, an estimation error distribution can be estimated

according to the following formula:

S_am = S _X) g(Q_x (t) - Qx (t \ t - d)) (2) CDF(S_{a m} ) (3) where S^, or S_; for a distribution i, is a set of samples of a real valued variable, e.g an estimation error distribution, and CDF^S^) is the empirical cumulative distribution function of the set of samples S_t^. The parameter Q is a measured quality at time t, where quality may be e.g. interference or signal strength/ gain; Q is the estimated or predicted quality estimated at time t-d, where d is the delay between the selection of an MCS for a certain

transmission and the reception of the transmission R>r example, in T F1, d is approximately 6 ms; Xis the resource for which the estimation is done. In T R, X may typically be a resource block (RB) pair, a set of RB pairs, or RB pairs over the entire bandwidth. [0041] The function σ is typically a mapping σ(Χ) = Xor a(X) = 1 for all X me first case, where σ(Χ) = X, an estimation error distribution is derived for each resource X, while in me second case, where σ(Χ⁾ = 1 , only one estimation error distribution is derived, which covers all resource allocations X When Xis a fiiequency resource, such as an LIE RB pair, the mapping σ(Χ⁾ = X enables fiiequency selectivity, while the mapping σ(Χ⁾ = 1 implies one estimation error distribution for the entire frequency band, i.e. all LIE RBs concerned. When Q is related to signal strength/ gain, σ(Χ) = 1 may be a preferred alternative, while if Q is related to interference, and frequency selective interference is expected, due to for example inter-cell interference coordination (ICIC) or scheduler behavior, σ(Χ⁾ = Xcan be a preferred alternative for the interference.

The parameter g in equation (2) is a mapping to the set S_t^. Pbr example, if S is implemented as a bucket based histogram, g is typically a mapping to the closest bucket, i.e. quantization

[0042] iurther, the statistics of the interference experienced in a network node is common for all mobile terminals served by the network node. Thus, the uplink interference can be jointly and continuously measured, e.g. at predetermined intervals, which enables an accurate and up to date error distribution per resource block, which captures interference variations. The measurements of received signal strength in a network node, on the other hand, are made separately for each mobile terminal, and can only be performed when the respective mobile terminals transmit

[0043] ibr a mobile terminal scheduled over KB = {j...k}, an appropriate IA margin estimate can be derived by first estimating an estimation error distribution for the mobile terminal, using a receiver model, ibr example, a receiver model "f can be formulated as follows: f(S ) = Aat≡L (3) where the signal strength received by different receiver antennas are combined.

[0044] An appropriate SINRIA margin can then be estimated or derived by mapping into the CDF of the error distribution of the SlNRata certain point, I¾rc_¾Iget, where I¾rc_¾Igetis a constant or a function dependent on a predetermined radio link quality target, e.g. a selected BIER target By "mapping" is meant reading or coirputing an x-value corresponding to a certain y-value in a distribution An example of deriving an LA margin estimate from CDPs will be described further below, with reference to figure 4.

[0045] The above described function SlNR=f(S,l), equation (1), is suitable for receivers having no ability to suppress or cancel interference. R>r receivers having such an ability to suppress or cancel interference, e.g. receivers applying so called MMSE (Minimum Mean Square Error) reception, SIC (Successive Interference Cancelation) or IRC (Interference Rejection

Combining), there will also be estimation errors depending on receiver characteristics. R>r example, even if a SlNRis estimated with approximately 100% accuracy, the estimated LA margin would still notbe appropriate for e.g. an MMSE receiver, since such a receiver may suppress the interference, and thereby experience an interference, after suppression, which is lower than the actual measured, and possibly accurately estimated, interference. The interference estimation error depending on receiver characteristics is then the difference between the received interference and the experienced interference after interference suppression Receivers could also improve the experienced received signal strength/ gain, e.g. by using antenna arrays, or the like, which also will re suit in S1NR estimation errors. Another example of receiver characteristics, which may induce e.g. S1NR estimation errors, is the use of an adaptive amount of receiver antennas.

[0046] Erors depending on receiver characteristics could, at least partly, be compensated for by the introduction of a receiver efficiency error distribution term. Instead of assuming perfect compliance between model and real receiver performance, an error term e is added to the original equation (1), and f is replaced with an estimate of f, / .

SINR = f(S,I) + e (4)

The model errors could then be calculated as the difference between received processed S1NR, including interference rejection/ suppression and antenna array gains, and the estimated SlNRbased on the measured "unprocessed" interference and signal strength/ gain. An estimate of a safety margin to be applied to the estimated SlNRis then based on the combination of the signal strength/ gain estimation error distribution, the interference estimation error distribution and the modeling error distribution.

[0047] In the uplink, the modeling error distribution can be estimated per

mobile terminal or per cell. It may be preferred to make the estimation per cell, since the modeling error mainly depends on the interference

characteristics and the cell antenna setup, which is common for all mobile terminals in a cell. Using one uplink modeling error distribution per cell enables measurement updates whenever any mobile terminal is scheduled within the cell.

[0048] Similarly to the compensation for receiver characteristics described

above, the effect of performance improving transmitter characteristics, such as pre-coding/ bearnforming, could be taken into account and be compensated for in the estimation of an IA margin The effect, e.g. ofMU-M O (Multi User - Multiple Input Multiple Output) could also be taken into account, for example by altering the interference measurement depending on if

simultaneous transmission is scheduled or not, or by adding a MU-MMO error distribution to the model. 49] Thus, one error term e_m could be added for each error generating

feature m, that is necessary and/ or desired to take into account in a certain embodiment, as follows:

where m=(0,l,...,n)

[0050] An embodiment of the procedure of supporting link adaptation will now be described with reference to the flowchart in figure 2 with steps performed by a network entity. Initially, a signal transmitted from another network entity and interference are received in a step 202. The signal could comprise, e.g. user data or reference signals. Alternatively, the signal could comprise estimates related to the quality of a radio link Then, one or more estimates of parameters related to the quality of the radio link, such as e.g. S1NR, or separately, received signal strength S and interference I, are obtained in a step 204. ibr received signals comprising estimates of one or more parameters related to the quality of the radio link, which are made in another network entity, the obtaining in step 204 only involves e.g. extracting the desired parameter estimates from the received signal, and storing them, e.g. in a memory, for later use. Otherwise, i.e. for other signal contents and interference, the obtaining in step 204 involves estimation, i.e. prediction, of quality related parameters of the radio link, over which the received signal was transmitted. In order to estimate said parameters, the obtaining step should have access to measurements performed on the radio link

[0051] At the reception of a transmission over the radio link, measurements are performed in a step 206 on the one or more parameters related to the quality of the radio link, e.g. the S1NR, or separately, the signal strength of a signal received over the radio link, and the interference experienced attiie network entity. The interference parameter is present also when no

transmissions from other network entities arrive, and could therefore be measured continuously, e.g. at certain predetermined intervals. The signal strength/ gain parameter, on the other hand, could only be measured when receiving a transmission from another network entity.

[0052] Then, the one or more measured, actual, parameters are compared to the corresponding one or more estimated parameters, and from the difference between the actual and the estimated one or more parameters, one or more parameter estimation error distributions are derived in a step 208. From the one or more derived error distributions, an estimated link adaptation margin is derived in a step 210. One way to derive the estimated link adaptation margin is to derive it from the CDEs of the estimation error distributions at a certain quality target percentile, which will be further described later below with reference to figure 4. The derived link adaptation margin estimate is then provided, either e.g. to a functional unit within the network entity, or to another network entity, for use as link adaptation margin when selecting which MCS that should be used when transmitting over the radio link.

[0053] Depending on if the selection of MCS is performed in a step 211 within the network entity or in another network entity, a signal transmitted to another network entity in step 212 comprises either the link adaptation margin estimate or an indication of which MCS to use when transmitting over the radio link R>r example, when the selection of MCS is not performed within the network entity, the link adaptation margin estimate is transmitted in step 212 to another network entity, which is to transmit over the radio link concerned. When the selection of MCS is performed in a step 211 within the network entity, a selected MCS is transmitted in step 212 to another network entity, which is to transmit over the radio link concerned.

[0054] There may be variations in how the above procedure is to be

implemented in different systems, such as e.g. T F1, WiMAX, WCDMA, due e.g. to differences in how and where the measuring and reporting of channel parameters are implemented today, and how and where the link adaptation margin is currently selected.

[0055] Below, an example arrangement 300, adapted to enable the

performance of the above described procedure in a network entity 301, will be described with reference to figure 3. The network entity 301 comprises a receiving unit 302, which is adapted to receive signals transmitted from other network entities over a radio link The received signals could comprise e.g. user data or pilot symbols, such as sounding or demodulation reference symbols. Alternatively, the signals could comprise estimates related to the quality of a radio link The arrangement 300 comprises a prediction unit 310, which is adapted to obtain one or more estimated parameters related to the quality of the radio link, and a measuring unit 304, which is adapted to measure one or more actual parameters corresponding to the one or more estimated parameters. The measuring unit 304 may also be adapted to provide the result of measurement of parameters related to the quality of the radio link to the prediction unit 310, which then could estimate future radio link quality parameters based on the provided result [0056] The arrangement 300 further comprises a distribution unit 306, which is adapted to derive one or more error distributions based on the differences between the one or more estimated radio link parameters and the one or more measured actual radio link parameters. An exemplary quality related parameter, which could be estimated and measured, is the S1NR, or, alternatively, the interference and the signal strength/ gain. The arrangement 300 further comprises a link adaptation margin unit 318, which is adapted to derive a link adaptation margin estimate from the one or more derived error distributions, based on a predetermined radio link quality target The radio link quality target may typically be based on a BIER target, but could also be expressed independently of a BIER The link adaptation margin unit 318 is also adapted to support link adaptation for the radio link, e.g. by providing the link adaptation margin estimate for use as margin when selecting MCS in link adaptation. The selection of MCS could e.g. be performed m an MCS unit within the network entity 301, or be performed in another network entity

[0057] The network entity 301 further comprises a transmitting unit308, which is adapted to transmit signals to other network entities. The signals could e.g. comprise the link adaptation margin estimate, and/ or the signals could comprise an indication of an MCS to be used by another network entity when transmitting over the radio link The receiving unit 302 and the transmitting unit 308 could also be referred to the arrangement 300, i.e. be regarded as a partof the arrangement 300. However, in this example, the receiving and transmitting units 302 and 308 are regarded as regular components of the network entity 301

[0058] In figure 3 is also illustrated a computer program product 314 which comprises instructions 316, which when executed by a processor 312 , or similar, will cause the units 304-306, 310 and 318, respectively, to perform their tasks according to an embodiment of the described procedure.

Connections between the processor 312 and the units 304-306, 310 and 318 are illustrated by a dashed arrow.

[0059] It should be noted that figure 3 merely illustrates various functional units in the network entity 301 in a logical sense. However, the skilled person is free to implement these functions in practice using any suitable software and/ or hardware means, such as e.g. ASICs (Application-Specific Integrated Circuit), FBGAs (ileld-HOgrammable Gate Array) and DS1¾ (Digital Signal frocessor). Thus, the invention is generally not limited to the shown structure of the network entity 301.

[0060] Below, an embodiment of deriving a link adaptation margin estimate from two parameter estimation error distributions for a radio link, based on a BIER target, will be described with reference to figure 4. Hgure 4 shows the CDFs of a signal strength/ gain- and an interference estimation error, respectively. In this example, the signal strength/ gain and the interference are estimated and measured separately. The distributions of the signal strength/ gain estimation error and the interference estimation error are assumed to be uncorrelated. The desired radio link quality, BIER^_t, is 10% HARQ BIER Thus, the target CDFlevel, B3rc_teIget, could be derived as l¾rc_teIKet = (Ι- ΒΙΈ^^^-ίΙ- ΒΙΈ^_^ ^_^^Ο,Θ^Ο,δΙ. In the diagrams in figure 4, the target CDFlevel 0,81 is indicated by a bold dashed line. The signal strength/ gain estimation error corresponding to the target CDF-level 0,81 is approximately 1 dB, and the interference estimation error corresponding to the target CDFlevel 0,81 is approximately 4 dB. These values are indicated by vertical arrows in the diagrams. Since the signal strength/ gain estimation error and the interference estimation error are assumed to be uncorrelated in Ibis example, a link adaptation margin estimate can be formed by adding the two estimation errors corresponding to the target CDFlevel B3rc_teIget, resulting in a link adaptation margin estimate of approximately (1+4) dB = 5 dB. This link adaptation margin estimate can then be used when selecting a transport format, such as an MCS, for a radio link

[0061] An additional adjustment factor k may be needed in order to reach a desired quality target, due to other error factors, such as bias in the error estimation method in general, correlation between S and I errors, receiver dependent errors, M O and others, as described above, if not specifically taken into account For example, Berc_teIKet , which was set to (l-BIER^,,)² in the previous example, can be set to B3rc_teIget= (1 - BIER^,,)² - k, where k is an empirical adjustment factor that can be acquired, e.g. through simulation and/ or by using statistics from an outer loop. B3rc_teIgetcould also be expressed independently of a BIERtarget, e.g. as B3rc_teIget= k, where k may be a function of a BLER target or other quality target, such as bit error rate, residual packet error rate after HARQ retransmission, delay or niaximum number ofHAEQ retransmissions. The acquirement of an adjustment factor k through simulation is further described below with reference to figures 5-7.

[0062] iigure 5 shows the results of a set of downlink simulations applying one embodiment of the invention. Simulations have been performed for two different antenna separations, 1λ and 10λ, where λ is the carrier

wavelength, and for a range of mobile terminal speeds, 0, 3, 10, 30 and 120 kph. In the simulations, the estimation error distribution of the SINRis used to derive a link adaptation margin estimate, i.e. notthe separate signal strength/ gain- and interference estimation error distributions, as in previous examples. The diagrams in figure 5 show the CDF of the S1NR estimation error in percent (%) to the SINR estimation error in decibel (dB). Hie variance of the SINR estimation error increases with increasing mobile terminal speed in the simulations

[0063] Figure 6 shows the SINR estimation error in dB at CDF 80% for the different antenna separations and mobile terminal speeds, which could also be retrieved from the diagrams in figure 5. The level 80% in the CDF corresponds to R3rc_teIget = 80%, i.e. that the actual SINR should be equal to or better than the estimated SlNRin at least 80% of the transmissions, and the percentile 80% is selected as a rough estimate of a suitable R3rc_teIget, corresponding to a predetermined radio link quality target, in this simulation approximately 10 % BIER Thus, figure 6 shows a rough estimate of which link adaptation margin that could be suitable to use at the different antenna separations and mobile terminal speeds, respectively.

[0064] Figure 7 shows the measured BLERafter simulated transmission over a radio link where the estimated values shown in figure 6 have been used as link adaptation margin. With a BLER_^^ of 10%, empirical variations of the factor k in the simulations gave thatk=0,l was a good selection of k in this scenario. The selection of k=0,l showed to work for the two simulated base station antenna configurations and for a range of mobile terminal speeds, which can be seen in figure 7.

[0065] The results of the simulation discussed above confirm the

correspondence between a percentile of the error distribution and a HARQ BIER

[0066] As previously mentioned, the embodiments of the invention could be applied in both the uplink and the downlink Currently, there are no LIE standard related issues preventing the implementation of different

embodiments in the uplink Regarding the LIE downlink, the existing parameters RSSI (Received Signal Strength Indicator) including interference, and RSRP (Reference Signal Received Ibwer) excluding interference, could presumably be utilized in the estimation of an IA margin. However, the reporting mechanisms for these parameters are currently not adapted to serve embodiments of the invention. In order to achieve an efficient downlink solution, new channel quality reporting and some minor changes of the T F1 standard would be recommended, ibr example, one possible standard change could be to split the CQI-reporting in two, by analogy with the RSSI and RSRP, thereby enabling separate interference estimation. However, the described method can be applied in a mobile terminal for do wrilink within the standard, when excluding the impact of delay on error, i.e. when setting d=0 in equation (2). When d is set to d=0, Q in equation (2) represents the result of the estimation algorithm for CQ I reporting of a mobile terminal, and Q is the quality at reception of downlink data. 67] Other possible additional features of the embodiments could e.g. be filtering of the measurements, weighting new measurements higher than old, for example by applying some forgetting factor. Such filtering would handle statistical changes in interference and fading patterns, such as increased variance due to increased terminal speed, or changed interference due to load changes in neighboring cells, Further, the interference could be measured after antenna combiriing, capturing receiver interference rejection capability. The signal strength could be measured per antenna and the error distribution per antenna could then be combined according to receiver method, iurther, mobile terminals arriving to a cell, having no relevant measures of received signal strength/ gain or interference, to which an error distribution could be formed, could initially use a cell-common error distribution instead. [0068] Exemplary advantages of embodiments of the invention are e.g. the enabling of a faster convergence of the estimation and adjustment of an IA margin, as compared to conventionally used regular outer-loop algorithms. Further, the interference can be measured jointly for all user terminals in the uplink, which gives a more accurate result than measurements per user terminal. In the uplink, the actual received signal strength/ gain can be measured for user data transmissions, but also for e.g. sounding and scheduled C SI transmissions, which is not the case in currently used solutions. This possibility enables a faster updating of an IA margin estimate, than when using previously known solutions. The access to more accurate IA margin estimates results in higher system throughput, lower transmission delay and a more stable system operation in wireless communication system. A further advantage is that the separation of S1NR measurements into separate measurements of the SlNR-components signal strength/ gain and interference corresponds well to how channel quality measurements are implemented in many network entities today, which enables a more efficient IA margin estimation process.

[0069] While the invention has been described with reference to specific

example embodiments, the description is in general only intended to illustrate the inventive concept and should notbe taken as limiting the scope of the invention The different features of the exemplary embodiments above may be combined in different ways according to need, requirements or

preference. Although the description has mainly been exemplified in terms of HE, the scope of the invention is not limited hereto, but could also be applied in other systems employing link adaptation, such as e.g. WiMAXand

WCDMA. The invention is generally defined by the following independent claims.

Claims

CIAJMS

1. A method in a network entity, for supporting link adaptation in a wireless communication system, the method comprising: btaining (204) one or more predicted parameters related to the quality of a radio link,

treasuring (206) one or more actual parameters, corresponding to the one or more predicted parameters,

-deriving (208) one or more error distributions based on the difference between the predicted and actual one or more parameters, -deriving (210) a link adaptation margin estimate from the one or more error distributions, based on a predetermined radio link quality target, and

-supporting (210, 212) link adaptation for the radio link, based on the link adaptation margin estimate.

2. Method according to claim 1, wherein the predetermined quality targetis related to a block error rate, BIER, target

3. Method according to claim 1 or 2, wherein the predicted and measured one or more parameters comprise:

■received signal strength and interference, or

-signal-to nterference-and-noise ratio, S1NR

4. Method according to any of the preceding claims, further comprising:

-deriving an error distribution of at least one correction factor e, which correction factor is based on one or more of the following: interference suppression or cancellation,

-characteristics of one or more network entities,

multi user multiple input multiple output, M O,

-a multiple input multiple output, MMO, transmission format,

-an error in modeling the receiver performance,

-a receiver method.

Method according to claim 4, wherein different error distributions are derived for each receiver method being used, and/ or for each multiple input multiple output, MMO, transmission format being used.

Method according to any of the preceding claims, further comprising:

-combining values from at least two error distributions when deriving the link adaptation margin estimate.

Method according to claim 6, further comprising:

-combining at least two error distributions to one error distribution before deriving an estimated link adaptation margin.

Method according to any of the preceding claims, wherein the network entity is a network node or a mobile terminal.

Method according to any of the preceding claims, wherein the atleastone measurement is frequency selective and the link adaptation is performed in a fraction of the frequency band.

Method according to any of the preceding claims, wherein the action of supporting link adaptation involves providing the link adaptation margin estimate for use as link adaptation margin when selecting a modulation and coding scheme, MCS, to be used for transmissions over the radio link.

11. Arrangement in a network entity, adapted to support link adaptation in a wireless communication system, the arrangement comprising:

-a prediction unit (310), adapted to obtain one or more predicted parameters related to the quality of a radio link,

-a measuring unit (304), adapted to measure one or more actual parameters, corresponding to the one or more predicted parameters,

-a distribution unit (306), adapted to derive one or more error distributions based on the difference between the predicted and actual one or more parameters,

-a link adaptation margin unit (318), adapted to derive a link adaptation margin estimate from the one or more error distributions, based on a predetermined radio link quality target, and to support link adaptation for the radio link based on the link adaptation margin estimate.

12. Arrangement according to claim 11 , wherein the link adaptation margin unit (318), is adapted to derive the link adaptation margin estimate from the one or more error distributions, based on a predetermined radio link quality target, which is related to a BIER target

13. Arrangement according to claim 11 or 12, adapted to predict and

measure received signal strength and interference; or signal-to interference- ratio, S1NR

14. Arrangement according to any of the claims 11-13, wherein the

distribution unit (306) is further adapted to derive an error distribution of at least one correction factor e, which correction factor is based on one or more of the following: interference suppression or cancellation,

-characteristics of one or more network entities,

multi user multiple input multiple output, M O,

-a multiple input multiple output, MMO, transmission format,

-an error in modeling the receiver performance,

-a receiver method

Arrangement according to any of the claim 11-14, wherein the distribution unit is adapted to derive an error distribution for each receiver method being used, and/ or for each multiple input multiple output, MMO, transmission format being used.

Arrangement according to any of the claims 11-15, wherein the link adaptation margin unit (318) is adapted to combine values from at least two error distributions when deriving the link adaptation margin estimate.

Arrangement according to claim 16, wherein the link adaptation margin unit(318) is adapted to combine atleasttwo error distributions to one error distribution before deriving an estimated link adaptation margin.

Arrangement according to any of the claims 11-17, wherein the network entity is a network node or a mobile terminal.

Arrangement according to any of the claims 11-18, adapted to perform the at least one measurement in a frequency selective manner.

Arrangement according to any of the claims 11-19, wherein the arrangement is adapted to provide the link adaptation margin estimate, either within the network entity or to another network entity, for use as link adaptation margin when selecting an modulation and coding scheme, MCS, to be used for transmissions over the radio link