Radio network node and method performed therein

Abstract

A method performed by radio network node for enabling channel handling of a channel between a wireless device and the radio network node in a wireless communication network. The channel is defined in continuous time and a sampling rate of the channel is non-uniform. The radio network node predicts a channel gain using a first sampling descriptor indicating a first momentary sampling frequency and a second sampling descriptor indicating a second momentary sampling frequency, wherein the first sampling descriptor operates on a different segment of continuous time than the second sampling descriptor. The predicted channel gain enables channel handling such as channel estimation and link adaptation.

Claims

1. A method performed by radio network node for enabling channel handling of a channel between a wireless device and the radio network node in a wireless communication network wherein the channel is defined in continuous time and a sampling rate of the channel is non-uniform; the method comprising predicting a channel gain using a first sampling descriptor indicating a first momentary sampling frequency and a second sampling descriptor indicating a second momentary sampling frequency, wherein the first sampling descriptor operates on a different segment of continuous time than the second sampling descriptor, which predicted channel gain enables channel handling.

2. The method according to claim 1, wherein the predicting the channel gain comprises obtaining the channel gain by a linear prediction, wherein a continuous time estimated parameter vector is multiplied with a regression vector obtained using the first and second sampling descriptors reflecting varying sampling periods.

3. The method according to claim 1, wherein the predicted channel gain at time t, (t), is defined by
{circumflex over (y)}(t).sup.T(t){circumflex over ()}(tk.sub.2h)+c(t) where (t) is a regression vector at time t {circumflex over ()}(tk.sub.2h) is a channel estimate at a time taking the second sampling descriptor into account; and c(t) is a parameter independent part of the prediction.

4. The method according to claim 3, wherein the regression vector (t) is defined as
(t)=(k.sub.2h(y(tk.sub.2h)y(t(k.sub.1+k.sub.2)h))k.sub.1k.sub.2h.sup.2y(t(k.sub.1+k.sub.2)h)).sup.T where h is a fundamental sampling period; and y is a measured channel gain.

5. The method according to claim 3, wherein the parameter independent part of the prediction c(t) is defined as $c\left(t\right)=y\left(t-k_{2}h\right)+\frac{k_2}{k_1}\left(y\left(t-k_{2}h\right)-y\left(t-\left(k_1+k_2\right)h\right)\right)$ where h is a fundamental sampling period; and y is a measured channel gain.

6. The method according to claim 1, wherein the channel is defined in terms of a parameter vector, being a continuous time estimated parameter vector, as
=(a.sub.1a.sub.2).sup.T where a.sub.1 a.sub.2 are the continuous time parameters.

7. The method according to claim 1, further comprising, performing a channel estimation using the predicted channel gain.

8. The method according to claim 1, when the sampling rate being below a threshold rate, further comprising feeding channel samples into a multirate predictor performing the prediction.

9. The method according to claim 8, when the sampling rate being equal or above the threshold rate, further comprising, averaging the channel samples instead to obtain an averaged channel gain estimate.

10. The method according to claim 1, further comprising, calculating an average of the channel gain by filtering channel samples; calculating a first magnitude of difference between received channel samples and the calculated average of the channel gain; calculating a second magnitude of difference between the received channel samples and the predicted channel gain; and selecting the predicted channel gain or the calculated average of the channel gain which ever gives a minimum expected error based on the calculated first and second magnitudes.

11. A radio network node for enabling channel handling of a channel between a wireless device and the radio network node in a wireless communication network wherein the channel is defined in continuous time and a sampling rate of the channel is non-uniform; the radio network node being configured to: predict a channel gain using a first sampling descriptor indicating a first momentary sampling frequency and a second sampling descriptor indicating a second momentary sampling frequency, wherein the first sampling descriptor operates on a different segment of continuous time than the second sampling descriptor, which predicted channel gain enables channel handling.

12. The radio network node according to claim 11, wherein the radio network node is further configured to predict the channel gain by obtaining the channel gain by a linear prediction, wherein the radio network node is configured to multiply a continuous time estimated parameter vector with a regression vector obtained using the first and second sampling descriptors reflecting varying sampling periods.

13. The radio network node according to claim 11, further being configured to predict channel gain at time t, (t), from
{circumflex over (y)}(t).sup.T(t){circumflex over ()}(tk.sub.2h)+c(t) where (t) is a regression vector at time t {circumflex over ()}(tk.sub.2h) is a channel estimate at a time taking the second sampling descriptor into account; and c(t) is a parameter independent part of the prediction.

14. The radio network node according to claim 13, wherein the regression vector (t) is defined as
(t)=(k.sub.2h(y(tk.sub.2h)y(t(k.sub.1+k.sub.2)h))k.sub.1k.sub.2h.sup.2y(t(k.sub.1+k.sub.2)h)).sup.T where h is a fundamental sampling period; and y is a measured channel gain.

15. The radio network node according to claim 13, wherein the parameter independent part of the prediction c(t) is defined as $c\left(t\right)=y\left(t-k_{2}h\right)+\frac{k_2}{k_1}\left(y\left(t-k_{2}h\right)-y\left(t-\left(k_1+k_2\right)h\right)\right)$ where h is a fundamental sampling period; and y is a measured channel gain.

16. The radio network node according to claim 11, wherein the channel is defined in terms of a parameter vector, being a continuous time estimated parameter vector, as
=(a.sub.1a.sub.2).sup.T where a.sub.1 a.sub.2 are the continuous time parameters.

17. The radio network node according to claim 11, further being configured to: perform a channel estimation using the predicted channel gain.

18. The radio network node according to claim 11, further being configured to, when the sampling rate being below a threshold rate, feed channel samples into a multirate predictor performing the prediction.

19. The radio network node according to claim 18, further being configured to, when the sampling rate being equal or above the threshold rate, average the channel samples instead to obtain an averaged channel gain estimate.

20. The radio network node according to claim 11, further being configured to; calculate an average of the channel gain by filtering channel samples; calculate a first magnitude of difference between received channel samples and the calculated average of the channel gain; calculate a second magnitude of difference between the received channel samples and the predicted channel gain; and select the predicted channel gain or the calculated average of the channel gain which ever gives a minimum expected error based on the calculated first and second magnitudes.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Features and advantages of the embodiments will become readily apparent to those skilled in the art by the following detailed description of exemplary embodiments thereof with reference to the accompanying drawings, wherein:

(2) FIG. 1 shows a channel gain process and an uplink scheduling process;

(3) FIG. 2 shows a graph depicting the delay spread of a signal;

(4) FIG. 3 shows the channel gain with a Doppler effect;

(5) FIG. 4 shows an SINR estimation error for three different methods for channel estimation;

(6) FIG. 5 shows an overview depicting a wireless communication network according to embodiments herein;

(7) FIG. 6 shows a combined flowchart and signaling scheme according to embodiments herein;

(8) FIG. 7 shows non-uniformly sampled data, on average 1/25.sup.th of all samples are used, i.e. the average sampling rate is 40 Hz;

(9) FIG. 8 shows a graph depicting prediction error;

(10) FIG. 9 shows a link adaptation apparatus according to embodiments herein;

(11) FIG. 10 shows a channel estimation process according to embodiments herein;

(12) FIG. 11 shows a flowchart depicting a method performed by a radio network node according to embodiments herein; and

(13) FIG. 12 shows a block diagram depicting a radio network node according to embodiments herein.

DETAILED DESCRIPTION

(14) The figures are schematic and simplified for clarity, and they merely show details which are essential to the understanding of the embodiments presented herein, while other details have been left out. Throughout the disclosure, the same reference numerals are used for identical or corresponding parts or actions.

(15) Embodiments herein relate to wireless communication networks in general. FIG. 5 is a schematic overview depicting a wireless communication network 1. The wireless communication network 1 may for example be a network such as a Long-Term Evolution (LTE), e.g. LTE Frequency Division Duplex (FDD), LTE Time Division Duplex (TDD), LTE Half-Duplex Frequency Division Duplex (HD-FDD), LTE operating in an unlicensed band, Wideband Code Division Multiple Access (WCDMA), Universal Terrestrial Radio Access (UTRA) TDD, Global System for Mobile communications (GSM) network, GSM/Enhanced Data Rate for GSM Evolution (EDGE) Radio Access Network (GERAN) network, Ultra-Mobile Broadband (UMB), EDGE network, network comprising of any combination of Radio Access Technologies (RATs) such as e.g. Multi-Standard Radio (MSR) base stations, multi-RAT base stations etc., any 3rd Generation Partnership Project (3GPP) cellular network, WFi networks, Worldwide Interoperability for Microwave Access (WiMax), 5G system or any cellular network or system. Thus, although terminology from 3GPP LTE may be used in this disclosure to exemplify embodiments herein, this should not be seen as limiting the scope of the embodiments herein to only the aforementioned system.

(16) In the wireless communication network 1, a wireless device 10, also known as a mobile station, a user equipment and/or a wireless terminal, communicates via a Radio Access Network (RAN) to one or more core networks (CN). It should be understood by the skilled in the art that wireless device is a non-limiting term which means any wireless terminal, user equipment, Machine Type Communication (MTC) device, a Device to Device (D2D) terminal, or node e.g. Personal Digital Assistant (PDA), laptop, mobile phone, sensor, relay, mobile tablets or even a small base station communicating within respective cell.

(17) The wireless communication network 1 covers a geographical area which is divided into cell areas, e.g. a cell 11 being served by a radio network node 12. The radio network node 12 may be a radio base station e.g. a NodeB, an evolved Node B (eNB, eNode B), a base transceiver station, an Access Point Base Station, a base station router, a WI-FI access point, or any other network unit capable of communicating with a wireless device within the cell served by the radio network node depending e.g. on the radio access technology and terminology used. The radio network node 12 may serve one or more cells or areas, such as the cell 11.

(18) A cell is a geographical area where radio coverage is provided by radio equipment at e.g. a base station site or at remote locations in Remote Radio Units (RRU). The cell definition may also incorporate frequency bands and radio access technology used for transmissions, which means that two different cells may cover the same geographical area but using different frequency bands. Each cell is identified by an identity within the local radio area, which is broadcast in the cell. Another identity identifying the cell 11 uniquely in the whole wireless communication network 1 may also be broadcasted in the cell 11. The radio network node 12 communicates over a radio interface, also referred to as air interface, operating on radio frequencies with the wireless device 10 within range of the radio network node 12. The wireless device 10 transmits data over the radio interface to the radio network node 12 in Uplink (UL) transmissions and the radio network node 12 transmits data over the radio interface to the wireless device 10 in Downlink (DL) transmissions.

(19) Embodiments herein disclose a channel gain prediction method and in some embodiments a corresponding channel estimator and link adaptor taking the prediction into account, which channel gain prediction automatically handles multiple and varying sampling rates. The channel estimation using the channel gain prediction produces the same parameter values, irrespective of the sampling rate applied, a fact that makes optimal channel gain prediction straightforward. The computational complexity is low and is similar as for the algorithm described by equation (eq. 9) above.

(20) A Multi-Rate Channel Prediction Method

(21) A problem indicated above with the complexity associated with the multiple sampling rates, is herein solved by providing the new channel gain prediction algorithm i.e. a channel gain prediction model with one or more of the following distinguishing features, The parameters of the channel gain prediction model are in continuous time, so called continuous time parameters. A regression vector of the channel gain prediction model reflects the time varying actual sampling period. The continuous time parameters of the channel gain prediction model may be estimated on-line, typically with e.g. a new recursive least squares algorithm. A prediction of the channel gain, e.g. complex amplitude or power, is obtained by e.g. a linear prediction, where a continuous time estimated parameter vector is multiplied with a regression vector that reflects the varying sampling period.

(22) As stated in the prior art section, the Doppler effect of the channel can be expressed in the frequency domain as a power spectrum, where a highest Doppler frequency corresponds to the speed of the wireless device 10. To model this spectrum the following continuous model may be used

(23) $\begin{array}{cc}y\left(t\right)=\frac{1}{A\left(p\right)}e\left(t\right)& \left(\mathrm{eq}.10\right)\\ A\left(p\right)=p^n+a_1{p}^{n-1}+\mathrm{.Math.}+a_n& \left(\mathrm{eq}.11\right)\\ \mathrm{py}\left(t\right)=\frac{\mathrm{dy}\left(t\right)}{\mathrm{dt}}& \left(\mathrm{eq}.12\right)\end{array}$

(24) Here p denotes a differentiation operator and a.sub.i, i=1, . . . , n are the continuous time parameters. y(t) denotes the output, either complex channel amplitude or power. Here A(p) is the spectral polynomial, that defines the Doppler spectrum in (eq. 10), and where the (eq. 12) is two equivalent ways of expressing a time derivative of the signal y(t).

(25) The measurements are the channel output, e.g. the channel output is here defined to be either the real part of the complex channel, the imaginary part of the complex channel, or the power of the channel, i.e. the sum of the squared real and imaginary parts, at the uneven sampling instances, i.e.
y(t.sub.0),y(t.sub.0+k.sub.1h),y(t.sub.0+(k.sub.1+k.sub.2)h), . . . y(t) (eq. 13)

(26) Here a fundamental sampling period, for VoLTE this is the 1 ms TTI, is given by h, while k.sub.1 and k.sub.2 are integers that model the momentary sampling period applied for e.g. VoLTE. k.sub.1 and k.sub.2 are also referred to herein as a first sampling descriptor indicating a first momentary sampling frequency and a second sampling descriptor indicating a second momentary sampling frequency,

(27) The next step is to replace the differentiation operator p of (eq. 10)-(eq. 12) with sequential approximations. Since the intention here is to obtain a low computational complexity, and since simulations have shown that an order of n=2 is sufficient, this approximation is illustrated for order 2. The extension to higher orders follows the same method, and embodiments should therefore not be limited to orders less than or equal to 2.

(28) To begin, it holds that

(29) $\begin{array}{cc}\mathrm{py}\left(t_0\right)\frac{q^{k_1}-1}{k_{1}h}y\left(t_0\right)=\frac{y\left(t_0+k_{1}h\right)-y\left(t_0\right)}{k_{1}h}& \left(\mathrm{eq}.14\right)\end{array}$

(30) where a shift operator q shifts the time one fundamental sampling period h ahead in time. Proceeding in this way results in

(31) $\begin{array}{cc}\mathrm{py}\left(t_0+k_{1}h\right)\frac{q^{k_2}-1}{k_{2}h}y\left(t_0+k_{1}h\right)=\frac{y\left(t_0+\left(k_1+k_2\right)h\right)-y\left(t_0+k_{1}h\right)}{k_{2}h}& \left(\mathrm{eq}.15\right)\\ p^2\left(y\left(t_0\right)\right)=p\left(\mathrm{py}\left(t_0\right)\right)\frac{q^{k_1}-1}{k_{1}h}\left(\mathrm{py}\left(t_0\right)\right)\frac{\mathrm{py}\left(t_0+k_{1}h\right)-\mathrm{py}\left(t_0\right)}{k_{1}h}\frac{\frac{q^{k_2}-1}{k_{2}h}{q}^{k_1}y\left(t_0\right)-\frac{q^{k_1}-1}{k_{1}h}y\left(t_0\right)}{k_{1}h}=\frac{1}{h^2}\left(\frac{1}{k_1{k}_2}{q}^{k_1+k_2}-\left(\frac{1}{k_1{k}_2}+\frac{1}{k_1^2}\right)q^{k_1}+\frac{1}{k_1^2}\right)y\left(t_0\right)& \left(\mathrm{eq}.16\right)\end{array}$

(32) It can be noted that the choice k.sub.1=k.sub.2=1 results in the familiar three point approximation of the second derivative of a signal.

(33) To obtain a discrete time model, from (eq. 10)-(eq. 12), the following approximations are introduced
py(t)py(t.sub.0)(eq. 17)
p.sup.2y(t)p.sup.2y(t.sub.0)(eq. 18)
These approximations state that the first and second derivatives at time r can be well approximated at time t.sub.0. This is reasonable for low Doppler frequencies.

(34) Employing (eq. 17) and (eq. 18) in (eq. 10), multiplying the resulting equation by q.sup.k.sup.1.sup.+k.sup.2, and using the fact that in the case n=2 (eq. 13) implies q.sup.k.sup.1.sup.+k.sup.2y(t.sub.0)=y(t), q.sup.k.sup.1y(t.sub.0)=q.sup.k.sup.2y(t), y(t.sub.0)=q.sup.(k.sup.1.sup.+k.sup.2)y(t), then results in the filter equation

(35) $\begin{array}{cc}y\left(t\right)=-a_1{k}_{2}h\left(y\left(t-k_{2}h\right)-y\left(t-\left(k_1+k_2\right)h\right)\right)-a_2{k}_1{k}_2{h}^{2}y\left(t-\left(k_1+k_2\right)h\right)+y\left(t-k_{2}h\right)+\frac{k_2}{k_1}\left(y\left(t-k_{2}h\right)-y\left(t-\left(k_1+k_2\right)h\right)\right)+k_1{k}_2{h}^{2}e\left(t\right)& \left(\mathrm{eq}.19\right)\end{array}$

(36) The final step in the derivation of the discrete time model is then to write equation (eq. 19) in a linear regression form as

(37) $\begin{array}{cc}y\left(t\right)={}^T\left(t\right)+c\left(t\right)+v\left(t\right)& \left(\mathrm{eq}.20\right)\\ ={\left(\begin{array}{cc}{a}_1& a_2\end{array}\right)}^T& \left(\mathrm{eq}.21\right)\\ \left(t\right)={\left(-k_{2}h\left(y\left(t-k_{2}h\right)-y\left(t-\left(k_1+k_2\right)h\right)\right)-k_1{k}_2{h}^{2}y\left(t-\left(k_1+k_2\right)h\right)\right)}^T& \left(\mathrm{eq}.22\right)\\ c\left(t\right)=y\left(t-k_{2}h\right)+\frac{k_2}{k_1}\left(y\left(t-k_{2}h\right)-y\left(t-\left(k_1+k_2\right)h\right)\right)& \left(\mathrm{eq}.23\right)\\ v\left(t\right)=k_1{k}_2{h}^{2}e\left(t\right)& \left(\mathrm{eq}.24\right)\end{array}$

(38) The equations (eq. 20)-(eq. 24) are now directly suitable for prediction and on-line estimation. It can be noted that the estimation algorithm (eq. 20)-(eq. 24) will include the prediction (eq. 20) as one step.

(39) Embodiments may use a so called recursive least squares algorithm. However, it should be noted that other alternatives exist and embodiments should not be limited to the use of the recursive least squares algorithm. The recursive least squares algorithm follows from standard results in the literature of estimation. The result is

(40) $\begin{array}{cc}K\left(t\right)=\frac{P\left(t\right)\left(t\right)}{+{}^T\left(t\right)P\left(t\right)\left(t\right)}& \left(\mathrm{eq}.25\right)\\ \hat{y}\left(t\right)={}^T\left(t\right)\hat{}\left(t-k_{2}h\right)+c\left(t\right)& \left(\mathrm{eq}.26\right)\\ \hat{}\left(t\right)=\hat{}\left(t-k_{2}h\right)+K\left(t\right)\left(y\left(t\right)-\hat{y}\left(t\right)\right)& \left(\mathrm{eq}.27\right)\\ P\left(t\right)=\frac{\left(P\left(t-k_{2}h\right)-P\left(t-k_{2}h\right)\left(t\right){}^T\left(t\right)P\left(t-k_{2}h\right)\right)}{\left(+{}^T\left(t\right)P\left(t-k_{2}h\right)\left(t\right)\right)}& \left(\mathrm{eq}.28\right)\end{array}$

(41) Above, (eq. 25) computes an update gain K(t) in terms of a covariance matrix P(t), the regression vector (t) (eq. 22) and a forgetting factor . The channel prediction, (t), is then computed in (eq. 26) by vector multiplication of the estimated channel parameters {circumflex over ()}(tk.sub.2h) of the previous steps, i.e. (eq. 22) and (eq. 23). Using the last measurement y(t) the new channel estimate is then updated in (eq. 27). Finally, the covariance matrix P(t) is updated in (eq. 28).

(42) This completes the description of the algorithm for adaptive channel gain prediction.

(43) FIG. 6 is a combined flowchart and signaling scheme according to some embodiments herein.

(44) Action 601. The wireless device 10 transmits a signal over a channel to the radio network node 12.

(45) Action 602. The radio network node 12 predicts a channel gain for the channel using a first sampling descriptor k, and a second sampling descriptor k.sub.2. k.sub.1 indicates a first momentary sampling frequency and k.sub.2 indicates a second momentary sampling frequency. The first sampling descriptor k.sub.1 operates on a different segment of continuous time than the second sampling descriptor k.sub.2, hence, these descriptors reflect varying sampling periods. The predicted channel gain enables channel handling such as channel estimation and/or link adaptation as the predicted channel gain is used in these processes.

(46) Action 603. The radio network node 12 may then use the predicted channel gain in a channel estimation e.g. making an SINR estimation.

(47) Action 604. The radio network node 12 may use the channel estimation when performing a link adaptation and hence the radio network node 12 may perform link adaptation based on the predicted channel gain.

(48) To illustrate the performance of embodiments herein, a case with a Doppler frequency of 5 Hz was selected. The sampling rate was random. The data is shown in FIG. 7. FIG. 7 shows non-uniformly sampled data, on average 1/25.sup.th of all samples are used, i.e. the average sampling rate is 40 Hz. Time is defined along a horizontal axis and power with noise [normalized] is defined along a vertical axis.

(49) FIG. 8. The predicted channel gain and the actual channel gain appear in the top figure, while the prediction error is shown in the bottom picture. Time is defined along a horizontal axis and measured and predicted power [normalized] is defined along a vertical axis in the top figure. Time is defined along a horizontal axis and prediction error [normalized] is defined along a vertical axis in the bottom figure. There are some spikes in the predicted errors which are due to channel samples taken too close to one another.

(50) Channel Measurement Embodiment

(51) In some embodiments herein several instances of the algorithm defined by (eq. 20)-(eq. 28) are run. All quantities of the two algorithms are independent. A first instance of the algorithm represents the real part of the complex channel gain and it is driven by measurements of this quantity. A second instance of the algorithm represents the imaginary part of the complex channel and it is driven by measurements of this quantity. Note that in this case the complex channel differs between antenna elements, hence two instances of (eq. 20)-(eq. 28), i.e. two channel gain predictors, are needed per antenna element. This embodiment gives a better result being more accurate than when using merely one channel gain predictor.

(52) Power Measurement Embodiment

(53) In this embodiment, one instance, one channel gain predictor, of the algorithm (eq. 20)-(eq. 28) is run. The algorithm represents the channel power and is driven by measurements of this quantity. Only one channel gain predictor for all antenna elements is used leading to a non-complex solution but with less accurate results than the one above.

(54) FIG. 9 is a block diagram depicting a link adaptation apparatus being an example of the radio network node 12. The wireless device 10 transmits signals to the radio network node 12. E.g. UL physical layer (UL PHY) in the wireless device 10 provides the radio network node 12 with complex channel coefficients and allocation size to a channel gain filter 901 in the radio network node 12. Furthermore, an interference measurement is provided to a SINR estimator 902 at the radio network node 12 from the UL PHY. A UL Medium Access Control layer at the wireless device 10 may provide power headroom reports (PHR) to the SINR estimator 902. This PHR may also be provided to the channel gain filter 901. In the channel gain filter 901 the methods described above is performed resulting in a predicted channel gain. This predicted channel gain is then fed to the SINR estimator 902. The SINR estimator 902 performs the channel estimation resulting in SINR for available PRBs. These available PRBs are provided from a Link Adaptor 903. The link adaptor 903 obtains the SINR from the SINR estimator 902 and also packet size from within the radio network node 12. The link adaptor 902 then determines TBS, MCS, and selected PRBs for transmission based on the estimated SINR and thus the predicted channel gain.

(55) In some embodiments herein certain safety measures may be applied to the methods and apparatuses to e.g. avoid the spikes in FIG. 8. These safety measures are acting as safety nets, in order to provide a solution with enhanced and robust performance avoiding making poor channel gain predictions.

(56) FIG. 10 depicts an example of a channel gain predictor arrangement and the method of which may be implemented as a part of the channel gain filter 901 in FIG. 9.

(57) Channel samples are fed into the radio network node 12. The radio network node 12 may update a rate limitation or sampling rate limitation at a rate limiting module 1001. As mentioned above the sampling of the channel becomes non-uniform when retransmissions and segmentation are used (see table I). A typical scenario is that a VoLTE packet is segmented into two segments and transmitted in two consecutive TTIs and the next transmission is performed 20 ms or 40 ms later. The transmissions will result in channel measurements, and due to the short time between the segmented transmissions any measurement noise will be amplified when predicting the channel quality e.g. 40 ms later. A method to avoid this noise amplification is to limit the rate at which measurements are fed into one or more multirate channel gain predictor(s) 1002. This limitation can be done in some different ways, e.g. a. Only feed channel samples to the multirate channel gain predictor when the time since the latest sample is longer than a threshold t.sub.th1, indicating a threshold rate. Channel samples too close one another in time gives a large prediction error, see FIG. 8. b. Average channel samples that are received within a certain time window and only feed the averaged channel samples to the multirate predictor.

(58) In case the measurements are fed to the multirate channel gain predictor 1002, the multirate channel gain predictor 1002 predicts channel gain according to embodiments herein.

(59) The radio network node 12 may further fall back to the average the channel gain when appropriate. When the Doppler spread for the wireless device 10 is high or the time between uplink transmissions is large the channel autocorrelation approaches zero. In these cases it is preferable to fall back to an average of the channel gain of the channel samples. This averaged channel gain may be computed in a calculator 1003.

(60) The radio network node 12 may further comprise a channel gain selector 1004 for selecting the predicted channel gain or the averaged channel gain. A criterion for using the averaged channel gain compared to the predicted channel gain may be based on an estimated error from the two different methods according to the selecting process described in actions 1010-1013.

(61) Action 1010. The radio network node 12, e.g. the calculator 1003, may calculate average of the channel gain by filtering the channel samples.

(62) Action 1011. The radio network node 12, e.g. the channel gain selector 1004, may calculate the magnitude of the difference, in dB, between received channel samples, i.e. an actual channel gain of previous channel samples, and the averaged channel gain.

(63) Action 1012. The radio network node 12, e.g. the channel gain selector 1004, may calculate the magnitude of the difference, in dB, between received channel samples, i.e. an actual channel gain of previous channel samples, and the predicted channel gain.

(64) Action 1013. The radio network node 12, e.g. the channel gain selector 1004, may then select the method, prediction or average, that gives the minimum expected error based on the results from actions 1011 and 1012 above.

(65) The output from action 1013 may then be the channel gain prediction that is used to calculate the SINR used for Link Adaptation of uplink transmissions.

(66) The method actions in the radio network node 12 for enabling channel handling of a channel between the wireless device 10 and the radio network node 12 in the wireless communication network 1 according to some embodiments will now be described with reference to a flowchart depicted in FIG. 11. Channel handling covers e.g. channel estimation and link adaptation. The channel is defined in continuous time and a sampling rate of the channel is non-uniform e.g. the sampling rate covers different sampling periods such as the five measurement rates of the different states in Table I above. Actions performed in some embodiments are marked with dashed boxes.

(67) Action 1100. The radio network node 12 may feed, when the sampling rate being below a threshold rate, the channel samples into a multirate predictor performing the prediction. Hence, the radio network node 12 limits the rate of sampling to be below the threshold rate. When the sampling rate is equal or above the threshold rate, the radio network node 12 may average the channel samples instead to obtain an averaged channel gain estimate. I.e. when a condition is fulfilled, the radio network node 12 may average the channel gain of the channel samples instead to obtain an averaged channel gain estimate.

(68) Action 1101. The radio network node 12 predicts the channel gain using a first sampling descriptor indicating a first momentary sampling frequency and a second sampling descriptor indicating a second momentary sampling frequency. The first sampling descriptor operates on a different segment of continuous time than the second sampling descriptor. The predicted channel gain enables channel handling. The radio network node 12 may predict the channel gain by obtaining the channel gain by a linear prediction, wherein a continuous time estimated parameter vector is multiplied with a regression vector obtained using the first and second sampling descriptors reflecting varying sampling periods.

(69) The predicted channel gain at time t, (t), is defined by
{circumflex over (y)}(t).sup.T(t){circumflex over ()}(tk.sub.2h)+c(t)

(70) where (t) is the regression vector at time t; {circumflex over ()}(tk.sub.2h) is a channel estimate at a time taking the second sampling descriptor into account; and c(t) is a parameter independent part of the prediction.

(71) The regression vector (t) may be defined as
(t)=(k.sub.2h(y(tk.sub.2h)y(t(k.sub.1+k.sub.2)h))k.sub.1k.sub.2h.sup.2y(t(k.sub.1+k.sub.2)h)).sup.T

(72) where h is a fundamental sampling period; and y is a measured channel gain.

(73) The parameter independent part of the prediction, c(t), may be defined as

(74) 0$c\left(t\right)=y\left(t-k_{2}h\right)+\frac{k_2}{k_1}\left(y\left(t-k_{2}h\right)-y\left(t-\left(k_1+k_2\right)h\right)\right)$

(75) where h is a fundamental sampling period; and y is a measured channel gain.

(76) The channel may be defined in terms of a parameter vector, being a continuous time estimated parameter vector, as
=(a.sub.1a.sub.2).sup.T

(77) where a.sub.1 a.sub.2 are the continuous time parameters.

(78) The channel may be supporting segmenting and retransmission of packets, and may additionally or alternatively support discontinuous reception. The channel may be for carrying VoLTE packets.

(79) Action 1102. The radio network node 12 may then perform a channel estimation using the predicted channel gain. E.g. action 11021, the radio network node 12 may measure power over the channel for channel estimation or, action 11022, the radio network node 12 may measure a real part of a complex channel gain and an imaginary part of the complex channel gain for channel estimation. In order to select a channel gain to use in the channel estimation the radio network node 12 may calculate the average of the channel gain by filtering channel samples; calculate a first magnitude of difference between received channel samples and the calculated average of the channel gain; calculate a second magnitude of difference between the received channel samples and the predicted channel gain; and select the predicted channel gain or the calculated average of the channel gain which ever gives a minimum expected error based on the calculated first and second magnitudes.

(80) Action 1103. The radio network node 12 may further perform a link adaptation using the predicted channel gain.

(81) Embodiments herein provide a solution where the predicted channel gain is closer to the actual channel gain resulting in a VoLTE capacity increasemore users per cell; VoLTE performance enhancementbetter audio quality for users; Low implementation complexityone or two multirate predictors per wireless device; and/or High channel tracking bandwidthall data fused by one or two multirate predictors.

(82) In order to perform the methods disclosed herein a radio network node 12 is provided. FIG. 12 is a block diagram depicting the radio network node 12 for enabling channel handling of a channel between the wireless device 10 and the radio network node 12 in the wireless communication network 1. The channel is defined in continuous time and a sampling rate of the channel is non-uniform.

(83) The radio network node 12 is configured to predict a channel gain using a first sampling descriptor indicating a first momentary sampling frequency and a second sampling descriptor indicating a second momentary sampling frequency. The first sampling descriptor operates on a different segment of continuous time than the second sampling descriptor, which predicted channel gain enables channel handling. The radio network node 12 may further be configured to predict the channel gain by obtaining the channel gain by a linear prediction, wherein the radio network node 12 is configured to multiply a continuous time estimated parameter vector with a regression vector obtained using the first and second sampling descriptors reflecting varying sampling periods. The radio network node 12 may further be configured to predict channel gain at time t, (t), from
{circumflex over (y)}(t).sup.T(t){circumflex over ()}(tk.sub.2h)+c(t) where (t) is a regression vector at time t {circumflex over ()}(tk.sub.2h) is a channel estimate at a time taking the second sampling descriptor into account; and c(t) is the parameter independent part of the prediction.

(84) The regression vector (t) may be defined as
(t)=(k.sub.2h(y(tk.sub.2h)y(t(k.sub.1+k.sub.2)h))k.sub.1k.sub.2h.sup.2y(t(k.sub.1+k.sub.2)h)).sup.T where h is a fundamental sampling period and y is a measured channel gain.

(85) The parameter independent part of the prediction, c(t) may be defined as

(86) $c\left(t\right)=y\left(t-k_{2}h\right)+\frac{k_2}{k_1}\left(y\left(t-k_{2}h\right)-y\left(t-\left(k_1+k_2\right)h\right)\right)$

(87) where h is a fundamental sampling period; and y is a measured channel gain.

(88) The channel may be defined in terms of a parameter vector, being a continuous time estimated parameter vector, as
=(a.sub.1a.sub.2).sup.T

(89) where a.sub.1 a.sub.2 are the continuous time parameters.

(90) The channel may be supporting segmenting and retransmission of packets, and may additionally or alternatively support discontinuous reception. The channel may be for carrying VoLTE packets.

(91) The radio network node 12 may further be configured to perform a channel estimation using the predicted channel gain. Then, the radio network node 12 may further be configured to measure power over the channel for channel estimation, or the radio network node 12 may further be configured to measure a real part of a complex channel gain and an imaginary part of the complex channel gain for channel estimation.

(92) The radio network node 12 may further be configured to perform a link adaptation using the predicted channel gain.

(93) The radio network node 12 may further be configured to, when the sampling rate being below a threshold rate, feed channel samples into a multirate predictor performing the prediction. The radio network node 12 may further be configured to, when the sampling rate being equal or above the threshold rate, average the channel samples instead to obtain an averaged channel gain estimate.

(94) The radio network node 12 may further be configured to calculate an average of the channel gain by filtering channel samples; calculate a first magnitude of difference between received channel samples and the calculated average of the channel gain; calculate a second magnitude of difference between the received channel samples and the predicted channel gain; and to select the predicted channel gain or the calculated average of the channel gain which ever gives a minimum expected error based on the calculated first and second magnitudes.

(95) The radio network node 12 may comprise processing circuitry 1201. The radio network node 12 may further comprise a predicting module 1202, e.g. the channel gain filter 901. The predicting module 1202 and/or the processing circuitry 1201 may be configured to predict a channel gain using a first sampling descriptor indicating a first momentary sampling frequency and a second sampling descriptor indicating a second momentary sampling frequency. The first sampling descriptor operates on a different segment of continuous time than the second sampling descriptor, which predicted channel gain enables channel handling. The predicting module 1202 and/or the processing circuitry 1201 may further be configured to predict the channel gain by obtaining the channel gain by a linear prediction, wherein the predicting module 1202 and/or the processing circuitry 1201 is configured to multiply a continuous time estimated parameter vector with a regression vector obtained using the first and second sampling descriptors reflecting varying sampling periods. The predicting module 1202 and/or the processing circuitry 1201 may further be configured to predict channel gain at time t, (t), from
{circumflex over (y)}(t).sup.T(t){circumflex over ()}(tk.sub.2h)+c(t) where (t) is a regression vector at time t {circumflex over ()}(tk.sub.2h) is a channel estimate at a time taking the second sampling descriptor into account; and c(t) is the parameter independent part of the prediction.

(96) The regression vector (t) may be defined as
(t)=(k.sub.2h(y(tk.sub.2h)y(t(k.sub.1+k.sub.2)h))k.sub.1k.sub.2h.sup.2y(t(k.sub.1+k.sub.2)h)).sup.T where h is a fundamental sampling period and y is a measured channel gain.

(97) The parameter independent part of the prediction, c(t) may be defined as

(98) $c\left(t\right)=y\left(t-k_{2}h\right)+\frac{k_2}{k_1}\left(y\left(t-k_{2}h\right)-y\left(t-\left(k_1+k_2\right)h\right)\right)$

(99) where h is a fundamental sampling period; and y is a measured channel gain.

(100) The channel may be defined in terms of a parameter vector, being a continuous time estimated parameter vector, as
=(a.sub.1a.sub.2).sup.T

(101) where a.sub.1 a.sub.2 are the continuous time parameters.

(102) The channel may be supporting segmenting and retransmission of packets, and may additionally or alternatively support discontinuous reception. The channel may be for carrying VoLTE packets.

(103) Furthermore, the radio network node 12 may comprise a channel estimating module 1203, e.g. the SINR estimator 902. The channel estimating module 1203 and/or the processing circuitry 1201 may be configured to perform a channel estimation using the predicted channel gain. Then, the channel estimating module 1203 and/or the processing circuitry 1201 may further be configured to measure power over the channel for channel estimation, or the channel estimating module 1203 and/or the processing circuitry 1201 may further be configured to measure a real part of a complex channel gain and an imaginary part of the complex channel gain for channel estimation.

(104) The radio network node 12 may comprise a link adaptation module 1204, e.g. a scheduler and/or the link adaptor 903. The link adaptation module 1204 and/or the processing circuitry 1201 may be configured to perform a link adaptation using the predicted channel gain.

(105) The radio network node 12 may comprise a feeding module 1205. The feeding module 1205 and/or the processing circuitry 1201 may further be configured to, when the sampling rate is below a threshold rate, feed channel samples into a multirate predictor performing the prediction. The radio network node 12 may comprise an averaging module 1206. The averaging module 1206 and/or the processing circuitry 1201 may further be configured to, when the sampling rate is equal or above the threshold rate, average the channel samples instead to obtain an averaged channel gain estimate.

(106) The radio network node 12 may further comprise a calculating module 1207 and a selecting module 1208. The calculating module 1207 and/or the processing module 1201 may be configured to calculate an average of the channel gain by filtering channel samples, and to calculate a first magnitude of difference between received channel samples and the calculated average of the channel gain. The calculating module 1207 and/or the processing module 1201 may further be configured to calculate a second magnitude of difference between the received channel samples and the predicted channel gain. The selecting module 1208 and/or the processing circuitry 1201 may be configured to select the predicted channel gain or the calculated average of the channel gain which ever gives a minimum expected error based on the calculated first and second magnitudes.

(107) The radio network node 12 further comprises a memory 1209. The memory 1209 comprises one or more units to be used to store data on, such as channel gains, predicted channel gains, channel estimations, SINRs, averaged channel gains, link adaptation values, applications to perform the methods disclosed herein when being executed, and similar.

(108) The methods according to the embodiments described herein for the radio network node 12 may be implemented by means of e.g. a computer program 1210 or a computer program product, comprising instructions, i.e., software code portions, which, when executed on at least one processor, cause the at least one processor to carry out the actions described herein, as performed by the radio network node 12. The computer program 1210 may be stored on a computer-readable storage medium 1211, e.g. a disc or similar. The computer-readable storage medium 1211, having stored thereon the computer program, may comprise the instructions which, when executed on at least one processor, cause the at least one processor to carry out the actions described herein, as performed by the radio network node 12. In some embodiments, the computer-readable storage medium may be a non-transitory computer-readable storage medium.

(109) As will be readily understood by those familiar with communications design, that functions means or modules may be implemented using digital logic and/or one or more microcontrollers, microprocessors, or other digital hardware. In some embodiments, several or all of the various functions may be implemented together, such as in a single application-specific integrated circuit (ASIC), or in two or more separate devices with appropriate hardware and/or software interfaces between them. Several of the functions may be implemented on a processor shared with other functional components of radio access network node, for example.

(110) Alternatively, several of the functional elements of the processing means discussed may be provided through the use of dedicated hardware, while others are provided with hardware for executing software, in association with the appropriate software or firmware. Thus, the term processor or controller as used herein does not exclusively refer to hardware capable of executing software and may implicitly include, without limitation, digital signal processor (DSP) hardware, read-only memory (ROM) for storing software, random-access memory for storing software and/or program or application data, and non-volatile memory. Other hardware, conventional and/or custom, may also be included. Designers of communications receivers will appreciate the cost, performance, and maintenance tradeoffs inherent in these design choices.

(111) It will be appreciated that the foregoing description and the accompanying drawings represent non-limiting examples of the methods and radio node taught herein. As such, the radio node and techniques taught herein are not limited by the foregoing description and accompanying drawings. Instead, the embodiments herein are limited only by the following claims and their legal equivalents.