Determining similarity between channel estimations

TECHNICAL FIELD

The present disclosure relates to determining similarity between channel estimations, for example in order to optimize a present reference signal configuration.

BACKGROUND

Employing a reference signal (RS) is a fundamental technique in wireless networks, where an RS is sent as a predefined sequence of bits to support different main functions. In 4G networks, there are cell-specific reference signals which always are enabled, while in the 5G networks reference signals are user-specific and more configurable. In 5Gtechnology, a demodulation reference signal (DMRS) is sent along with payload data for channel estimation purpose at the receiving device to aid it in demodulating the payload data. Since the properties of the radio channel can vary rapidly, such as moving user equipment (UE), the UE and/or a base station can be configured to employ the reference signals with higher density in the time-domain for a more recent channel estimation. For example, for 5G technology, up to four DMRSs can be used per time slot.

To support the scenarios with a fast changing channel, in 5G technology it is possible to configure more frequent reference signals for having more accurate channel estimation. However, reference signals add to the overhead and consequently reduce the number of available resource elements for user data allocation. Accordingly, it is better to reconfigure the RS configuration based on how fast the channel changes, otherwise, using the fixed RS configuration under all channel conditions, can result in reducing the peak rate. Hence, an adaptive reconfiguration is desirable to change the RS configuration based on the channel changes to achieve a peak rate.

US 2017311188 discloses a base station that can determine a Doppler metric associated with a wireless channel and UE where the Doppler metric is used to select one or both of a reference signal density and a channel estimation technique for the wireless channel and associated UE.

However, in US 2017311188 no Doppler metric is defined, and it is difficult to define a Doppler metric since the Doppler shift is difficult to estimate alone in the practical scenario as frequency deviation also can be due to frequency error in the transceiver. In addition, the Doppler spread in non-line of sight channels will also impact the Doppler shift, while this is not significant in the line-of-sight channels.

There are many other applications that can take advantage of an estimation of how and when a channel changes. For example, the statistics of how and when a channel changes can be used for cell planning. Knowledge of how and when a channel changes can furthermore be used when selecting a specific frequency band by comparing estimated channel properties in different frequency bands.

It is therefore desired to provide a means and a method for estimating how and when a channel changes where limited information regarding channel properties is used. The result can for example be used for optimizing a present reference signal configuration, but can also be applied to other applications.

SUMMARY

It is an object of the present disclosure to provide means and a method for estimating how and when a channel changes using only limited information regarding channel properties.

This object is obtained by means of a control unit that is adapted to provide a plurality of channel estimates which are separated in time during a certain first time period and to represent each channel estimate by a corresponding complex channel vector in a complex vector space. Each channel estimate constitutes an estimated channel for a radio channel provided between at least two wireless communication nodes. The radio channel constitutes a medium via which signals transmitted between nodes are transferred. The control unit is further adapted to calculate a difference angle between two of the complex channel vectors which are separated in time during a certain second time period, forming a pair of complex channel vectors, where the difference angle is a Kahler angle.

In this way, an estimation of how and when a channel has changed can be determined without any prior knowledge of its characteristics such as, e.g., line-of-sight or non-line-of-sight, only based on available channel estimations. For example, a fast-moving node such as a UE can be configured with suitable parameters without knowing or estimating its speed.

According to some aspects, the plurality of channel estimates comprises at least three channel estimates, where the control unit is adapted to calculate the Kahler angle between at least two different pairs of complex channel vectors corresponding to the at least three channel estimates such that a plurality of Kahler angles is determined and to determine a Kahler angle metric for each Kahler angle. The control unit is further adapted to determine an average of the Kahler angle metrics for the first time period, where the average of the Kahler angle metrics forms a total metric that indicates to which degree the estimated channel has changed during the first time period.

According to some aspects, the plurality of channel estimates comprises at least three channel estimates, and where the control unit is adapted to calculate the Kahler angle between at least two different pairs of complex channel vectors corresponding to the at least three channel estimates such that a plurality of Kahler angles is determined, and to determine an average Kahler angle for the first time period. The control unit is further adapted to determine a Kahler angle metric for the average Kahler angle, where the Kahler angle metric forms a total metric that indicates to which degree the estimated channel has changed during the first time period.

This means that the estimation of how and when a channel has changed can be determined in several secure and reliable ways.

According to some aspects, the Kahler angle metric is one of:

- the Kahler angle itself,

- sines for the Kahler angle, sin 9K, or

- cosines for the Kahler angle, cos 9K.

According to some aspects, the Kahler angle metric is a monotonic function of the Kahler angle.

This means that the Kahler angle metric can be chosen to suit its further implementation.

According to some aspects, the control unit is adapted to increase a reference signal density if the total metric indicates that the estimated channel has changed to a degree that exceeds a first threshold value, and to decrease a reference signal density if the total metric indicates that the estimated channel has changed to a degree that falls below a second threshold value. Otherwise, the control unit is adapted to maintain the present reference signal density, where the reference signal density relates to a density for reference signals exchanged between at least two of the wireless communication nodes.

In this way, a measurement for channel similarity can be used to reconfigure additional reference signals only when it is necessary, focusing on using obtained channel estimations corresponding to reference signals at different instances and assess whether they are similar or dissimilar. If they are determined to be similar to a certain extent, superfluous reference signals can be de-configured. If they are not determined to be similar to a certain extent, extra reference signals can be configured at shorter time intervals.

According to some aspects, the control unit is comprised in one of the wireless communication nodes.

According to some aspects, the control unit is comprised in a remote server that is accessible for at least one of the wireless communication nodes. This means that the estimation of how and when a channel has changed can be determined can be performed both centrally and locally, wherever suitable.

This object is also obtained by means of methods, wireless communication nodes and computer program product which are associated with the above advantages.

BRIEF DESCRIPTION OF THE DRAWINGS

The present disclosure will now be described more in detail with reference to the appended drawings, where:

Figure 1 schematically illustrates a first wireless communication node and a second wireless communication node with an intermediate channel;

Figure 2A schematically illustrates a first example of channel estimates provided during a timeline;

Figure 2B schematically illustrates a second example of channel estimates provided during a timeline;

Figure 3 shows a block diagram for calculation steps;

Figure 4 schematically illustrates thresholds for a degree of estimated channel change;

Figure 5 shows a flowchart for a procedure according to the present disclosure;

Figure 6 shows a flowchart for methods according to the present disclosure;

Figure 7 schematically illustrates a control unit; and

Figure 8 schematically illustrates a computer program product.

DETAILED DESCRIPTION

Aspects of the present disclosure will now be described more fully hereinafter with reference to the accompanying drawings. The different devices, systems, computer programs and methods disclosed herein can, however, be realized in many different forms and should not be construed as being limited to the aspects set forth herein. Like numbers in the drawings refer to like elements throughout. The terminology used herein is for describing aspects of the disclosure only and is not intended to limit the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

With reference to Figure 1, there is a base station (BS) 3 and a User Equipment (UE) 4 that can be any type of suitable user terminal. The BS 3 and the UE 4 are generally constituted by wireless communication nodes 3, 4 and communicate wirelessly via a radio channel 2. In order to support the communication, using reference signals (RS) 5, 6 are sent as a predefined sequence of bits to support different main functions and can for example be in the form of cell-specific reference signals which are always enabled and user-specific reference signals that are more configurable.

In all communication, signals transmitted between nodes 3, 4 go through a medium, the radio channel 2, where the signals get distorted. To properly decode received signals, it is often necessary to know the characteristics of the channel 2. The process of quantifying the characteristics is generally referred to as channel estimation. In this, a parameterized mathematical model is used to correlate a transmitted signal and a received signal. By transmitting a known signal, usually in the form of one or more reference signals, and analyzing the subsequently received signal, it is possible to obtain estimates of the parameters of the model, these estimates constituting channel estimates. To account for effects such as attenuation and phase shifting for different frequencies, channel estimates are commonly expressed as arrays of complex numbers. For a channel that is changing in time, channel estimates obtained at one instant are likely to be different from the ones obtained at a later instant.

This means that the properties of the channel 2 can vary rapidly, for example due to a moving UE, a moving obstacle or heavy rain. In such a case, the UE 4 can be configured to employ the reference signals with higher density in the time-domain for a more recent channel estimation.

Since reference signals require system resources, and at the same time are needed to increase the transmission rate, it is desired to configure reference signals to have a sufficient but not superfluous density.

By obtaining channel estimations at different instances, and assess whether they are similar or dissimilar the reference signal density can be changed accordingly. If the channel estimations are similar, superfluous reference signals can be de-configured, decreasing reference signal density, and if the channel estimations are not similar, additional reference signals can be configured at shorter time intervals, increasing reference signal density. With reference also to Figure 2A that illustrates channel estimates provided during a timeline, the BS 3 comprises a control unit 1 that is adapted to provide a plurality of channel estimates ci, C2, C3; CN-I, CN, which are separated in time during a certain first time period ti, and to represent each channel estimate by a corresponding complex channel vector in a complex vector space. Each channel estimate ci, C2, C3; CN-I, CN constitutes an estimated channel for the radio channel 2. The aim is to estimate the channel coefficient, and this can be done using different types of channel estimators such as for example minimum mean square error (MMSE) channel estimation.

The control unit 1 is further adapted to calculate a difference angle between two of the complex channel vectors, forming a first pair of complex channel vectors. In this example, these complex channel vectors correspond to a first channel estimate ci and a second channel estimate C2, where the channel estimates ci, C2 are separated in time during a certain second time period t2a, forming a first pair of channel estimates ci, C2 that corresponds to the first pair of complex channel vectors, where the difference angle is a Kahler angle 9K.

The Kahler angle is previously known and is for example discussed in the paper K. Scharnhorst, Angles in Complex Vector Space, Acta Applicandae Mathematicae, 2001. The Kahler angle is also known under other designations such as for example the Kahler function, Kahler form, characteristic deviation, holomorphy angle and Wirtinger angle.

Generally, the second time period t2a, t2b, t2x relates to a time between two channel estimates ci, C2; C2, C3; CN-I, CN and two corresponding complex channel vectors. Figure 2B illustrates a case where the first time period ti is equal to the second time period t2, where the two channel estimates ci, C2 of the second time period t2 are the only channel estimates present in the first time period ti and results in providing only two complex channel vectors and one Kahler angle.

According to some aspects, and as illustrated in Figure 2 A, there is a plurality of second time periods t2a, t2b, t2x which are not equal to the first time period ti. There is furthermore a plurality of pairs of channel estimates ci, C2; C2, C3; CN-I, CN that correspond to a plurality of pairs of complex channel vectors where a plurality of Kahler angles are calculated, one for each pair of complex channel vectors. According to some aspects, an average is then formed for the Kahler angle 9K or for any suitable Kahler angle metric 9KM.

A Kahler angle metric 9KM can be regarded as a similarity coefficient and is associated with the Kahler angle 9K. The Kahler angle metric 9KM can for example be constituted by the Kahler angle 9K itself, sines for the Kahler angle 9K, sin 9K, which increases the more the estimated channel changes, or cosines for the Kahler angle 9K, COS 9K, which decreases the more the estimated channel changes.. Any one of cos 9K or sin 9K can thus according to some aspects be used as a Kahler angle metric 9KM, and in principle, any monotonic function of the Kahler angle 9K, and of course the Kahler angle 9K itself, can be used as a Kahler angle metric 9KM.

This means that the plurality of channel estimates comprises at least three channel estimates ci, C2; C2, C3; CN-I, CN, where the control unit 1, 1’ is adapted to calculate the Kahler angle 9K between at least two different pairs of corresponding complex channel vectors and to determine an average Kahler angle metric 9KM for the first time period ti. Then the second time periods t2a, t2b, t2x differ from the first time period ti since there are at least two different pairs of corresponding complex channel vectors provided by means of the channel estimates of the first time period.

Each pair of the channel estimates is provided during a second time period, where, as shown in Figure 2A, a plurality of second time periods t2a, t2b, t2x are comprised in the first time period ti. Generally there is least one second time period with two channel estimates, enabling at least one Kahler angle 9K to be calculated for the corresponding pair of complex channel vectors. In the case of a plurality of second time periods t2a, t2b, t2x, these need not be of equal length, and can be regarded as a second type time period.

According to some aspects and as indicated above, if the Kahler angle 9K itself is not the metric, a Kahler angle metric 9KM is formed for each Kahler angle 9K and then an average is formed for the Kahler angle metric 9KM. The average of the Kahler angle metrics 9KM forms a total metric M a that indicates to which degree the estimated channel has changed during the first time period ti.

Alternatively, according to some further aspects, an average of the Kahler angles 9K, an average Kahler angle 9KA, may be calculated for a plurality of Kahler angles 9K, and if the Kahler angle 9K itself is not the metric, a Kahler angle metric 9KM is then formed for the average Kahler angle 9KA, the Kahler angle metric 9KM constituting a total Kahler angle metric M b that indicates to which degree the estimated channel has changed during the first time period ti.

If the Kahler angle itself is the metric, the average is calculated for the plurality of Kahler angles, and a Kahler angle metric 9KM is then formed that equals the average Kahler angle 9KA, and no further metric is involved. The average Kahler angle 9KA then forms the total metric M b that indicates to which degree the estimated channel has changed during the first time period ti.

According to some aspects, the Kahler angle metric is one of:

- the Kahler angle 9K itself,

- sines for the Kahler angle 9K, sin 9K, or

- cosines for the Kahler angle 9K, COS 9K. In the following, it is assumed that a metric is formed for each Kahler angle, and then an average is formed for the metrics.

More in detail, with reference to the block diagram in Figure 3, the control unit 1 is according to some aspects adapted to perform calculations in a number of steps A, B, C where the similarity between channel estimations at different times i denoted by A. The number of steps performed depends on what the determined result is intended to be used for. A first step A is always performed.

In the first step A, by considering a first complex vector a = h _± as a channel estimation from a first reference signal and b = h ₂ as a channel estimation from a second reference signal, the Kahler angle denoted by 9K, is calculated as follows:

First, the Hermitian product is defined for any pair of complex vectors where a _k and b _k denote the components of the vectors and a _k is the complex conjugate of a _k . In the following, magnitude of the vectors is calculated as

Second, a complex- valued angle, 0 _c (a, b), is calculated as cos 9 _c (a, b) =

Finally, the Kahler angle, 9K, <9K<K, is obtained by:

After calculating the similarity, i.e., in the form of the Kahler angle metric 9KM = cos 0 _K (a, b), between first two channel estimations per specific time interval such that a similarity measure Ml is obtained, according to some aspects, a second step B is performed. At the second step B, the control unit 1 is adapted to perform averaging over several time intervals such that an averaged measure M2, corresponding to the total metric Mr _a in the example above, is obtained for an average of a plurality Kahler angle metrics 9KM. According to some aspects, a third step C is performed. At the third step C, the control unit 1 is adapted to check the averaged measure M2 against a predetermined threshold such that a comparison measure M3 is provided, where the comparison measure M3 is forwarded for further use as will be discussed in the following.

According to some aspects, the comparison measure is used for determining if and to what extent, reference signal density should be re-configured, and is according to some aspects made between a Kahler angle metric 9KM, or an average of a plurality Kahler angle metrics 9KM, and one or more thresholds. Depending on which Kahler angle metric 9KM that is used, the Kahler angle metric 9KM, or average of Kahler angle metrics 9KM, can either exceed a threshold or fall below a threshold when the estimated channel has been determined to change to such an extent that an action needs to be taken, for example an increase of reference signal density. When there is only one Kahler angle 9K, there is only one Kahler angle metric 9KM that of course can be the Kahler angle 9K itself.

According to some aspects, as schematically illustrated in Figure 4, in the case of more than one Kahler angle 9K being determined, the control unit 1 is adapted to determine a total metric Mi _a , Mu in a suitable manner, for example in accuracy with any one of the examples above. The control unit is then adapted to:

- increase a reference signal density if the total metric M _a , Mu indicates that the estimated channel has changed to a degree that exceeds a first threshold value Tl, or to

- decrease a reference signal density if the total metric Mu, Mu indicates that the estimated channel has changed to a degree that falls below a second threshold value T2, otherwise

- maintain the present reference signal density.

The reference signal density relates to a density for reference signals 5, 6 exchanged between at least two of the wireless communication nodes 3, 4 and can related to number of reference signals per time unit or complexity of reference signal. The reference signal density can be regarded as how many reference signals that are used in a pre-defined time slot, and can according to some aspects be defined as the number of symbols used for a reference signal in a pre-defined slot in 5G.

An example of a procedure for controlling reference signal density by means of the present disclosure will now be described with reference to Figure 5. Channel similarity measurements can be used to find an appropriate point when to reconfigure the reference signal in fast-changing channel scenarios. After a procedure Start, channel estimations hi, 112 from first and second reference signal are selected 19, whereby the similarity between them is calculated 11. There are now two options depending on if the channel estimations hi, I12 are determined 12 to be similar or not. If they are not determined to be similar, and it is determined 13 that a present RS configuration is not a maximum RS configuration, the RS density is increased 14 in time, e.g., a 2 RS configuration is changed to a 3 RS configuration or a 3 RS configuration is changed to a 4 RSs configuration. However, additional reference signals could be configured if a current RS configuration does not consist of the maximum RS configuration. The procedure then reaches an End. If the present RS configuration is determined to be a maximum RS configuration, the procedure reaches the End.

If they are determined to be similar, and current configuration is determined 15 to be more than minimum RS configuration, it is needed to decrease 16 the RS density in time, e.g., change a 3 RS configuration to a 2 RS configuration or change a 4 RS configuration to a 3 RS configuration. After this step, it should be checked if the new configuration density is good enough to have an accurate channel estimation or not. To this end, the algorithm should be repeated by selecting 10 new channel estimations of first and second reference signal and in the new RS configuration. By calculating 11 their similarity, and if they are not determined 12 to be similar, it means the previous configuration with a higher RS density should be reconfigured 14, otherwise the steps of determining 15 if the current configuration is more than a minimum RS configuration, and if not, the RS density is decreased 16 in time, are repeated. Should it be determined 15 that the current configuration is more than a minimum RS configuration, the procedure reaches the End.

The procedure is according to some aspects performed by means of the control unit 1. The control unit 1 is according to some aspects comprised in one of the wireless communication nodes 3, 4. The control unit 1’ is according to some aspects comprised in a remote server 7 that is accessible for at least one of the wireless communication nodes 2, 3.

By means of the present disclosure:

- A fast-moving UE can be configured with suitable parameters without knowing or estimating its speed.

- Additional overhead can be avoided and consequently the number of available resource elements for user data allocation can be preserved.

- No additional implementation is needed as an estimated channel change can be determined based on available channel estimations.

With reference to Figure 6, the present disclosure also relates to a method in a wireless communication network, where the method comprises providing SI a plurality of channel estimates ci, C2, C3; CN-I, CN which are separated in time during a certain first time period ti and representing S2 each channel estimate ci, C2, C3; CN-I, CN by a corresponding complex channel vector in a complex vector space. Each channel estimate ci, C2, C3; CN-I, CN constitutes an estimated channel for a radio channel 2 provided between at least two wireless communication nodes 3, 4, where the radio channel 2 constitutes a medium via which signals transmitted between nodes 3, 4 are transferred. The method further comprises calculating S3 a difference angle between two of the complex channel vectors which are separated in time during a certain second time period t2a, t2b, t2x, forming a pair of complex channel vectors, where the difference angle is a Kahler angle 9K.

According to some aspects, the plurality of channel estimates ci, C2, C3; CN-I, CN comprises at least three channel estimates, where the method comprises calculating S4 the Kahler angle 9K between at least two different pairs of complex channel vectors corresponding to the at least three channel estimates such that a plurality of Kahler angles 9K is determined and determining S5 a Kahler angle metric 9KM for each Kahler angle 9K. The method further comprises determining S6 an average of the Kahler angle metrics 9KM for the first time period ti, where the average of the Kahler angle metrics 9KM forms a total metric M a that indicates to which degree the estimated channel has changed during the first time period ti.

According to some aspects, the plurality of channel estimates ci, C2, C3; CN-I, CN comprises at least three channel estimates, where the method comprises calculating S4 the Kahler angle 9K between at least two different pairs of complex channel vectors corresponding to the at least three channel estimates such that a plurality of Kahler angles 9K is determined and determining S7 an average Kahler angle 9KA for the first time period ti. The method further comprises determining S8 a Kahler angle metric 9KM for the average Kahler angle 9KA, where the Kahler angle metric 9KM forms a total metric MID that indicates to which degree the estimated channel has changed during the first time period ti.

According to some aspects, the method comprises increasing S9 a reference signal density if DI the total metric M a, Mu indicates that the estimated channel has changed to a degree that exceeds a first threshold value TI, and decreasing S19 a reference signal density if D2 the total metric M _a , Mu indicates that the estimated channel has changed to a degree that falls below a second threshold value T2. Otherwise, the method comprises maintaining SI 1 the present reference signal density, where the reference signal density relates to a density for reference signals 5, 6 exchanged between at least two of the wireless communication nodes 3, 4.

According to some aspects, the Kahler angle metric is one of:

- the Kahler angle 9K itself,

- sines for the Kahler angle 9K, sin 9K, or

- cosines for the Kahler angle 9K, COS 9K.

According to some aspects, the Kahler angle metric is a monotonic function of the Kahler angle 9K. Figure 7 schematically illustrates a control unit 1, 1’ according to aspects of the present disclosure. It is appreciated that the above described methods and techniques may be realized in hardware. This hardware is then arranged to perform the methods, whereby the same advantages and effects are obtained as have been discussed above.

Processing circuitry 101 is provided using any combination of one or more of a suitable central processing unit (CPU), multiprocessor, microcontroller, digital signal processor (DSP), etc., capable of executing software instructions stored in a computer program product, e.g. in the form of a storage medium 102. The processing circuitry 101 may further be provided as at least one application specific integrated circuit (ASIC), or field programmable gate array (FPGA).

Particularly, the processing circuitry 101 is configured to cause the control unit 1, 1’ to perform a set of operations, or steps, for example the methods described above. For example, the storage medium 102 may store the set of operations, and the processing circuitry 101 may be configured to retrieve the set of operations from the storage medium 102 to cause the control unit to perform the set of operations. The set of operations may be provided as a set of executable instructions. Thus, the processing circuitry 101 is thereby arranged to execute methods as herein disclosed.

The storage medium 102 may also comprise persistent storage, which, for example, can be any single one or combination of magnetic memory, optical memory, solid state memory or even remotely mounted memory.

The control unit 1, 1’may further comprise a communications interface 103 for communications with at least one external device. As such the communication interface 103 may comprise one or more transmitters and receivers, comprising analogue and digital components and a suitable number ports for wireline or wireless communication.

The processing circuitry 101 controls the general operation of the control unit 1, 1’, e.g. by sending data and control signals to the communication interface 103 and the storage medium 102, by receiving data and reports from the communication interface 103, and by retrieving data and instructions from the storage medium 102. Other components, as well as the related functionality, of the unit are omitted in order not to obscure the concepts presented herein.

Figure 8 schematically illustrates a computer program product 200 comprising a computer program 201 according to the disclosure above, and a computer readable storage medium 202 on which the computer program is stored. The present disclosure is not limited to the above, but may vary freely within the scope the appended claims. For example, according to some aspects, determining if and to which degree a radio channel changes can be used for many applications other than the described reference signal configuration. For example, the statistics of how and when a channel changes can be used for cell planning. Knowledge of how and when a channel changes can furthermore be used when selecting a specific frequency band by comparing estimated channel properties in different frequency bands.

Generally, the present disclosure relates to a control unit 1, 1’ that is adapted to provide a plurality of channel estimates ci, C2, C3; CN-I, CN which are separated in time during a certain first time period ti and to represent each channel estimate ci, C2, C3; CN-I, CN by a corresponding complex channel vector in a complex vector space, where each channel estimate ci, C2, C3; CN-I, CN constitutes an estimated channel for a radio channel 2 provided between at least two wireless communication nodes 3, 4, where the radio channel 2 constitutes a medium via which signals transmitted between nodes 3, 4 are transferred, wherein the control unit 1, 1’ further is adapted to calculate a difference angle between two of the complex channel vectors which are separated in time during a certain second time period t2a, t2b, t2x, forming a pair of complex channel vectors, where the difference angle is a Kahler angle 9K.

According to some aspects, the plurality of channel estimates ci, C2, C3; CN-I, CN comprises at least three channel estimates, and where the control unit 1, 1’ is adapted:

- to calculate the Kahler angle 9K between at least two different pairs of complex channel vectors corresponding to the at least three channel estimates such that a plurality of Kahler angles 9K is determined,

- to determine a Kahler angle metric 9KM for each Kahler angle 9K, and

- to determine an average of the Kahler angle metrics 9KM for the first time period ti, where the average of the Kahler angle metrics 9KM forms a total metric M a that indicates to which degree the estimated channel has changed during the first time period ti.

According to some aspects, the plurality of channel estimates ci, C2, C3; CN-I, CN comprises at least three channel estimates, and where the control unit 1, 1’ is adapted:

- to calculate the Kahler angle 9K between at least two different pairs of complex channel vectors corresponding to the at least three channel estimates such that a plurality of Kahler angles 9K is determined,

- to determine an average Kahler angle 9KAfor the first time period ti, and

- to determine a Kahler angle metric 9KM for the average Kahler angle 9KA, where the Kahler angle metric 9KM forms a total metric M that indicates to which degree the estimated channel has changed during the first time period ti. According to some aspects, the Kahler angle metric is one of:

- the Kahler angle 9K itself,

- sines for the Kahler angle 9K, sin 9K, or

- cosines for the Kahler angle 9K, COS 9K.

According to some aspects, the Kahler angle metric is a monotonic function of the Kahler angle 9K.

According to some aspects, the control unit 1, 1’ is adapted to increase a reference signal density if the total metric M a, M b indicates that the estimated channel has changed to a degree that exceeds a first threshold value Tl, to decrease a reference signal density if the total metric M a, M indicates that the estimated channel has changed to a degree that falls below a second threshold value T2, otherwise to maintain the present reference signal density, where the reference signal density relates to a density for reference signals 5, 6 exchanged between at least two of the wireless communication nodes 3, 4.

According to some aspects, the control unit 1 is comprised in one of the wireless communication nodes 3.

According to some aspects, the control unit 1’ is comprised in a remote server 7 that is accessible for at least one of the wireless communication nodes 3.

Generally, the present disclosure also relates to a wireless communication node 3 comprising a control unit 1 according to the above.

Generally, the present disclosure also relates to a computer program product 299 comprising computer readable medium 292 carrying a computer program 291, the computer program 291 comprising computer executable instructions, the computer executable instructions 291 being configured such that, on execution by processing circuitry 191, the processing circuitry 191 executes any one of, or a combination of, the methods according to the above.