METHODS AND APPARATUSES FOR FEEDBACK REPORTING IN A

WIRELESS COMMUNICATIONS NETWORK

TECHNICAL FIELD

The present disclosure relates to the field of wireless communications, and in particular, to methods and apparatuses for efficient feedback reporting for at least a New Radio- (NR-) based wireless communication network system, which feedback includes Channel State Information (CSI).

BACKGROUND

In a wireless communications system, such as New Radio, also called 3GPP Fifth Generation wireless communications system or 5G for short, downlink (DL) and uplink (UL) signals convey data signals, control signals comprising DL control information (DCI) and/or uplink control information (UCI), and a number of reference signals (RSs) used for different purposes. A radio network node or a radio base station or a gNodeB (or gNB or gNB/TRP (Transmit Reception Point)) transmits data and DCI through the so-called physical downlink shared channel (PDSCH) and the physical downlink control channel (PDCCH), respectively.

A UE transmits data and UCI through the so-called physical uplink shared channel (PUSCH) and physical uplink control channel (PUCCH), respectively. Moreover, the DL or UL signal(s) of the gNB respectively the user equipment (UE or a radio device) may contain one or multiple types of RSs including a channel state information RS (CSI-RS), a demodulation RS (DM-RS), and a sounding RS (SRS). The CSI-RS (SRS) is transmitted over a DL (UL) system bandwidth part and used at the UE (gNB) for CSI acquisition. The DM-RS is transmitted only in a bandwidth part of the respective PDSCH/PUSCH and used by the UE/gNB for data demodulation.

One of many key features of 5G is the use of multi-input multi-output (Ml MO) transmission schemes to achieve high system throughput compared to previous

generations of mobile systems. MIMO transmission generally demands the availability of accurate CSI used at the gNB for a signal precoding using a precoding matrix of the data and control information. The current third Generation Partnership Project Release 15 specification (3GPP Rel. 15) therefore provides a comprehensive framework for CSI reporting. The CSI is acquired in a first step at the UE based on received CSI-RS signals transmitted by the gNB. The UE determines in a second step based on the estimated channel matrix a precoding matrix from a predefined set of matrices called ‘codebook’. The selected precoding matrix is reported in a third step in the form of a precoding matrix identifier (PM I) and rank identifier (Rl) to the gNB.

In the current Rel. -15 NR specification, there exist two types (Type-I and Type-II) for CSI reporting, where both types rely on a dual-stage (i.e., two components) W1W2 codebook. The first codebook, or the so-called first stage precoder, W1, is used to select a number of beam vectors from a Discrete Fourier Transform-based (DFT-based) matrix which is also called the spatial codebook. The second codebook, or the so-called second stage precoder, W2, is used to combine the selected beams. For Type-I and Type-II CSI reporting, W2 contains phase-only combining coefficients and complex combing coefficients, respectively. Moreover for Type-II CSI reporting, W2 is calculated on a subband basis such that the number of columns of W2 depends on the number of configured subbands. Here, a subband refers to a group of adjacent physical resource blocks (PRBs). Although Type-II provides a significant higher resolution than Type-I CSI feedback, one major drawback is the increased feedback overhead for reporting the combining coefficients on a subband basis. The feedback overhead increases approximately linearly with the number of subbands, and becomes considerably large for large numbers of subbands. To overcome the high feedback overhead of the Rel. -15 Type-II CSI reporting scheme, it has recently been decided in 3GPP RAN#81 [2] (3GPP radio access network (RAN) 3GPP RAN#81) to study feedback compression schemes for the second stage precoder W2.

As will be described in according with some embodiments herein, a problem of how to compress and efficiently quantize the combining coefficients of W2 is addressed.

But before going into the detailed description of the solution(s) of the present embodiments, an informative description is provided in order to better understand the problems of the prior art followed by a described how said problems are solved according to the embodiments of the present disclosure.

3GPP Rel.-15 dual-stage precoding and CSI reporting

Assuming a rank-L (L may be up to two) transmission and a dual-polarized antenna array at the gNB with configuration (N1, N2, 2), the Rel.-15 double-stage precoder for the s-th subband for a layer is given by

where the precoder matrix W has 2 N1N2 rows corresponding to the number of antenna ports, and S columns for the reporting subbands/PRBs. The matrix
is thewideband first-stage precoder containing 2 U spatial beams for

both polarizations which are identical for all S subbands, and WA is a diagonal matrix containing 2 U wideband amplitudes associated with the 2 U spatial beams, and is the second-stage precoder containing 2 U subband (subband

amplitude and phase) complex frequency-domain combining-coefficients associated with the 2 U spatial beams for the s-th subband.

According to [1], the reporting and quantization of the wideband amplitude matrix W A and subband combining coefficients in are quantized and reported as

follows:

- The wideband amplitude corresponding to the strongest beam which has an amplitude value of 1 is not reported. The wideband amplitude values associated with the remaining 2U — 1 beams are reported by quantizing each amplitude value with 3 bits.

- The subband amplitudes and phase values of the coefficients associated with the first leading beam are not reported (they are assumed to be equal to 1 and 0).

- For each subband, the amplitudes of the B coefficients associated with the first B — 1 leading beams (other than the first leading beam) are quantized with 1 bit (quantization levels [sqrt(0.5), 1]). The amplitude values of the remaining 2U — B beams are not reported (they are assumed to be equal to 1).

- For each subband, the phase values of the B — 1 coefficients associated with the first B — 1 leading beams (other than the first leading beam) are quantized with 3 bits. The phase values of the remaining 2U — B beams are quantized with 2 bits.

- The number of leading beams for which the subband amplitude is reported is given by B = 4, 4 or 6 when the total number of configured spatial beams U = 2, 3, or 4, respectively.

- SUMMARY AND SOME DETAILED DESCRIPTION

In view of the drawbacks disclosed earlier, there is provided a communication device or a radio device or a user equipment (UE) and a method therein for providing a channel state information (CSI) feedback in a wireless communication system including at least the UE and a gNB or a radio network node. The UE comprising a processor and a memory, said memory containing instructions executable by said processor whereby said UE is operative by means of e.g. a transceiver to receive from a transmitter (e.g. the gNB or any suitable network node and/or radio communication device) a radio signal via a MIMO channel, where the radio signal contains DL reference signals according to a DL reference signal configuration. The UE is further operative, by means of e.g. the processor to:

- estimate the MIMO channel between the gNB and the UE based on the received DL reference signals for the configured resource blocks,

- calculate, based on a performance metric, a precoder matrix, for a number of antenna ports of the gNB and configured subbands, the precoder matrix being based on two codebooks and a set of combination coefficients for

complex scaling/combining one or more of vectors selected from a first codebook and a second codebook, wherein:

o the first codebook contains one or more transmit-side spatial beam components of the precoder, and

o the second codebook contains one or more delay components of the precoder, and

the UE is operative to report a CSI feedback and/or a PMI and/or a PMI/RI, used to indicate the precoder matrix for the configured antenna ports and resource blocks.

In accordance with some exemplary embodiments, the first codebook comprises a first DFT- or oversampled DFT-codebook-matrix of size N1N2 x 01,1N101,2N2 containing the spatial beam components (N1N2 x 1 vectors) of the precoder matrix. Here, N1 and N2 refer to the number of antenna ports of the same polarization in the first and second dimension of the antenna array, respectively.

In general, for a two-dimensional (2D) antenna array, N1 and N2 are both greater than 1 , whereas for a linear (or one-dimensional (1 D)) either N1 or N2 is one. The total number of antenna ports for dual-polarized antenna array that may be considered for better understanding is 2 N1N2. Furthermore, 01,1 ∈ {1,2,3, ..} and 01,2 ∈ {1,2,3, ..} refer to the oversampling factors of the codebook matrix with respect to the first and second dimension, respectively. The second codebook comprises a second DFT, or discrete cosine transform (DCT-), or oversampled DFT-, or oversampled DCT-codebook matrix of size N3 x N3O2 containing the delay components (represented by N3 x 1 DFT-/DCT-vectors) of the precoder matrix, where O2 refers to the oversampling factor O2 = 1,2, .... of the second codebook matrix. Each DFT/DCT vector of the second codebook is associated with a delay (in the transformed domain), as each DFT/DCT vector may model a linear phase increase over the N3 subbands. Therefore, herein we may refer to DFT/DCT vectors of the second codebook in the following as delay vectors or simply delays.

In accordance with some exemplary embodiments, the precoder matrix F(l) of the l-th transmission layer is represented by a three-stage structure
where

- contains U(l) selected beam components/beam vectors from the first codebook of the l- th layer for the 2 N1N2 antenna ports,

- contains selected delay vectors from the second codebook of the u- th

beam for the configured N3 subbands, where the number of delay vectors per
beam may be identical or different over the beams, and

- contains a number of complex-combining coefficients used to combine the

selected U(l) beam vectors and delay vectors per layer.

According to an embodiment, the precoder matrix of the l-
th transmission for the configured 2 N1N2 antenna ports and N3 subbands may also be represented in a double sum notation for the first polarization of the antenna ports as

and for the second polarization of the antenna ports as

where represents the u-th spatial beam vector (contained

in matrix selected from the first codebook, is the

delay vector (contained in matrix associated with the u-th beam and p-th

polarization selected from the second codebook, is the complex combining

coefficient (contained in matrix associated with the u- th beam, d-th delay and

p-th polarization, and α (l) is a normalizing scalar.

For brevity, in the following embodiments the delay vectors and are

exemplified as identical across two polarizations, such that

However, the embodiments herein are not restricted to this example, which means that the embodiments may also be applicable when delay vectors are not identical over both polarizations.

Configuration of the second codebook (N3, O2)

In accordance with exemplary embodiments, the UE may be configured to receive from the gNB the higher layer (such as Radio Resource Control (RRC) layer or medium access control-control element (MAC-CE)) or physical layer (Layer 1 or L1) parameter oversampling denoted N3 for the configuration of the second codebook. The specific value of the number of subbands N3 may depend on the maximum expected delay spread of the radio channel and the computational complexity spent at the UE for calculating the combining coefficients of the precoder matrix. Therefore, the specific value of N3 may depend on parameters related to or associated with the radio channel (such as the channel delay spread) and different design aspects of the precoder. In one example, the value of N3 may be identical to the number of configured channel Quality Indicator (CQI) subbands (low computational complexity approach). In another example, the value of N3 may be identical to the number of configured PRBs (high computational complexity approach), although not necessary for the functioning of the embodiments herein.

In accordance with some exemplary embodiments, the value of N3 may be defined by/as the total number of subbands with subband size NPRB, wherein PRB stands for physical resource block, where NPRB denotes the number of PRBs per subband. The value of NPRB may depend on the parameters of a orthogonal frequency division multiplexing (OFDM) transmission signal such as a configured subcarrier spacing (SCS) and a channel delay spread of the channel. Two exemplary values for NPRB are 4 and 2 for 15 KHz and 30 KHz SCSs, respectively.

In accordance with some exemplary embodiments, the UE may be configured or operative to receive from the gNB a higher layer (RRC or MAC-CE) or physical layer (L1) parameter oversampling factor O2 for the configuration of the second codebook. The oversampling factor defines the grid size of the delay components of the precoder. A large oversampling factor may result in a very fine grid for the delay components of the precoder and enhanced performance, but it also increases the codebook size and the computational complexity for selecting the delay components of the precoder.

In accordance with some exemplary embodiments, the UE is configured or is operative to select the oversampling factor used for the configuration of the second codebook and signal to the gNB by higher layer (RRC or MAC-CE) or physical layer (L1) the oversampling factor O2.

In accordance with some exemplary embodiments, the UE is configured or is operative to use an a priori known (default) oversampling factor(s) O2 for the configuration of the second codebook. In such a case, the oversampling factor may depend on the total number of configured PRBs (e.g. the total system bandwidth), where a higher oversampling factor (e.g., O2 = 8 or O2 = 16) may be applied when the total number of PRBs is larger than a specific pre-determined value and a lower oversampling factor (e.g., O2 = 4, O2 = 2 or Ot = 1) otherwise.

In accordance with some exemplary embodiments, the UE may be configured or may be operative to signal its capability with respect to the oversampling factor of the second codebook. For example, a UE with a limited computational power may not support oversampling of the second codebook and may signal O2 = 1. Hence, signaling UE capabilities may be advantageous in case the UE has limited computational power or capacity or CPU power.

Beam configuration and Reporting of selected beam indices

In accordance with some exemplary embodiments, the UE is configured to or is operative to receive from the gNB a higher layer (RRC or MAC-CE) or physical layer (L1) parameter U(l), representing the number of spatial beams for the l- th transmission layer. The number of spatial beams U(l) and the selected spatial beam vectors from the first codebook are typically different for each transmission layer. However, the reporting of different spatial beam vectors for each transmission layer may result in a high feedback overhead. In order to reduce the feedback overhead in accordance with embodiments herein, the UE may be

configured to or may be operative to select identical beam vectors from the first codebook for a subset of the transmission layers which is advantageous. For example, the UE may be configured to or be operative to select identical spatial beam vectors for the first and second transmission layers and different (but possibly identical) spatial beam vectors for the third and fourth transmission layers.

Delay configuration and Reporting of selected delay vectors

The configured U(l) beam vectors and the delay vectors per beam of the

precoder matrix are aligned with the multipath components of the MIMO propagation channel. The multipath components of the radio channel generally occur in the form of multipath clusters, where a multipath cluster may be understood as a group of multipath components with similar channel propagation parameters such as angle-of-arrival, angle-of-departure and delay [3], Depending on the cluster distribution in the spatial and delay domains of the radio channel, each beam vector of the precoder matrix may be associated with a single cluster or few clusters, where each cluster may have a different delay. Some of the beam vectors of the precoder matrix shall therefore be associated with a small number of delays/delay vectors and some of the beam vectors shall be associated with a large number of delays/delay vectors.

In accordance with some exemplary embodiments, the UE may be configured with a different number of delays per beam vector, or with subsets of beam vectors

having an identical number of delays and with a different number of delays per subset. The number of configured delays may increase (decrease) with a beam or subgroup beam index. The selected delay vectors by the UE may be non-identical, partially identical, or fully identical over the beam indices and/or layer indices. Hence, the embodiments herein are not restricted to any specific delay vectors.

There is also provided a method performed by the UE as previously described. The method includes:

- estimating the MIMO channel (as previously described) between the gNB and the UE based on the received DL reference signals for the configured resource blocks,

- calculating, based on a performance metric, a precoder matrix, for a number of antenna ports of the gNB and configured subbands, the precoder matrix being based on two codebooks and a set of combination coefficients for complex scaling/combining one or more of vectors selected from a first codebook and a second codebook, wherein:

o the first codebook contains one or more transmit-side spatial beam components of the precoder, and

o the second codebook contains one or more delay components of the precoder, and

the UE reporting, to the gNB, a CSI feedback and/or a PMI and/ora PMI/RI, used to indicate the precoder matrix for the configured antenna ports and resource blocks.

According to an exemplary embodiment, the method further comprises receiving from the gNB the higher layer (such as Radio Resource Control (RRC) layer or medium access control-control element (MAC-CE)) or physical layer (Layer 1 or L1) parameter oversampling denoted N3 for the configuration of the second codebook.

According to another exemplary embodiment, the method further comprises receiving from the gNB a higher layer (RRC or MAC-CE) or physical layer (L1) parameter oversampling factor O2 for the configuration of the second codebook.

In accordance with some exemplary embodiments, the method may further comprises receiving from the gNB a higher layer (RRC or MAC-CE) or physical layer (L1) parameter U(l), representing the number of spatial beams for the l- th transmission layer. The number of spatial beams U(l) and the selected spatial beam vectors from the first codebook are typically different for each transmission layer. However, the reporting of different spatial beam vectors for each transmission layer may result in a high feedback overhead. In order to reduce the feedback overhead in accordance with embodiments herein, the method comprises selecting identical beam vectors from the first codebook for a subset of the transmission layers which is advantageous. For example, for the UE, the

method may be configured to select identical spatial beam vectors for the first and second transmission layers and different (but possibly identical) spatial beam vectors for the third and fourth transmission layers.

As described, the configured U(l) beam vectors and the delay vectors per

beam of the precoder matrix are aligned with the multipath components of the MIMO propagation channel. The multipath components of the radio channel generally occur in the form of multipath clusters, where a multipath cluster may be understood as a group of multipath components with similar channel propagation parameters such as angle-of-arrival, angle-of-departure and delay [3], Depending on the cluster distribution in the spatial and delay domains of the radio channel, each beam vector of the precoder matrix may be associated with a single cluster or few clusters, where each cluster may have a different delay. Some of the beam vectors of the precoder matrix shall therefore be associated with a small number of delays/delay vectors and some of the beam vectors shall be associated with a large number of delays/delay vectors.

In accordance with some exemplary embodiments, the method performed by the UE may include that the UE be configured with a different number of delays
per beam vector, or with subsets of beam vectors having an identical number of delays and with a different number of delays per subset. The number of configured delays may increase (decrease) with a beam or subgroup beam index. The selected delay vectors by the UE may be non-identical, partially identical, or fully identical over the beam indices and/or layer indices. Hence, the embodiments herein are not restricted to any specific delay vectors.

There is also provided a computer program comprising instructions which when executed on at least one processor of the UE according to the method related or associated with the UE described above, cause the at least said one processor to carry out the method according to anyone of the method subject-matter disclosed earlier. A carrier is also provided containing the computer program wherein the carrier is one of a computer readable storage medium; an electronic signal, optical signal or a radio signal.

There is also provided a method performed by the gNB or a radio network node or a radio base station and a radio network node or a gNB. The gNB is configured to perform at least the steps disclosed earlier. The method performed by the gNB includes in method terms, what has been defined as “configured to. As an example, the method in the gNB may include receiving from the UE a CSI feedback and/or a PMI and/or a PMI/RI, used to indicate the precoder matrix for the configured antenna ports and resource blocks.

According to an exemplary embodiment, the method, by the gNb may include transmitting to the UE a higher layer (such as Radio Resource Control (RRC) layer or medium access control-control element (MAC-CE)) or physical layer (Layer 1 or L1) parameter oversampling denoted N3 for the configuration of the second codebook.

According to another exemplary embodiment, the method further comprises transmitting to the UE a higher layer (RRC or MAC-CE) or physical layer (L1) parameter oversampling factor O2 for the configuration of the second codebook.

In accordance with some exemplary embodiments, the method may further comprise transmitting to the UE a higher layer (RRC or MAC-CE) or physical layer (L1) parameter U(l) representing the number of spatial beams for the l- th transmission layer. The number of spatial beams U(l) and the selected spatial beam vectors from the first codebook are typically different for each transmission layer. However, the reporting of different spatial beam vectors for each transmission layer may result in a high feedback overhead. In order to reduce the feedback overhead in accordance with embodiments herein, the method comprises selecting identical beam vectors from the first codebook for a subset of the transmission layers which is advantageous. For example, for the UE, the method may be configured to select identical spatial beam vectors for the first and second transmission layers and different (but possibly identical) spatial beam vectors for the third and fourth transmission layers.

In accordance with some exemplary embodiments, the method performed by the gNB may include configuring the UE with a different number of delays per

beam vector, or with subsets of beam vectors having an identical number of delays and with a different number of delays per subset. The number of configured delays may increase (decrease) with a beam or subgroup beam index. The selected delay vectors by the UE may be non-identical, partially identical, or fully identical over the beam indices and/or layer indices. Hence, the embodiments herein are not restricted to any specific delay vectors.

According to another aspect of embodiments herein, there is also provided a radio base station or gNB, the radio base station comprising a processor and a memory, said memory containing instructions executable by said processor whereby said gNB is operative to perform any one of the subject-matter of method steps described above.

There is also provided a computer program comprising instructions which when executed on at least one processor of the gNB according to the method related or associated with the gNB described above, cause the at least said one processor to carry out the method according to anyone of the method subject-matter disclosed earlier. A carrier is also provided containing the computer program wherein the carrier is one of a computer readable storage medium; an electronic signal, optical signal or a radio signal.

BRIEF DESCRIPTION OF THE DRAWINGS

Examples of embodiments and advantages of the embodiments herein are described in more detail with reference to attached drawings in which:

Figures 1-4 depict several examples of delay configurations for the precoder matrix of a layer with different computational complexities and feedback overheads for selecting and reporting the delay vectors per beam are provided.

Figures 5-12 depicts examples of the number of feedback bits for amplitude reporting according to some exemplary embodiments herein

Figure 13 is an exemplary block diagram depicting a radio base station or gNB or network node according to exemplary embodiments herein.

Figure 14 is a block diagram depicting a UE or communication device or radio device according to exemplary embodiments herein.

DETAILED DESCRIPTION

In order to perform the previously described process or method steps related to the radio network node (e.g. a radio base station or gNB), some embodiments herein include a network node for receiving feedback from a UE as previously described. As shown in Figure 13, the network node or radio base station or gNB 800 comprises a processor 810 or processing circuit or a processing module or a processor or means 810; a receiver circuit or receiver module 840; a transmitter circuit or transmitter module 850; a memory module 820 a transceiver circuit or transceiver module 830 which may include the transmitter circuit 850 and the receiver circuit 840. The network node 800 further comprises an antenna system 860 which includes antenna circuitry for transmitting and receiving signals to/from at least the UE. The antenna system employs beamforming as previously described.

The network node 500 may belong to any radio access technology including 2G, 3G, 4G or LTE, LTE-A, 5G, WLAN, and WiMax etc. that support beamforming technology.

The processing module/circuit 810 includes a processor, microprocessor, an application specific integrated circuit (ASIC), field programmable gate array (FPGA), or the like, and may be referred to as the “processor 810.” The processor 810 controls the operation of the network node 800 and its components. Memory

(circuit or module) 820 includes a random access memory (RAM), a read only memory (ROM), and/or another type of memory to store data and instructions that may be used by processor 810. In general, it will be understood that the network node 800 in one or more embodiments includes fixed or programmed circuitry that is configured to carry out the operations in any of the embodiments disclosed herein.

In at least one such example, the network node 800 includes a microprocessor, microcontroller, DSP, ASIC, FPGA, or other processing circuitry that is configured to execute computer program instructions from a computer program stored in a non-transitory computer-readable medium that is in, or is accessible to the processing circuitry. Here, “non-transitory” does not necessarily mean permanent or unchanging storage, and may include storage in working or volatile memory, but the term does connote storage of at least some persistence. The execution of the program instructions specially adapts or configures the processing circuitry to carry out the operations disclosed herein including anyone of method steps already described. Further, it will be appreciated that the network node 800 may comprise additional components not shown in Figure 13.

Details on the functions and operations performed by the network node have already been described and need not be repeated again.

In order to perform the previously described process or method steps related to the UE or communication device or radio device, some embodiments herein include a UE for providing efficient feedback reporting for at least a New Radio-(NR) based wireless communication network system, which feedback includes Channel State Information (CSI).

CLAIMS

1. A method performed by a User Equipment, UE, the method comprising:

- receiving from a network node, a radio signal via a Multiple Input Multiple Output, MIMO, channel, wherein the radio signal contains at least one

DownLink, DL, reference signal according to a DL reference signal configuration;

- estimating said MIMO channel based on said received at least one DL reference signal for configured resource blocks;

- calculating a precoding matrix for a number of antenna ports of the network node and configured subbands; the precoding matrix being based on a first codebook and on a second codebook and a set of combination coefficients for complex scaling/combining one or more of vectors selected from the first codebook and the second codebook, wherein the first codebook contains one or more transmit-side spatial beam components/vectors of the precoding matrix and the second codebook contains one or more delay components/vectors of the precoding matrix; and

- reporting, to the network node, a Channel State Information, CSI, feedback and/or a Precoder matrix Indicator, PMI and/or a PM I/Rank Indicator, PMI/RI, used to indicate the precoding matrix for the configured antenna ports and subbands, wherein the report contains a bitmap for indicating at least selected delay vectors and spatial beam vectors associated with non-zero combining coefficients of said set of combination coefficients.

2. The method according to claim 1 wherein the precoding matrix,

of a l-th transmission layer is represented by a

double sum notation for a first polarization of the antenna ports,

and for a second polarization of the antenna ports,

wherein representing U(l) selected beam

components or Discrete Fourier Transform, DFT-based beam vectors selected from the first codebook for N1N2 antenna ports, where N1 and N2 refer to the number of antenna ports of a same polarization in a first and second dimension of an antenna array of the network node, respectively, representing D(l) selected delay components

or Discrete Fourier Transform, DFT-based delay vectors for the u-th beam selected from the second codebook, wherein the number of DFT-based delay vectors D(l) is identical for all the beams, are the complex

combining coefficients associated with the U(l) selected beam vectors and D(l) selected delay vectors, and a(l) is a normalizing scalar.

3. The method according to claim 1 or claim 2 further comprising receiving from said network node a higher layer parameter K indicating a maximum number of non-zero combining coefficients to be reported by the UE per layer.

4. The method to claim 3 wherein the bitmap contains K or less than K number of "1"s per layer.

5. The method according to anyone of claims 1-4 wherein a bit value 1 in the bitmap indicates that the non-zero combining coefficient with associated vectors selected from the first and second codebooks is reported, and a bit value 0 indicates that the corresponding combining coefficient is not reported.

6. The method according to anyone of claims 1-5 further comprising receiving from said network node a parameter, N3, for a configuration of the number of subbands for the second codebook, and wherein the value of said parameter depends on a maximum expected delay spread of a radio channel and a computational complexity spent at the UE for calculating the combination coefficients of the precoding matrix.

7. The method according to anyone of claims 1-6 comprising reporting for each beam component or each beam group a delay indicator per transmission layer for the delay vectors selected from the second codebook.

8. The method according to claim 1 or claim 2 further comprising selecting identical beam vectors from the first codebook for a subset of transmission layers.

9. The method according to claim 1 or claim 2 comprising selecting identical spatial beam vectors for a first and a second transmission layer and different or identical spatial beam vectors for a third and a fourth transmission layer.

10. The method of claim 1 or claim 2 comprising receiving from the network node, via higher layer signaling, a configuration, including the parameter U(l), indicating the number of spatial beam vectors, and the parameter D(l) indicating the number of delay vectors.

11. The method according to anyone of claims 1-10 comprising selecting for each beam delay vector(s) from a common set of D(l) non-identical delay vectors selected by the UE from the second codebook, and reporting to said network node a single delay indicator indicating selected non-identical delay vectors of the common set.

12. The method according to claim 11 comprising reporting, to the network node, for each beam a delay indicator for the delay vectors selected by the UE, wherein the selected delay vectors are associated with non-zero combining coefficients of the beam, and wherein the delay indicator indicates the selected delay vectors from the common set of non-identical delay vectors.

13. The method according to claim 11 comprising selecting D(l) delay vectors from the second codebook for the common set, wherein N out of D(l) delays are fixed and known at the UE and network node, and wherein D(l) - N selected delays are indicated by the delay indicator.

14. The method according to claim 13 wherein N=1.

15. The according to anyone of claims 1-14 comprising normalizing the combining coefficients with respect to the strongest combining coefficient in amplitude and phase such that the strongest combining coefficient is given by the value one and not reported.

16. The method according to anyone of claims 1-15 wherein a beam associated with a combining coefficient having a largest amplitude over all combining coefficients is the leading beam.

17. The method according to claim 16 comprising subtracting a reference delay which is identical to the first delay associated with the leading beam from the selected delays of the common set.

18. The method according to claim 11 and claim 17 comprising not reporting an index associated with a first delay vector of the leading beam by removing the index associated with the first delay vector from the delay indicator.

19. The method according to claim 1 or claim 2 comprising selecting delay vectors per beam and layer from a subset of delay vectors from the second codebook.

20. The method according to claim 19 wherein the subset of delay vectors is defined by the first Z vectors, or a number, Z1 of first vectors and a number, Z2, of last vectors of a DFT-based matrix.

21. The method according to claim 19 wherein the subset of delay vectors is defined by i1: i2 orthogonal vectors of a DFT-based matrix.

22. The method according to claim 19 or claim 20 wherein the UE is configured with the subset of delay vectors from the second codebook by higher layer signaling or with an a prior known subset of delay vectors from the second codebook.

23. The method according to claim 20 wherein the parameters Z, or Z1 and Z2 that indicate the subset of delay vectors from the second codebook are a priori known.

24. The method according to anyone of claim 1 or claim 2 comprising quantizing and reporting the combining coefficients per beam of the precoding matrix, wherein each combining coefficient is a product of

three coefficients al,p,i, bl,p,i,j and dl,p,i,j, and is given by:

where al,p,i, is a real-valued coefficient representing a common amplitude across all combining coefficients associated with a i- th beam, p-th polarization and l-th layer, bl,p,i,j is a real-valued normalized combining- coefficient representing the amplitude associated with the i-th beam, j-th delay vector, p-th polarization and l-th layer, and
is a coefficient to indicate the phase of

25. The method according to claim 1 or claim 2 comprising quantizing and reporting the combining coefficients per beam of the precoding matrix, wherein each combining coefficient is a product of three coefficients

cl,p,j, bl,p,i,j and dl,p,i,j.

where cl,p,j is a polarization-dependent real-valued coefficient representing a common amplitude across all combining coefficients associated with the j-th delay vector and l- th layer, bl,p,i,j is a real-valued normalized combining-coefficient representing the amplitude associated with the i-th beam, y-th delay vector, p-th polarization and l-th layer, and
is a coefficient to indicate

the phase of

26. The method according to claim 24 wherein the quantization of the amplitudes al,p,i, and/or bl,p,i,j is identical for all combining coefficients of a layer.

27. The method according to claim 25 wherein the quantization of the amplitudes cl,p,j and/or bl,p,i,j is identical for all combining coefficients of a layer.

28. The method according to claim 24 wherein the amplitudes al,p,i are partitioned, per layer, into at least two disjoint subsets, and each subset is assigned a single and different value for said quantization.

29. The method according to claim 25 wherein the amplitudes cl,p,j are partitioned, per layer, into at least two disjoint subsets, and each subset is assigned a single and different value for said quantization.

30. The method according to claim 28 wherein each subset contains the amplitudes al,p,i with respect to a single polarization.

31. The method according to claim 29 wherein each subset contains the amplitudes cl,p,j with respect to a single polarization.

32. The method according to claim 28 wherein the amplitudes al,p,i of the first set contains the strongest amplitude and is quantized with 0 bits and not reported, and the amplitudes al,p,i of the second set are quantized with N=1 or 2 or 3 or 4 bits and reported.

33. The method according to claim 29 wherein the amplitudes cl,p,j of the first set contains the strongest amplitude and is quantized with 0 bits and not reported, and the amplitudes cl,p,j of the second set are quantized with N=1 or 2 or 3 or 4 bits and reported.

34. The method according to claim 24 or claim 25 comprising partitioning the amplitudes bl,p,i,j, per layer, into at least two disjoint subsets per layer and each subset is assigned a single value for quantization of the amplitudes bl,p,i,j .

35. The method according to claim 34 wherein the first set of said distinct subsets contains the amplitudes bl,p,i,j, corresponding to a number less or equal of K selected non-zero combining coefficients, indicated by the bitmap, and the second set contains the remaining amplitude coefficients.

36. The method according to claim 35 wherein the amplitudes of the first set are quantized with N=2, or 3 bits and reported, and the amplitudes of the second set are quantized with 0 bits and not reported.

37. The method according to claim 24 or claim 25 wherein quantization of phases dl,p,i,j is identical for all combining coefficients using a single value for a l-th layer, wherein the single value is known to the UE and is identical for all layers.

38. The method according to claim 24 and claim 25 comprising partitioning the phases dl,p,i,j into at least two disjoint subsets, per layer, and each subset is assigned a single value for phase quantization.

39. The method according to claim 38 wherein the first set contains the phases corresponding to a number less or equal of K selected non-zero combining coefficients, indicated by the bitmap, and the second set contains the remaining phases, and wherein the phases of the first set are quantized with N=2 or 3 or 4 bits and reported, and the phases of the second set are quantized with 0 bits and not reported.

40. The method according to claim 35 and claim 39 wherein the bitmap is used to indicate reported phases from the first set and the second set and wherein the same bitmap is used for indicating the amplitudes bl,p,i,j of the first set and the second set.

41. The method according to claim 24 comprising normalizing the amplitudes al,p,i and reporting the amplitudes al,p,i except for the strongest amplitude.

42. The method according to claim 25 comprising normalizing the amplitudes cl,p,j and reporting the amplitudes cl,p,j except for the strongest amplitude.

43. A method performed by a network node the method comprising:

- transmitting to a User Equipment, UE, a radio signal via a Multiple Input Multiple Output, MIMO, channel, wherein the radio signal contains at least one DownLink, DL, reference signal according to a DL reference signal configuration; and

- receiving from the UE a report including a Channel State Information, CSI, feedback and/or a Precoder Matrix Indicator, PM I and/or a PM I/Rank

Indicator, PMI/RI, used to indicate a precoding matrix for configured antenna ports and configured subbands, the precoding matrix being based on a first codebook and on a second codebook and a set of combination coefficients for complex scaling/combining one or more of vectors selected from the first codebook and the second codebook, wherein the first codebook contains one or more transmit-side spatial beam components/vectors of the precoding matrix and the second codebook contains one or more delay components/vectors of the precoding matrix; and

wherein the report contains a bitmap for indicating at least selected delay vectors and spatial beam vectors associated with non-zero combining coefficients of said set of combination coefficients.

44. The method according to claim 43 wherein the precoding matrix,

of a l-th transmission layer is represented by a

double sum notation for a first polarization of the antenna ports,

and for a second polarization of the antenna ports,

wherein (u = 0, ..., U(l) — 1) representing U(l) selected beam

components or Discrete Fourier Transform, DFT-based beam vectors selected from the first codebook for N1N2 antenna ports, where N1 and N2 refer to the number of antenna ports of a same polarization in a first and second dimension of an antenna array of the network node, respectively,

( d = 0, ... ,D (l) — 1) representing D(l) selected delay components

or Discrete Fourier Transform, DFT-based delay vectors for the u-th beam selected from the second codebook, wherein the number of DFT-based delay vectors D(l) is identical for all the beams, are the complex

combining coefficients associated with the U(l) selected beam vectors and D(l) selected delay vectors, and α(l) is a normalizing scalar

45. The method according to claim 43 or claim 44 further comprising transmitting to the UE, a higher layer parameter K indicating a maximum number of non-zero combining coefficients to be reported by the UE per layer.

46. The method according to anyone of claims 43-45 comprising transmitting to the UE a parameter, N3, for a configuration of the number of subbands for the second codebook, and wherein the value of said parameter depends on a maximum expected delay spread of a radio channel and a computational complexity spent at the UE for calculating the combination coefficients of the precoding matrix.

47. The method according to claim 43 or claim 44 comprising configuring the UE with the parameter U(l), indicating the number of spatial beam vectors, and with the parameter D(l) indicating the number of delay vectors via higher layer signaling.

48. A User Equipment, UE, (900) comprising a processor (910) and a memory (920), said memory (920) containing instructions executable by said processor (920) whereby said UE (900) is operative to perform any one of the subject-matter of method claims 1-42.

49. A network node (800) comprising a processor (810) and a memory (820), said memory (820) containing instructions executable by said processor (810) whereby said network node (800) is operative to perform any one of the subject-matter of method claims 43-47.