HK1247450B - Codebook subset restriction signaling - Google Patents

Description

Codebook subset restriction signaling

Related applications

This application claims priority to U.S. Provisional Patent Application Serial No. 62/103,101, filed January 14, 2015, the entire contents of which are incorporated herein by reference.

Technical Field

The present application relates generally to a network node and a wireless communication device for operation in a wireless communication system, and more particularly to a network node signaling to a wireless communication device which precoders in a codebook are restricted for use.

Background Art

The use of multiple antennas at the transmitter and/or receiver of a wireless communication system can significantly improve the capacity and coverage of the wireless communication system. Such MIMO systems can exploit the spatial dimensionality of the communication channel. For example, several information-carrying signals can be sent in parallel using the transmit antennas and remain separate through signal processing at the receiver. By adapting the transmission to the current channel conditions, significant additional gain can be achieved. One form of adaptation is to dynamically adjust the number of signals carrying information streams transmitted simultaneously from one TTI to another to the number supported by the channel. This is often referred to as (transmit) rank adaptation. Precoding is another form of adaptation, in which the phase and amplitude of the aforementioned signals are adjusted to better match the current channel properties. The signals are formed into vector-valued signals, and the adjustments can be viewed as multiplications using a precoder matrix. A common approach is to select the precoder matrix from a finite, indexed set (the so-called codebook). This type of codebook-based precoding is an integral part of the LTE standard and is also used in many other wireless communication standards.

Codebook-based precoding can be considered a form of channel quantization. A typical approach (see LTE and MIMO HSDPA) is for the receiver to recommend a suitable precoder matrix to the transmitter by signaling a precoder matrix indicator (PMI) over the feedback link. If the feedback link has limited capacity, it is generally important to keep the codebook size as small as possible to limit signaling overhead. However, this needs to be balanced against performance impact, as a larger codebook may better match the current channel conditions.

For example, in the LTE downlink, the user equipment (UE) reports the precoding matrix indicator (PMI) to the eNodeB periodically on the physical uplink control channel (PUCCH) or aperiodically on the physical uplink shared channel (PUSCH). The former is a relatively narrow bit pipe (e.g., using a few bits), whereas channel state information (CSI) feedback is reported in a semi-statically configured and periodic manner. CSI feedback in this regard includes one or more channel quality indicators (CQI), PMI, and/or transmission rank (e.g., indicating multiple transmission layers). On the other hand, reporting on the PUSCH is dynamically triggered as part of the uplink grant. Therefore, the eNodeB can dynamically schedule CSI transmission. In contrast to the PUCCH, where the number of physical bits is currently limited to 20, reports on the PUSCH can be considerably larger. Therefore, a small codebook size is desirable for feedback on the PUCCH to suppress signaling overhead. However, for feedback on PUSCH, a larger codebook size is desirable to increase performance, since the capacity on the feedback channel is not so limited in this case.

The desired size of the codebook may also depend on the transmission scheme used. For example, a codebook used in multi-user multiple input multiple output (MU-MIMO) operation can benefit more from having a larger number of elements than a codebook used in single-user multiple input multiple output (SU-MIMO) operation. In the former case, a large spatial resolution is important to allow sufficient UE separation.

A convenient way to support different codebook sizes is to use a large codebook with many elements by default, and apply codebook subset restriction in situations where a smaller codebook is beneficial. With codebook subset restriction, the subset of precoders in the codebook is restricted so that the UE has a smaller set of possible precoders to choose from. This effectively reduces the size of the codebook, meaning that the search for the best PMI can be made on a smaller, unrestricted set of precoders, thereby also reducing the UE computational requirements for this particular search.

Typically, the eNodeB will signal the codebook subset restriction to the UE via a bitmap in the dedicated message portion of the AntennaInfo information element (see RRC specification, TS 36.331), with one bit for each precoder in the codebook, where a 1 will indicate that the precoder is restricted (meaning the UE is not allowed to select and report that precoder). Thus, for a codebook with N elements, a bitmap of length N will be used to signal the codebook subset restriction. This allows full flexibility for the eNodeB to restrict every possible subset of the codebook. Therefore, there are ^2N possible codebook subset restriction configurations.

For large antenna arrays with many antenna elements, the effective beam becomes narrower, and a codebook containing many precoders is required for the desired coverage area. Furthermore, for two-dimensional antenna arrays, the codebook size increases quadratically, since the precoders in the codebook need to span two dimensions (typically, the horizontal and vertical domains). Consequently, the codebook size (i.e., the total number of possible precoding matrices W) can be very large. Conventional signaling of the codebook subset restriction using a bitmap with one bit per precoder can therefore impose significant overhead, especially if the codebook subset restriction (CSR) is frequently updated or if there are many users served by a cell that each must receive a CSR.

Summary of the Invention

One or more embodiments herein include a method implemented by a network node for signaling to a wireless communication device which precoders in a codebook are restricted for use. The method includes generating codebook subset restriction signaling that, for each of one or more groups of precoders, jointly restricts the precoders in the group by restricting a component shared by the precoders in the group. The method further includes sending the generated signaling from the network node to the wireless communication device.

Embodiments herein also include a method implemented by a wireless communication device for decoding signaling from a network node indicating which precoders in a codebook are restricted for use. The method includes receiving codebook subset restriction signaling that, for each of one or more groups of precoders, jointly restricts the precoders in the group by restricting a component shared by the precoders in the group. The method further includes decoding the received signaling to indicate the joint restriction of the precoders in each of the one or more groups of precoders.

In some embodiments, the codebook subset restriction signaling is rank-agnostic signaling that jointly restricts the precoders in a group regardless of the transmission rank of the precoders.

In some embodiments, a component includes a beam precoder. In some embodiments, for example, the beam precoder is a Kronecker product of different beamforming vectors associated with different dimensions of a multi-dimensional antenna array. In this case, the different beamforming vectors may include discrete Fourier transform (DFT) vectors.

In other embodiments where a component includes a beam precoder, the beam precoder is a beamforming vector for transmission on a specific layer of a multi-layer transmission. Different scaled versions of that beamforming vector are transmitted on different polarizations.

In still other such embodiments, the beam precoder is a beamforming vector that is used to transmit: on multiple different layers of a multi-layer transmission; on multiple different layers of a multi-layer transmission where the layers are sent on orthogonal polarizations; or on a specific layer and on a specific polarization.

In some embodiments, if at least one of one or more beam precoders comprising a precoder of the one or more beam precoders is restricted, the precoder is restricted.

In any of these embodiments, the codebook subset restriction signaling may include a bitmap with different bits in the bitmap correspondingly dedicated to indicating whether different beam precoders are restricted from use.

Alternatively or additionally, the beam precoder may be a Kronecker product of first and second beamforming vectors with first and second indices. In this case, the first and second beamforming vectors may be associated with different dimensions of a multi-dimensional antenna array, and the codebook subset restriction signaling may jointly restrict precoders in a group of precoders having the same pair of values for the first and second indices.

In some embodiments, each precoder comprises one or more beam precoders. In some of these embodiments, each beam precoder comprises a plurality of different components corresponding to different dimensions of the multi-dimensional antenna array. In this case, a component may comprise a component of the beam precoder.

In some embodiments, the codebook subset restriction signaling jointly restricts the precoders in the group of precoders to transmit at least partially towards a certain angular pointing direction by restricting a certain component having that angular pointing direction.

Embodiments herein also include another method implemented by a network node for signaling to a wireless communication device which precoders in a codebook are restricted for use. The method includes multiple steps for each of one or more groups of precoders in the codebook. The steps include identifying one or more reference configurations for the group. Each reference configuration is one of different possible configurations that restrict the use of different subsets of precoders in the group. The steps also include identifying an actual configuration to be signaled for the group from the different possible configurations for the group. The steps also include generating signaling for indicating the actual configuration for the group by generating the signaling as a bit pattern, the length of the bit pattern depending on: (i) whether the actual configuration matches a reference configuration of the one or more reference configurations, and/or (ii) which reference configuration the actual configuration matches. The method further includes sending the generated signaling to the wireless communication device.

Embodiments herein further include another corresponding method implemented by a wireless communication device for decoding signaling from a network node indicating which precoders in a codebook are restricted for use. The method includes receiving signaling from the network node. The method further entails a plurality of steps for each of one or more groups of precoders in the codebook. The steps include identifying one or more reference configurations for the group. Each reference configuration is one of different possible configurations that restrict the use of different subsets of precoders in the group. The steps further include identifying a bit pattern defined for each reference configuration signaled, and a length of that bit pattern. The steps further include detecting an actual configuration signaled for the group by detecting a bit pattern in the signaling, the length of the bit pattern being dependent on: (i) whether the actual configuration matches a reference configuration of the one or more reference configurations, and/or (ii) which reference configuration the actual configuration matches.

In some embodiments, the signaling is a short bit pattern when the actual configuration matches any one of the one or more reference configurations, and is a long bit pattern when the actual configuration does not match any of the one or more reference configurations. The long bit pattern has more bits than the short bit pattern. In this case, the one or more reference configurations for at least one of the one or more groups may include a single reference configuration, and different long bit patterns may be defined accordingly for signaling different configurations, different from the single reference configuration. Alternatively or additionally, the long bit pattern defined for signaling the actual configuration of the group may include: (i) a non-reference bit pattern defined for signaling that the actual configuration does not match the reference configuration for the group; and (ii) a bitmap including different bits correspondingly dedicated to indicating whether different precoders in the group are restricted from use.

In some embodiments, the one or more reference configurations for at least one of the one or more groups include multiple reference configurations. In this case, when the actual configuration matches a specific one of the multiple reference configurations, the signaling is a bit pattern having a length shorter than a length of a bit pattern generated when the actual configuration matches a different reference configuration of the multiple reference configurations.

In some embodiments, the one or more reference configurations for a group have an actual or assumed higher likelihood of being signaled than any other possible configuration that is not one of the one or more reference configurations.

In some embodiments, the method is performed for a plurality of different groups, each comprising a different portion of the precoder in the codebook. In this case, the signaling indicates the actual configurations for the groups in a defined order. The one or more reference configurations for each group include a single reference configuration, and the single reference configuration for any given group is the actual configuration, if any, signaled immediately before the one for the given group.

In some embodiments, the codebook is a Kronecker codebook defined for a multi-dimensional antenna array and includes different precoders indexed by different possible values of a single index parameter. In this case, the different possible values of the single index parameter are divided into different clusters of consecutively ordered values, and the precoders in different groups of the one or more groups are correspondingly indexed by different clusters of consecutively ordered values.

In some embodiments, the codebook is a Kronecker codebook defined for a multi-dimensional antenna array and includes different precoders indexed by different pairs of possible values of a first dimension index parameter and a second dimension index parameter. In this case, the precoders in each of the one or more groups are indexed by pairs having the same value for the first dimension index parameter or the second dimension index parameter.

Embodiments herein further encompass corresponding apparatus and computer program products.

In at least some embodiments, signaling codebook subset restrictions in this manner beneficially reduces the signaling overhead imposed by communicating the codebook subset restrictions, while still allowing flexibility in configuring different codebook subset restrictions.

Embodiments herein thus generally include a method for reducing the number of bits required to signal a codebook subset restriction configuration to a wireless communication device. The method in one or more of these embodiments does so by:

Utilizing explicit or implicit assumptions about which sets of precoders are more likely to be restricted, and/or associating groups of precoders to a single codebook subset restriction bit.

BRIEF DESCRIPTION OF THE DRAWINGS

1 is a logic flow diagram illustrating codebook subset restriction (CSR) signaling between a network node and a wireless communication device in accordance with one or more embodiments.

2 is a logic flow diagram of a method implemented by a network node for signaling to a wireless communication device which precoders in a codebook are restricted for use, according to some embodiments.

3 is a block diagram of a two-dimensional antenna array of cross-polarized antenna elements in accordance with some embodiments.

FIG4 is a graph illustrating angular pointing directions of precoders in a codebook according to some embodiments.

FIG5 is a logic flow diagram of a method implemented by a network node for signaling to a wireless communication device which precoders in a codebook are restricted for use according to other embodiments.

FIG6 is a block diagram of an exemplary codebook according to some embodiments.

FIG7 is a graph illustrating angular pointing directions of precoders in a codebook according to some embodiments.

FIG8 is a block diagram of precoder grouping according to some embodiments.

9 is a logic flow diagram of a method implemented by a wireless communication device for decoding signaling from a network node indicating which precoders in a codebook are restricted for use, in accordance with some embodiments.

10 is a logic flow diagram of a method implemented by a wireless communication device for decoding signaling from a network node indicating which precoders in a codebook are restricted for use, according to other embodiments.

FIG11 is a block diagram of a network node according to some embodiments.

FIG12 is a block diagram of a network node according to other embodiments.

13 is a block diagram of a wireless communication device in accordance with some embodiments.

FIG14 is a block diagram of a wireless communication device according to other embodiments.

DETAILED DESCRIPTION

According to the flow chart of FIG1 , a network node 10 in a wireless communication network (e.g., an eNB in the network) signals a codebook subset restriction (CSR) configuration 12 to a wireless communication device 14 (e.g., a UE). The device 14 then sends a channel state information (CSI) report 16 back to the network. This CSI report 16 advises the network on which of the different possible precoders in the codebook should be used for transmission to the device 14, but the CSI report 16 is restricted in the sense that there is a subset of precoders that cannot be reported by the device 14; that is, not all precoders in the codebook can be selected and reported by the device 14. This restriction is defined by the signaled CSR configuration 12.

In more detail, for a precoder codebook X consisting of N precoders, there are ^2N possible codebook subset restriction configurations, since each precoder can be individually allowed or restricted (restricted configurations are not allowed to be used). Each configuration can be represented by an N-bit bitmap, where each bit corresponds to a certain precoder, and the value of the bit then indicates whether the precoder is restricted. If each of the ^2N configurations is equally likely and independent, this is an optimized representation of the codebook subset restriction configuration with respect to the expected length of the representation (in bits), and it provides full flexibility.

However, embodiments herein recognize that if certain configurations are more likely to be used than others, and/or if the restrictions on one precoder are highly correlated with the restrictions on another precoder, this signaling results in unnecessarily high signaling overhead. One or more embodiments herein include methods for reducing this signaling overhead; that is, reducing the number of bits required to signal the codebook subset restriction configuration from the network to the wireless communication device 14. In some embodiments, for example, the methods utilize implicit assumptions about which sets of precoders are more likely to be restricted or which sets of precoders are more likely to be jointly restricted.

According to one embodiment shown in FIG. 2 , for example, a method for signaling to a wireless communication device 14 which precoders in a codebook are restricted for use is implemented by a network node 10 (e.g., a base station). For each of one or more groups of precoders in the codebook, the method includes identifying one or more reference configurations for the group (block 110). Each reference configuration is one of different possible configurations that restrict the use of different subsets of precoders in the group. One of the reference configurations for the group can be, for example, any one of the different possible configurations with the greatest likelihood of being signaled, e.g., as predicted or estimated based on empirical observation or implicit assumptions. Regardless, the method further includes identifying an actual configuration to be signaled for the group from the different possible configurations for the group (block 120).

The method also includes generating signaling for indicating the actual configuration for the group (block 130). This entails generating the signaling as a bit pattern whose length depends on: (i) whether the actual configuration matches a reference configuration of the one or more reference configurations; and/or (ii) which reference configuration the actual configuration matches. In some embodiments, for example, when the actual configuration matches any reference configuration, the length of the bit pattern is shorter than when the actual configuration does not match any reference configuration. In other embodiments, when the actual configuration matches a specific one of the multiple reference configurations, the length of the bit pattern is shorter than when the actual configuration matches a different one of the reference configurations. In any case, this process (blocks 110-130) is repeated for each group of one or more groups of precoders in the codebook (blocks 100, 140, and 150). Finally, the method includes sending the generated signaling to the wireless communication device 14 (block 160).

This approach can be considered, in a sense, a compression algorithm for CSR signaling. Indeed, over the course of a given time period, this approach advantageously reduces signaling overhead when the overhead savings achieved by signaling bit patterns with relatively shorter lengths outweigh the overhead costs imposed by signaling bit patterns with relatively longer lengths. Depending on the relative lengths of the bit patterns, this approach can, for example, reduce signaling overhead when the one or more reference configurations (or specific reference configurations of the one or more reference configurations) are signaled more often than not.

Thus, in at least some embodiments, a reference configuration has a higher likelihood or probability of being signaled than any other possible configuration that is not the reference configuration. For example, the one or more reference configurations for a group may include whichever of the different possible configurations for the group has the highest likelihood of being signaled. Different reference configurations with different signaled probabilities may be represented by bit patterns of different lengths, with reference configurations with higher probabilities being represented by bit patterns of shorter lengths. That is, certain configurations considered more likely may be represented by a smaller number of bits, while other configurations considered less likely to be used may be represented by a larger number of bits.

In some embodiments, the one or more reference configurations may be predefined as a specific one/some configurations of possible configurations, for example, based on an (implicit) assumption that a specific configuration with the highest likelihood is signaled. For example, an implicit assumption is made about how likely the network is to be configured. Thus, certain configurations are considered more likely than others, but there are no actual likelihood values estimated for different configurations.

However, in other embodiments, network node 10 determines the signaling probabilities of different configurations, for example, based on empirical observations and comparing those probabilities to identify the configuration with the highest probability. For example, in one embodiment, the signaling probabilities are estimated using a logging method of network data. Thus, it may be possible to estimate the actual probabilities for different configurations. Thus, generally speaking, knowledge of how "probable" a configuration is can be obtained in a number of ways.

In some embodiments, only a single reference configuration is defined for a group. In this case, signaling is generated as a short bit pattern when the actual configuration matches the reference configuration, and is generated as a long bit pattern when the actual configuration does not match the reference configuration. Different long bit patterns are defined accordingly in this regard for signaling different configurations (unlike the reference configuration, for which short bit patterns are defined for signaling). Long bit patterns naturally have more bits than short bit patterns (e.g., N bits versus 1 bit).

In other embodiments, multiple reference configurations are defined for a group. In this case, signaling can be generated as bit patterns of different lengths when the actual configuration matches different reference configurations. These lengths can correspond to how likely it is that a reference configuration will be signaled. The length of the bit pattern can be shortest when the actual configuration matches a specific one of the reference configurations (e.g., the one with the highest signaled probability), next shortest when the actual configuration matches a different reference configuration (e.g., the one with the next highest signaled probability), and longest when the actual configuration does not match any reference configuration.

In some embodiments, the bit pattern that signals the non-reference configuration is encoded as a combination of a so-called "non-reference bit pattern" and a "bitmap". The non-reference bit pattern is defined to signal that the actual configuration for the group does not match any reference configuration for the group. The non-reference bit pattern may, for example, be a complement to the bit pattern defined to signal the reference configuration. For example, when only a single reference configuration is defined for the group, the bit pattern that signals that reference configuration may simply be a single bit with a value of "1", while the non-reference bit pattern may be a single bit with a value of "0". In any case, the bitmap part of the bit pattern includes different bits that are respectively dedicated to indicating whether different precoders in the group are restricted for use.

In at least some embodiments, the method is performed for only one group. In one embodiment, this single group includes all precoders in the codebook.

In another embodiment, of course, a single group contains only a portion of the precoders in the codebook, so that signaling means are taken for only this portion, while other signaling means (eg, conventional bitmaps) are taken for the other portions.

In other embodiments, the method is performed for a plurality of different groups, each comprising a different portion of the precoder in the codebook. In one such embodiment, the signaling indicates the actual configurations for the groups in a defined ordering. In one embodiment, the one or more reference configurations for any given group include the actual configuration, if any, signaled immediately before the one for the given group (according to the defined ordering).

Consider the example of an arbitrary codebook of size N, where a single group contains all N precoders. One configuration from the ^2N possible codebook subset restriction configurations for a single group is considered more likely. This configuration is represented by a single bit, '1'. The other ^2N -1 configurations are represented by '0', followed by a bitmap of size N. One of the configurations is then represented by 1 bit, while the other is represented by N+1 bits. Because the configuration represented by one bit is signaled more frequently, it is assumed that the average number of bits required to convey the codebook subset restriction can be significantly less than N.

However, if the assumption that one of the possible codebook subset restriction configurations is more likely than the other configurations is incorrect for actual use of the codebook subset restriction configuration, the average number of bits required to convey the codebook subset restriction to the UE may be greater than N bits. One or more embodiments herein therefore aim to choose a good representation of the ^2N configurations. Various methods may represent the ^2N configurations differently depending on which sets of precoders are more likely to be restricted.

Consider, for example, an embodiment in which a codebook is defined for a multi-dimensional (e.g., two-dimensional) antenna array. Such an antenna array can be (partially) described by the number of _antenna columns Mℎ corresponding to the horizontal dimension , the number of antenna _rows Mvv corresponding to the vertical dimension, and _the number of dimensions Mp corresponding to different polarizations. The total number of _antennas is therefore M = MℎMvMp _. It should be noted that the concept of antenna is non-limiting in that it can refer to any virtualization (e.g., a linear mapping ₎ of a physical antenna element. For example, pairs of physical sub-elements can be fed the same signal and thus share the same virtualized antenna port.

An example of a 4x4 array with cross-polarized antenna elements is shown in Figure 3. Specifically, assuming that one antenna element corresponds to one antenna port, Figure 3 shows a two-dimensional antenna array _{( Mp} = 2 ₎ with cross-polarized antenna elements with Mℎ = 4 horizontal antenna elements and Mv = 4 vertical antenna elements.

Precoding can be interpreted as multiplying the signal by different beamforming weights for each antenna prior to transmission. A typical approach is to tailor the precoder to the antenna forming factor, ie, taking into account Mℎ , _{Mv ,} and Mp when designing the _precoder codebook.

According to some embodiments, the precoder codebook is tailored for a 2D antenna array by combining precoders tailored for the horizontal array and the vertical array respectively by means of a Kronecker product. This means that (at least part of) the precoder can be described as the following function:

W _H ⊗ W _V

where WH is _a horizontal precoder taken from _{a (} sub-)codebook XH _containing NH codewords, and similarly WV is a vertical precoder taken from _a (sub-)codebook XV containing NV codewords. _{The joint codebook denoted by XH ⊗ XV thus contains} NH · NV codewords. The elements of XH are indexed with k = 0, ..., NH _-1 , the elements _of XV are indexed with l = 0, ..., _{NV -1} , and the elements _{of the} joint codebook XH ⊗ XV are indexed with m = NV · k + _{l (} implying m = ₀ , ..., NH · NV _{-1 )} .

In some embodiments, for example, a (sub)codebook of the Kronecker codebook is composed of a DFT precoder. In this case, the horizontal codebook can be represented as, where Qℎ _is an integer horizontal oversampling factor and can accept values in the interval 0 to 1 in order to "shift" the beam pattern (=0.5 can be an interesting value for creating beam symmetry with respect to the broadside of the array). And the vertical codebook can be represented as, where Qv is _an integer vertical oversampling factor and is defined similarly as above.

It should be noted that the precoder codebook can be defined in several ways. For example, the Kronecker codebook mentioned above can be interpreted as a codebook indexed by a single PMI m . Alternatively, it can be interpreted as a single codebook indexed by two PMIs k and l . It can also be interpreted as two separate codebooks, indexed by k and l respectively. Furthermore, the Kronecker codebook discussed above may only describe a portion of the precoder, i.e., the precoder may also be a function of other parameters. In such an example, the precoder is also a function of another PMI n . Again, this can be interpreted as three separate codebooks with indices k , l , and n , respectively, or two separate codebooks with indices m = NV · k + l and n , _respectively . It can also be interpreted as a single joint codebook with a joint PMI. The embodiments herein should be considered agnostic with respect to how the codebook is defined.

With this understanding, the codebook at issue in FIG. 2 can be a Kronecker codebook comprising different precoders indexed (at least in part ₎ by different possible values of a single index parameter (e.g., index parameter m = 0, ..., NH · NV - ₁ ). In this case, the different possible values of the single index parameter are divided into different clusters of consecutively ordered values. And the precoders in different groups are respectively indexed (at least in part) by different clusters of consecutively ordered values. For example, precoders indexed by cluster m = 0, ..., m 1 belong to a first group, precoders indexed by cluster m = m 2, ..., m 3 belong to a second group, precoders indexed by cluster m = m 4, ..., m 5 belong to a third group, and so on. As an even more specific example, one or more embodiments exploit the Kronecker structure of the precoders by mapping _index m to index k and l when m = Nvk + l and grouping the precoders such that m = 0, ..., _NV -1 is the first group, m = _NV , ..., _2NV -1 is the second group, and so on.

In another embodiment, by contrast, the Kronecker codebook includes different precoders indexed (at least in part) by different pairs of possible values for a first dimension index parameter (e.g., k = 0, ..., NH _- 1) and a second dimension index parameter (e.g., l = 0, ..., NV _- 1). In this case, the precoders in each of the different groups are indexed (at least in part) by pairs ( k , l ) having the same value for the first dimension index parameter k and/or the second dimension index parameter l .

Two different embodiments, referred to as "row-like embodiments" and "column-like embodiments," will now be illustrated in this regard in the context of a Kronecker codebook, where only a single reference configuration is defined for a group. The Kronecker codebook in this example consists of precoders with different angular directions, spanning a two-dimensional angular region as seen from the transmitter. An important use case for codebook subset restriction in such embodiments can be to restrict the precoders in a certain angular region or angular interval (e.g., corresponding to the direction in which user hotspots of neighboring cells are located). If the precoders corresponding to the beam pointing in that direction are restricted, the eNodeB will subsequently reduce interference with the neighboring cells and the specific hotspot region. This is beneficial from a system capacity perspective.

Below, we consider a specific example in which codebook subset restriction is used on the Kronecker codebook to facilitate understanding of how different embodiments can be used to reduce signaling overhead. In this case, a 4x4 antenna array with a mechanical downtilt of 18 ^° is used. The Kronecker codebook consists of 8 vertical and 8 horizontal precoders, i.e., N _H = N _V = 8. The angular pointing directions of the precoders in the codebook are shown in Figure 4.

Codebook subset restriction is applied to restrict beams with pointing directions in the top interval [85 ^∘ ,95 ^∘ ] (shown with dashed lines). That is, codebook subset restriction is applied in the angular interval 85 ^∘ <θ <95 ^∘ , meaning that precoders with indices ( k , l ) = (0, 4), (3, 5), (4, 5), (7, 4) are restricted. These restricted beams are shown with 'o', while unrestricted beams are shown with 'x'. Beam index k in the horizontal codebook and beam index l in the vertical codebook are written next to the beam as ( k , l ). If this configuration of codebook subset restriction were _to be signaled via a conventional bitmap, then N = NH · NV ₌ 64 bits would be used.

"Similar row embodiment"

In one embodiment, by using compression of CSR signaling, the scheme is designed with the assumption that precoders ( k , l ) with adjacent l indices (i.e., ( k , l ₀ -1), ( k , l ₀ ) and ( k , l ₀ +1 )) are likely to have the same restriction settings, meaning that if ( k , l ₀ ) is restricted, then ( k , l ₀ +1) is also likely to be restricted, and vice versa. The scheme works as follows:

First, a bitmap of NH _bits is sent, indicating the codebook subset restriction for the "row" of precoders where l = 0 (see Figure 4), i.e., precoders ( k , l ) = (0,0), (1,0), ..., ( NH _- 1,0).

Then, the codebook subset restrictions for the second "row" of precoders where l = 1 are sent. If the restrictions are the same as the restrictions for the previous row of precoders, a '1' is sent. If the restrictions for this row are different from the restrictions for the previous row, a '0' is sent, followed by a bitmap indicating the restrictions for this row.

The previous steps are then repeated for each of the N _V "rows" of the precoder.

This embodiment is illustrated by way of example, considering the codebook subset restriction setting shown in FIG4 , ie, the restriction of precoders with indices ( k , l )=(0,4),(3,5),(4,5),(7,4) should be signaled.

For l = 0:

No precoder with index 0 should be restricted, so the bitmap '00000000' is sent.

For l = 1:

The limits for this row are the same as those for the previous row, bit ‘1’ is sent.

For l = 2:

The limits for this row are the same as those for the previous row, bit ‘1’ is sent.

For l = 3:

The limits for this row are the same as those for the previous row, bit ‘1’ is sent.

For l = 4:

The restrictions for this row are different from those for the previous row, so bit '0' is sent. A bitmap indicating the restrictions for this row should now be sent. Precoders (0, 4) and (7, 4) should be restricted. Therefore, a bitmap of '10000001' is sent.

For l = 5:

The restrictions for this row are different from those for the previous row, so bit '0' is sent. The bitmap indicating the restrictions for this row should now be sent. Precoders (3, 5) and (4, 5) should be restricted. Therefore, the bitmap '00011000' is sent.

For l = 6:

The restrictions for this row are different from those for the previous row, so bit '0' is sent. A bitmap indicating the restrictions for this row should now be sent. No precoder should be restricted. Therefore, a bitmap of '00000000' is sent.

For l = 7:

The restrictions for this row are the same as those for the previous row, bit ‘1’ is sent.

The string of bits to be signaled is therefore '00000000011101000000100001100000000000001', consisting of 39 bits. In general, the number of bits required for this scheme is

N _{bibits =} M · NH + NV _-1

where M is the number of times a row changes and the bitmap for that row must be transmitted; in _this example, M = 4. Analyzing the above expression, we note that 1 ≤ M ≤ NV . This means that for some of the possible codebook subset restrictions, the number of bits required to signal the codebook subset restriction using this scheme is less than N , while for other restrictions, such as when M = NV _, the number of bits required is greater than M.

Note that this is a small example for the sake of illustrating the embodiments. If _a larger codebook is used, say NH = NV = 30 and M = 4, the number of bits required for this scheme would _{be Nbibits} = M _NH + NV - ₁ = 149, compared to N = NH · NV = 900 in the case where _{just the} entire _bitmap is transmitted; this is therefore a substantial reduction in the number of bits required.

Finally, it is noted that all possible codebook subset restriction configurations can be represented by this encoding/decoding scheme, thus providing full flexibility.

"Similar columns" example

In another embodiment, the scheme discussed in the previous embodiment is modified by instead taking into account the assumption that precoders ( k , l ) with adjacent k indices (i.e., ( _{k0-1 ,} l ), ( k0 , l ), and ( k0 ₊ 1 , l )) are likely to have the same restriction settings, meaning that if ( k0 _, l ) is restricted, ( k0 ₊ 1, l ) is also likely to be restricted, and vice versa. The structure of the string of bits to be signaled will then work similarly as in the previously discussed embodiment, except that the precoder "column" k will be used instead.

In another embodiment, additional initial bits are inserted, where '1' indicates that encoding is done under the assumption that precoders ( k , l ) with adjacent l indices (i.e., ( k , l _{0 -1} ), ( k , l 0 ) and ( k , l ₀ +1)) are likely to have the same restrictions, so encoding is done row wise, and '0' indicates that precoders ( k , l ) with adjacent k indices (i.e., ( k ₀ -1, l ), ( k ₀ , l ) and ( k ₀ +1, l )) are likely to have the same restriction settings, so encoding is done column wise.

In another embodiment, an initial bit is inserted where '1' indicates that no precoder is restricted, '0' indicates that some precoders are restricted and '0' is followed by a number of bits representing the codebook subset restriction.

Thus, different "compression" techniques (whether based on similar rows, columns, or other means) can be employed for different groups of precoders in the same codebook, with the specific technique indicated to the device so that the device can decode the signaling. Alternatively, the same "compression" technique can be employed for each group of precoders, but the network evaluates the different possible techniques to identify the one that provides the best compression and then employs that technique (and indicates it to the device).

Of course, the embodiment shown in Figure 2 and its variants can be used to signal a restricted subset of precoders in any given codebook, whether Kronecker constructed or not. In addition, the signaling can be rank-specific, meaning that different signalings restrict different rank-specific codebooks.

According to another embodiment shown in FIG. 5 , a method is implemented in a network node 10 (e.g., a base station) for signaling to a wireless communication device 14 which precoders in a codebook are restricted (e.g., which Kronecker product precoders are restricted) to use. As shown, the method includes generating codebook subset restriction signaling that, for each of one or more groups of precoders, jointly restricts the precoders in the group, e.g., with a single signaling bit (block 210). In at least some embodiments, this signaling (i) is rank-agnostic so as to restrict the precoders regardless of their transmission rank; and/or (ii) jointly restricts the group of precoders (i.e., the precoders in the group) by restricting a component that the precoders (i.e., the precoders in the group) have in common. In any case, the method then includes sending the generated signaling to the wireless communication device 14 (block 220).

Consider an embodiment in which a group of precoders (i.e., the precoders in the group) is jointly restricted by restricting a certain component that the precoders have in common. Precoders share a certain component if the precoders are derived from or otherwise act as a function of that same component. In one embodiment, for example, by restricting a certain component b , the group of precoders W ( b ) that share that component b is jointly restricted. This restriction of component b can be signaled, for example, according to one or more indices for the component (e.g., ( m, where the component is indexed as bm or ( k , l ) where the component is indexed as _{bk , l} , where m , k , and l are indices into the Kronecker-constructed codebook as described above).

Note that embodiments herein contemplate precoders having one or more different "components" at any level of granularity (e.g., components of precoder factorability at a high level and/or components of precoder factorability at a lower level). For example, a precoder may include one or more different components b at one level of granularity. However, at a finer level of granularity, each of these components b may in turn be derived from or otherwise be a function of multiple subcomponents x _H and x _V , such that b ( x _{H ,} x _V ) . In this case, a group of precoders W ( x _{H ,} x _V ) that share a certain component x _H or x _V may be jointly constrained by constraining that component x _H or x _V. The constraint on this component x _H or x _V may be signaled, for example, by an index into the component (e.g., k or l , where component x _H is indexed as and component x _V is indexed as , x _H and x _V are the horizontal and vertical beamforming vectors, respectively, and k and l are indices into the Kronecker-constructed codebook as described above).

In some embodiments, a precoder at one level of granularity is composed of one or more distinct components, referred to as one or more so-called "beam precoders." Each precoder W, in this regard, is composed of one or more beamforming vectors, referred to _as beam precoders _, b0 , b1 , ..., bX . One or more embodiments herein jointly restrict the set of precoders W that share that beam precoder by restricting a particular beam precoder. _By restricting precoder W (based on the restriction of one or more of its constituent beam precoders as a whole), these embodiments advantageously generate CSR signaling according to beam-specific restrictions (i.e., restrictions on certain beam precoders) rather than according to precoder-specific restrictions (i.e., restrictions on precoder W as a whole). In some embodiments, device 14 assumes that precoder W is restricted if one or more of its beam precoders are restricted. In other embodiments, each beam precoder must be restricted for device 14 to assume that the overall precoder W is restricted.

In one embodiment, the beam precoder is a beamforming vector for transmission on a specific layer, where different scaled versions of that beamforming vector are transmitted on different polarizations. Different layers are transmitted on different beam precoders. In this case, the precoder W can be expressed as:

Here, W is an N x L precoder matrix, where N is the number of transmit antenna ports, L is the transmit rank (i.e., the number of transmitted spatial streams), b ₀ , b ₁ , …, b _{L -1} are beamforming vectors (referred to as beam precoders), and is an arbitrary complex number. Another precoder W with the same codebook as W above can be expressed as:

.

For example, by signaling b ₀ , only the former precoder is restricted, and by signaling b ₁ , both precoders will be restricted.

In some embodiments, the first antenna port is mapped to an antenna with one polarization, while subsequent antenna ports are mapped to antennas with the same position as the first antenna, but with orthogonal polarizations. In such embodiments, for each column of W (i.e., for the precoder of each spatial layer), the beam precoder b is transmitted on one polarization, and a scaled version of the same beam precoder is transmitted on a second polarization. Such scaling can affect the phase, amplitude, or both phase and amplitude of the beam precoder.

In another embodiment, the beam precoder is a beamforming vector used to transmit on multiple different layers, where the layers are sent on orthogonal polarizations. In this case, the precoder W can be expressed as:

Therefore, it should be noted that the beam precoders b ₀ , b ₁ , ..., b _{L −1} for each spatial layer may be different beam precoders, or some subsets of beam precoders may be the same, eg, b ₀ may be equal to b ₁ .

In yet another embodiment, a beam precoder is a beamforming vector for transmitting on a specific layer and on a specific polarization. That is, the beam precoder can be defined in a slightly different way than the above definition. The definition of the beam precoder can, for example, allow different beam precoders to be transmitted on different polarizations of the same layer, such as:

.

In yet another embodiment, the beam precoder can be defined by ignoring polarization as

.

Note _that the beam precoders b0 , b1 , ..., bL _{- 1} can be explicitly chosen from a set _of beam precoders (codebook), or they can be implicitly chosen when selecting the (overall) precoder W from the codebook X. It should be noted that the selection of the (overall) precoder W can be made using one or several PMIs. In the case where the selection of the overall precoder W is made using several PMIs, the resulting beam precoders for each layer can be a function of only a subset of the PMIs, or they can be a function of all PMIs.

However, regardless of the specific manner in which beam precoders are defined, one or more embodiments herein jointly restrict the set of precoders W that share that beam precoder by restricting a particular beam precoder. That is, in some embodiments, codebook subset restriction (CSR) can be signaled based on the set of possible beam precoders b , instead of signaling the CSR over the set of possible (total) precoders W. In some such embodiments, device 14 should assume that precoder W is restricted if one or more of the beam precoders b ₀ , b ₁ , ..., b _{L -1} for each layer are restricted. In other such embodiments, the beam precoder for each layer must be restricted for device 14 to assume that the total precoder W is restricted.

Consider the specific example of an 8TX codebook with a transmission rank of 2. In some embodiments, this codebook is defined as shown in FIG6 . Defined in this manner, each precoder W is partially formed by a beam precoder v _m (note the transition from b ₀ , b ₁ , …, b _{L -1} to v _m ). The beam precoder index m is the same for some precoders W, including, for example, precoders whose subcodebook index i ₂ is equal to 0, 1, 8, 9, 12, or 13 (because m = 2 i ₁ for those precoders). This means that those precoders W share the same beam precoder v _m . Therefore, some embodiments herein jointly restrict the group of precoders W that share that beam precoder v _m by restricting a specific beam precoder v _m (e.g., using a single bit). This restriction of the beam precoder v _m can be signaled, for example, according to the index m (e.g., the beam precoder indexed by a specific value of m is restricted). In this case, the signaling may constitute a bitmap, with different bits in the bitmap being dedicated to indicating whether different beam precoders are restricted for use. For example, the signaling may constitute a bitmap of m values, with different bits in the bitmap being dedicated to indicating whether beam precoders indexed with different m values are restricted for use.

In an alternative embodiment not shown in FIG6 , the beam precoder v _m is replaced by a beam precoder v _{k , l} , which is the Kronecker product of the vertical beamforming vector x _V with index k and the horizontal beamforming vector x _H with index l . For example, as noted above, these beamforming vectors may comprise DFT vectors. Regardless, the restriction of the beam precoder v _{k , l} may be signaled per index pair ( k , l ). In this case, the signaling may constitute a bitmap of ( k , l ) value pairs, with different bits in the bitmap correspondingly dedicated to indicating whether the beam precoder indexed by a different ( k , l ) value pair is restricted from use.

Instead of such a bitmap, the restrictions on one or more beam precoders vk _{, l are ,} in some embodiments, jointly signaled according to a "rectangle" defined by _two ( k , l ) value pairs, namely, ( k0 , l0 ) and ₍ k1 , l1 ). _In this case, beam precoders vk _{, l} with _{indices k0 < k < k1} and _l0 < l < _l1 are restricted .

As yet another alternative, in some embodiments, the restrictions on one or more beam precoders v _{k , l} are signaled as a bitmap of k values and/or a bitmap of l values. If a bitmap of only k values is signaled, then in some embodiments, the device assumes that any beam precoder v _{k , l} with certain k values is restricted, regardless of the l values of those precoders. If a bitmap of only l values is signaled, then in some embodiments, the device assumes that any beam precoder v _{k , l} with certain l values is restricted, regardless of the k values of those precoders. If both a bitmap of k values and a bitmap of l values are signaled, then in some embodiments, the device assumes that only beam precoders v _{k , l} with certain ( k , l ) value pairs, as correctly defined by those bitmaps, are restricted.

Even so, the limits specified in terms of k and/or l values can in some sense be thought of as limits at a finer level of granularity than even the beam precoders themselves. Indeed, as noted above, each beam precoder vk _{, l} is, in some embodiments, the Kronecker product _of a vertical beamforming vector xV with index k _and a horizontal beamforming vector xH _with index l . Therefore, signaling limits as k and/or l values effectively amounts to limiting the (sub)components xV or xH _.

Consider an example of these finer-grained embodiments, where codebook subset restriction is to be applied to beam precoders with l values of 3 or _{4. If this configuration of codebook subset restriction were to be signaled using a conventional bitmap, N = NH ·NV} = 64 bits would be used. By contrast, the schemes in these finer-grained embodiments consider restriction of an entire precoder "row," i.e., all precoders formed by beam precoders with the same l _index are turned on or off. Thus, in this example, to signal the codebook subset restriction, a bitmap of NV = 8 bits consisting of l values, '00011000', can be sent. With this scheme, a significant reduction in the number of bits required to signal the codebook subset restriction is seen. However, not all of the ^2N possible codebook subset restrictions can be signaled.

In a similar embodiment, the restriction is applied on a precoder "column" k , and the codebook subset restriction is signaled using a bitmap that is NH bits long, indicating the restriction for the entire precoder "column _" .

In another embodiment, additional initial bits are inserted, where ‘1’ indicates that the encoding is done “row-wise” as above, and ‘0’ indicates that it is done “column-wise”.

In yet another embodiment, the apparatus 14 should assume that if both the vertical and horizontal precoders in the Kronecker structure are limited, then the precoder W is limited. If only one of the vertical and horizontal precoders is limited, the apparatus 14 should not assume that the resulting precoder after the Kronecker operation is limited.

Therefore, one or more embodiments herein advantageously utilize the Kronecker structure of the codebook to generate the signaling of Figure 5 according to indices k , l , and/or m . In some embodiments, for example, the signaling is generated to jointly restrict (e.g., with a single bit) the set of precoders that (i) have the same value for index k ; (ii) have the same value for index l ; or (iii) have the same pair of values for index ( k , l ).

In some embodiments, the signaling that jointly restricts the group of precoders by restricting some component that the precoders have in common (e.g., beam precoders) is rank-agnostic. That is, the signaling jointly restricts the group of precoders regardless of the transmission rank of the precoders (i.e., regardless of which rank-specific codebook they belong to). For example, the embodiment _of restricting a single beam precoder b0 can be extended so that all precoders across all ranks containing _the restricted beam precoder b0 are restricted. Therefore, all precoders across all ranks _containing a certain beam precoder b0 are the group of precoders that can be jointly restricted. Therefore, according to some embodiments, a benefit of signaling CSR based on beam precoders is that we do not need to signal separate CSRs for precoders with different ranks (precoders with different ranks are restricted with the same CSR). This reduces signaling overhead.

Signaling that jointly limits the group of precoders by limiting a certain component that the precoders have in common has also proven effective for limiting precoders that transmit, in whole or in part, toward certain angular pointing directions. Indeed, according to some embodiments herein, network node 10 jointly limits the group of precoders that transmit, at least in part, toward certain angular pointing directions by limiting a certain component (e.g., beamforming precoders) that has a certain angular pointing direction. In this way, network node 10 avoids transmitting energy in a certain direction by signaling, via the CSR, to device 14 that device 14 should not calculate feedback for that particular direction.

More particularly in this regard, when each precoder W is formed from multiple beam precoders, precoder W, in one sense, has multiple angular pointing directions corresponding to the angular pointing directions of its constituent beam precoders (e.g., where each beam precoder has its own azimuth and apex pointing directions). However, in another sense, precoder W has an overall angular pointing direction that is a combination (e.g., an average) of the corresponding directions of its beam precoders. By limiting beam precoders having certain angular pointing directions, embodiments herein effectively limit precoders that transmit at least partially in those directions, and do so with reduced signaling overhead.

As an example, a set of rank 1 precoders with the same angular pointing direction but with different polarization properties, such as the entire set of rank 1 precoders

,

Limitation can be achieved by signaling a limit for a single beam precoder _, b0 . That is, when a limit is signaled for a particular beam precoder, the limit implicitly applies to all polarization phases of the signaled beam. Therefore, the group of rank-1 precoders illustrated above is associated with a single CSR bit and is therefore jointly limited. This reduces device complexity and CSR signaling overhead, as only the beam direction needs to be signaled.

In another example, the set of rank 1 precoders

,

can be jointly restricted by restriction signaling of a single beam precoder b _0. Thus, the group of rank 1 precoders illustrated above is associated with a single CSR bit and is therefore jointly restricted.

Restricting precoders with certain angular pointing directions can also be accomplished by specifying restrictions based on certain values of k and/or l . This is illustrated with reference to Figure 7, which shows the angular beam pointing directions for a rank-1 precoder in an example codebook. In this example, the network node has a 4x4 antenna array with no mechanical downtilt. The Kronecker codebook consists of 8 vertical and 8 horizontal precoders, i.e., N _H = N _V = 8. In this example, codebook subset restriction is applied to restrict beams with pointing directions in the top interval [80 ^∘ , 100 ^∘ ] (the interval is shown with a dashed line). That is, the codebook subset restriction is applied in the angular interval 80 ^∘ < θ < 100 ^∘ , so that precoders with indices 1 - 3 and 4 are restricted. Restricted beams are indicated by 'o', while unrestricted beams are indicated by 'x'. The beam index k in the horizontal codebook and l in the vertical codebook are written as ( k , l ) next to the beam. Therefore, to signal the codebook subset restriction in this example, a bitmap of l values '00011000' consisting of NV ₌ 8 bits can be sent. With this scheme, a significant reduction in the number of bits required to signal the codebook subset restriction is seen.

In another embodiment, the device 14 shall assume that the precoder is limited if both the vertical and horizontal precoders in the Kronecker structure are limited. This allows limiting a rectangular "window" of beamformer pointing angles as seen from the network node 10.

This can also be done by signaling the restriction of a "rectangle" of precoders as defined by _the index pairs ₍ k0 , l0 ) and ₍ k1 , l1 ₎ . With this scheme, precoders _with indices k0 _< k _< k1 and _l0 < l < l1 are restricted.

The component-based restriction of the precoder group is just one example of an embodiment for providing rank-agnostic CSR signaling. Other embodiments herein also provide such rank-agnostic signaling. For example, by generating signaling to (explicitly or implicitly) indicate a certain/certain angular pointing directions, some embodiments herein generate signaling to jointly indicate that the group of precoders transmitted in whole or in part in that/those angular pointing directions is restricted. The signaling may specify the restricted angular region or interval, for example, according to one or more angular parameters. This restriction may relate to the angular pointing directions of the precoder as a whole, or the angular pointing directions of any beam precoder forming the precoder.

In one embodiment, the angular region or interval may be represented by the corner points α and α, spanning a rectangle in the angular domain. Here, α and α are the azimuths and top angles respectively relative to the eNodeB. Although the present embodiment focuses on the case of a single rectangular region for simplicity, multiple such rectangular regions may be signaled. The device 14 may then compute the angular pointing directions of the precoders in the codebook and compare them to the restricted angular regions to derive the codebook subset restriction. The device 14 may need some additional information about what is to be assumed about the transmitter antenna array (which does not need to correspond to the antenna array actually used) to be able to compute the pointing directions of the precoders. Consider an exemplary embodiment in which the (sub)codebook of the Kronecker codebook consists of a DFT precoder, i.e.

The horizontal codebook can be expressed as

, , where Qℎ is an integer horizontal oversampling factor and can take values in the interval 0 to 1 in order to “shift” the beam pattern (=0.5 is an interesting value _for creating symmetric properties of the beam with respect to the broad side of the array).

The vertical codebook can be expressed as, where Qv is _an integer vertical oversampling factor and is defined similarly as above.

The pointing direction of the precoder ( k , l ) can be calculated by first calculating the pointing angle relative to the broadside of the antenna array:

Where dV and dH are the vertical and horizontal antenna element spacings _of the array in wavelengths, respectively. The mechanical downtilt angle is taken into account to calculate _the actual beam pointing angle as follows:

The device 14 needs to be signaled with additional information d _V , d _H and d H to be able to compute the beam steering directions of the precoder in the codebook. It is assumed that the device 14 already knows the parameters Q _v , M _v , Q _ℎ , M _ℎ and d V as part of the codebook structure.

The set of parameters , , , d _H , d _V , thus parameterizes the codebook subset restriction in this embodiment. When signaling the parameters, several strategies can be used.

In one embodiment, each parameter is uniformly quantized using a number of bits over a predefined interval. Examples are given in the table below.

parameter interval Quantization bits [0,180][degree] 6 [0, 2] 4 [-30, 30][degree] 6

In this embodiment, the number of bits required to signal the codebook subset restriction is 38. Note that this is independent of the codebook size.

In another embodiment, each parameter may take a value from a fixed set of possible values. Each possible value of the parameter is encoded using a different number of bits (depending on, for example, the perceived likelihood of the parameter taking that value). For example, the horizontal array element spacing d _H may be encoded as follows:

value 0.5 0.8 0.65 1 4 2 0.75 Bit 1 01 0011 0010 0001 00001 00000

In this embodiment, the encoding of d _H is designed to take into account that d _H = 0.5 is a common value for horizontal antenna element separation, so this value is encoded with a low number of bits. Other less common values are encoded with a larger number of bits. Note that the encoding of d _H in this embodiment constitutes a uniquely decodable code.

In another embodiment, some of the parameters are uniformly quantized using a number of bits over a predefined interval, while other parameters are encoded using a different number of bits as in the previous embodiment.

In some other embodiments, different sets of parameters relating to the restricted angular regions may constitute the parameters defining the codebook subset restriction. In one such embodiment, only the azimuth interval is restricted, and therefore, , may be transmitted. In another such embodiment, the restriction is only to the azimuth interval. In still another such embodiment, the angular interval may be open-ended, i.e., constitute a restriction.

In other embodiments, parameters related to the antenna array, such as d _H , d _V and d V are not part of the codebook subset restriction parameters, but instead they may be already known to the UE, or the UE assumes default values for said parameters, and the eNodeB chooses the restriction angles d H and d V in such a way that the intended precoder is restricted when the UE computes the restrictions based on the default values for said parameters, which may be different from the actual values of said parameters.

In other embodiments, more parameters may be included in the codebook subset restriction parameters. In one such embodiment, the roll angle of the antenna array may be included in the codebook subset restriction parameters.

In light of the above modifications and variations, we recognize that there are many ways in which CSR signaling can jointly constrain the precoders in a group. The signaling can be rank-agnostic or not. And the signaling can constrain a signal angle parameter associated with the group or some component common to the group. The signaling can take the form of a bitmap of beam precoder indices, an angle parameter, a pair of subcodebook indices, a bitmap of indices for a single subcodebook, and so on. However, regardless of these specific variations, CSR signaling overhead is reduced based on equivalent grouping of precoders or correlation of precoder restrictions. However, group-based joint restrictions mean that not all ^2N codebook subset restriction configurations may be delivered to device 14. Instead, only a subset of possible configurations can be selected.

Thus, by performing joint restrictions with respect to only a subset of the precoders in the codebook, at least some embodiments balance the loss in flexibility caused by such joint restrictions with the signaling overhead gain through joint restrictions. That is, codebook subset restrictions can be configured with full flexibility on a subset A of precoders in the codebook (meaning each precoder can be individually turned on or off), while only a few configurations can be chosen for the remaining set B of precoders. For example, the codebook subset restriction for the remaining set B of precoders can be expressed with only one bit, turning all precoders in the set on or off. This reduces CSR signaling overhead, which is beneficial.

As an example in the context of beam precoders, a codebook can consist of two sets of precoders. One set consists of precoders that can be equivalently expressed as a function of a layer-specific beam precoder (as defined above), while the other set can consist of arbitrary precoders. In this embodiment, the precoders in the first set can be configured with full flexibility, while the other precoders in the codebook can be configured with limited flexibility.

This embodiment is just one example of a grouping of precoders in a codebook, where the precoders belonging to set A are each represented by a single bit, while the precoders in set B are all jointly restricted using a single bit. This embodiment can be further extended by having multiple sets B, such as B_1, B_2, ..., B_N, where each set B_n, n=1, ..., N, contains at least two precoders and is associated with a CSR bit. An example is shown in Figure 8, where precoders 1 to 14 are each represented by a separate bit (set A), while all precoders in group B1 are represented by a single CSR bit, such as the bit for precoder 15.

The defined groups may also overlap, such that a given precoder is present in multiple groups. If this is the case, then priority or combining rules need to be defined so that the device 14 understands how to interpret the situation when a precoder is restricted by signaling from one group but not from another group to which it belongs.

Therefore, in a further detailed embodiment, the groups B_n in FIG8 can overlap, and the rules are specified in standard text regarding how the device 14 should interpret CSR signaling. For example, assume two groups B_1 and B_2, each represented by one bit, and a precoder belongs to both groups. One rule could be that if a precoder is restricted in any group it belongs to, the precoder should be assumed to be restricted. Another alternative is that the precoder must be restricted in both groups for the precoder to be assumed to be restricted.

In some embodiments of this disclosure, codebook subset restrictions are discussed using the terms precoder and codebook . It can be assumed that beam-specific restrictions are used in the embodiments described, and the terms beam precoder and set of beam precoders can be interchanged, depending on the granularity being discussed.

Note that although terminology from 3GPP LTE has been used in this disclosure to illustrate embodiments herein, this should not be seen as limiting the scope of the embodiments to only the aforementioned systems. Other wireless systems including WCDMA, WiMax, UMB and GSM may also benefit from utilizing the ideas covered in this disclosure.

Note also that terms such as eNodeB and UE should be considered non-limiting and do not specifically imply a hierarchical relationship between the two terms; generally speaking, "eNodeB" can be considered device 1 and "UE" is device 2, and the two devices communicate with each other over a radio channel. Here, we also focus on wireless transmission in the downlink, but the embodiments herein are equally applicable in the uplink.

The embodiments herein also include methods in the wireless communication device 14 corresponding to the methods described above in the network node 10. These methods receive and decode signaling generated by the network node 10 according to any of the above embodiments.

According to one embodiment shown in FIG9 , for example, a method implemented by a wireless communication device 14 (e.g., a UE) is used to decode signaling from a network node 10 indicating which precoders in a codebook are restricted for use. The method includes receiving the signaling (block 300 ). The method also includes, for each of one or more groups of precoders in the codebook, decoding the signaling to identify which of different possible configurations is actually signaled for that group. The different possible configurations restrict the use of different subsets of precoders in the group. This decoding is performed on a group-by-group basis, starting with the first group (block 310 ). Specifically, the decoding entails identifying one or more reference configurations for the first group, the bit pattern identified for each signaled reference configuration, and the length of that bit pattern (block 320 ). The reference configuration(s) may be predefined at the device 14 or may be signaled from the network node 10 . In any event, by detecting a bit pattern in the received signaling, the decoding then necessitates detecting the actual configuration for the group, the length of the bit pattern depending on: (i) whether the actual configuration matches a reference configuration of the one or more reference configurations; and/or (ii) which reference configuration the actual configuration matches (box 330).

This may entail, for example, determining the length B of the bit pattern defined for the signaled specific reference configuration and checking whether the B-length string of the following bits in the signaling corresponds to the bit pattern defined for the signaled reference configuration. This determination and check may be performed for each of the one or more reference configurations, after which (if no reference configuration is identified as being signaled) the default length string of the following bits in the signaling is decoded for detecting a non-reference configuration.

Regardless of the specific implementation of the decoding process (blocks 320-330), decoding is repeated for each of the one or more groups of precoders in the codebook (blocks 340, 350).

Those skilled in the art will appreciate that the device-side embodiment includes decoding of any of the network-side embodiments shown with reference to FIG. 3 , including, for example, the “row-like embodiment” and the “column-like embodiment.”

According to one or more other embodiments shown in FIG. 10 , a method is implemented by a wireless communication device 14 (e.g., a UE) for decoding signaling from a network node 10 indicating which precoders in a codebook are restricted for use (e.g., which Kronecker product precoders are restricted). As shown, the method includes receiving signaling from the network node 10 (e.g., a base station) (block 400). The method also includes decoding the signaling to jointly restrict the precoders in each of one or more groups of precoders (block 410). In at least some embodiments, such decoding involves decoding the signaling (i) to be rank-agnostic so as to restrict the precoders regardless of their transmission rank; and/or (ii) to jointly restrict the groups of precoders by restricting a component that the precoders have in common.

Those skilled in the art will appreciate that device-side embodiments encompass the decoding of any of the network-side embodiments described with reference to FIG5 . Thus, for example, device 14 in some embodiments decodes the signaling to jointly restrict the set of precoders that share a particular beam precoder by restricting that beam precoder. Furthermore, one or more device-side embodiments also advantageously utilize the Kronecker structure of the codebook to decode the signaling of FIG10 according to indices k , l , and/or m . In some embodiments, for example, the signaling is being decoded to jointly restrict (e.g., using a single bit) the set of precoders that: (i) have the same value for index k ; (ii) have the same value for index l ; or (iii) have the same pair of values for index ( k , l ).

With the above modifications and variations in mind, FIG11 illustrates additional details of a network node 500 (corresponding to network node 10) according to one or more embodiments. Network node 500 (e.g., via functional components or units 540-570) is configured to implement the process of FIG2 for signaling to wireless communication device 14 which precoders in the codebook are restricted for use. In some embodiments, network node 500, for example, includes a reference configuration identification component or unit 540 for identifying one or more reference configurations for each of one or more groups of precoders. In such cases, network node 500 further includes an actual configuration identification component or unit 550 for identifying an actual configuration for each of the one or more groups. Network node 500 also includes a signaling generation component or unit 560 for generating signaling indicating the actual configuration for each of the one or more groups by generating the signaling as a bit pattern, the length of the bit pattern depending on: (i) whether the actual configuration matches a reference configuration of the one or more reference configurations; and/or (ii) which reference configuration the actual configuration matches. The network node 500 finally comprises sending means or units 570 for notifying the generated signaling to the wireless communication device.

In at least some embodiments, the network node 500 includes one or more processing circuits 510 configured to implement this processing, such as by implementing functional components or units 540-570. In one embodiment, for example, the processing circuits 510 of the node implement the functional components or units 540-570 as corresponding circuits. The circuits in this regard may include one or more microprocessors connected to a memory 520 and/or circuits dedicated to performing a certain functional processing. In embodiments employing memory 520, the memory 520 may include one or more types of memory, such as read-only memory (ROM), random access memory, cache memory, flash memory devices, optical storage devices, etc., that stores program code that, when executed by one or more circuits for implementing one or more microprocessors, implements the techniques described herein.

In one or more embodiments, the network node 500 also includes one or more communication interfaces 530. The one or more communication interfaces 530 include various components (not shown) for sending and receiving data and control signaling. More specifically, the interface 530 includes a transmitter configured to use known signaling processing techniques, typically according to one or more standards, and configured to condition the signal for transmission (e.g., over the air via one or more antennas). Similarly, the interface 530 includes a receiver configured to convert the received signal (e.g., via an antenna) into digital samples for processing by the one or more processing circuits 510.

FIG12 illustrates additional details of a network node 600 according to one or more embodiments. Network node 600 (e.g., via functional components or units 640-650) is configured to implement the process of FIG5 to signal to a wireless communication device which precoders in a codebook are restricted from use. In some embodiments, network node 600 includes, for example, a generating component or unit 640 for generating codebook subset restriction signaling that, for each of one or more groups of precoders, jointly restricts the precoders in the group (e.g., using a single signaling bit). Network node 600 also includes a transmitting component or unit 650 for transmitting the generated signaling to the wireless communication device.

In at least some embodiments, the network node 600 includes one or more processing circuits 610 configured to implement this processing, such as by implementing functional components or units 640-650. In one embodiment, for example, the processing circuits 610 of the node implement the functional components or units 640-650 as corresponding circuits (similar to the circuits described above, for example, connected to the memory 620). In one or more embodiments, the network node 600 also includes one or more communication interfaces 630.

FIG13 illustrates additional details of a wireless communication device 700 (corresponding to wireless communication device 14) according to one or more embodiments. Device 700 (e.g., via functional components or units 740-760) is configured to implement the process of FIG9 for decoding signaling from a network node indicating which precoders in a codebook are restricted for use. In some embodiments, device 700 includes, for example, a receiving component or unit 740 for receiving signaling from the network node. Device 700 further includes an identification component or unit 750 configured to identify, for each of one or more groups of precoders, one or more reference configurations for the group, a bit pattern identified for signaling each reference configuration, and the length of that bit pattern. Finally, device 700 includes a detection component or unit 760 configured to detect the actual configuration signaled for the group by detecting a bit pattern in the received signaling, the length of the bit pattern being dependent on: (i) whether the actual configuration matches one of the one or more reference configurations; and/or (ii) which reference configuration the actual configuration matches.

In at least some embodiments, the device 700 includes one or more processing circuits 710 configured to implement such processing, such as by implementing functional components or units 740-760. In one embodiment, for example, the processing circuits 710 of the device implement the functional components or units 740-760 as corresponding circuits. The circuitry in this regard may include one or more microprocessors connected to a memory 720 and/or circuitry dedicated to performing a certain functional processing. In embodiments employing a memory 720, the memory 720 may include one or more types of memory, such as read-only memory (ROM), random access memory, cache memory, flash memory devices, optical storage devices, etc., that stores program code that, when executed by one or more circuits for implementing one or more microprocessors, implements the techniques described herein.

In one or more embodiments, the device 700 also includes one or more communication interfaces 730. The one or more communication interfaces 730 include various components (not shown) for sending and receiving data and control signaling. More specifically, the interface 730 includes a transmitter configured to use known signaling processing techniques, typically according to one or more standards, and configured to condition the signal for transmission (e.g., over the air via one or more antennas). Similarly, the interface 730 includes a receiver configured to convert the received signal (e.g., via an antenna) into digital samples for processing by the one or more processing circuits 710.

FIG14 illustrates additional details of apparatus 800 according to one or more other embodiments. Apparatus 800 is configured (e.g., via functional components or units 840-850) to implement the process of FIG10 for decoding signaling from a network node indicating which precoders in a codebook are restricted for use. In some embodiments, apparatus 800 includes, for example, a receiving component or unit 840 for receiving the signaling from the network node. Apparatus 800 further includes a decoding component or unit 850 for decoding the signaling into precoders in each of one or more groups of jointly restricted precoders.

In at least some embodiments, the device 800 includes one or more processing circuits 810 configured to implement this processing, such as by implementing functional components or units 840-850. In one embodiment, for example, the processing circuits 810 of the device implement the functional components or units 840-850 as corresponding circuits (similar to the circuits described above, for example, connected to the memory 820). In one or more embodiments, the device 800 also includes one or more communication interfaces 830.

Those skilled in the art will also appreciate that the embodiments herein further encompass corresponding computer programs.

The computer program includes instructions that, when executed on at least one processor of a wireless communication device or network node, cause the node or device to perform any of the corresponding processes described above. Embodiments further include a carrier embodying such a computer program. The carrier may include one of an electronic signal, an optical signal, a radio signal, or a computer-readable storage medium.

The computer program in this regard may comprise one or more code modules corresponding to the components or units described above.

General Examples

In the first embodiment, the UE can receive messages to turn on/off various codebooks. The following messages are valid for the set of possible messages:

At least one of the messages corresponding to a configuration from the 2^N possible configurations is represented by fewer than N bits.

The message will contain information for defining on/off for each individual codebook in the entire codebook.

Each message is uniquely decodable for the UE and will correspond to one of the 2^N possible configurations.

In a second embodiment, the UE of the first embodiment is configured such that codebook subset restriction is made on the beam precoder.

In a third embodiment, the UE of the first embodiment is configured such that the codebook subset restriction is configured with full flexibility for a subset of precoders in the codebook, while the codebook subset restriction is configured with limited flexibility for other precoders in the codebook.

In a fourth embodiment, the UE of the third embodiment is configured such that the set of precoders that the codebook subset restriction is configured for with full flexibility is a set of precoders that can be equivalently expressed as a function of a layer-specific beam precoder.

In the fifth embodiment, the UE of the first embodiment is configured so that N=N_H·N_V comes from the Kronecker structure.

In a sixth embodiment, the UE of any of the first to fifth embodiments is configured so that information for designing a set of messages consists of information on an angular interval that may be restricted.

In a seventh embodiment, the UE of the first embodiment is configured such that only a subset of the 2^N possible configurations can be configured.

In an eighth embodiment, the UE of the first embodiment is configured such that at least one message corresponding to a certain configuration from the 2^N possible configurations is represented by more than N bits.

In a ninth embodiment, the UE of the first embodiment is configured such that a set of messages is designed using information about the likelihood of certain configurations being chosen.

In a tenth embodiment, the UE of the first embodiment is configured such that the information on the likelihood of a certain configuration being chosen is only an implicit assumption of the likelihood.

In the eleventh embodiment, the UE of the first embodiment is configured so that a set of angles specifies the configuration.

Claims (30)

1. A method implemented by a network node (10) for signaling to a wireless communication device (14) which precoders in the codebook are restricted from use, the method being characterized in that: Generate (210) codebook subset restriction signaling, which, for each of one or more groups of precoders, jointly restricts the precoders in that group by restricting a common component shared by the precoders in that group, wherein the precoders are Kronecker product-based precoders, and wherein the signaling is rank-agnostic so as to restrict the precoders regardless of the precoder's transport rank; and The generated signaling is sent (220) from the network node (10) to the wireless communication device (14). 2. The method of claim 1, wherein one of the components includes a beam precoder. 3. The method of claim 2, wherein the beam precoder is based on the Kronecker product of Qh and Mh, where _Qh is an integer horizontal oversampling factor and _Mh is a horizontal dimension, and _Qv is an integer vertical oversampling factor and _Mv is an integer vertical dimension. 4. A method for implementing signaling via a wireless communication device (14) for decoding which precoders in an indicator codebook from a network node (10) are restricted from use, the method being characterized in that: Receive (400) codebook subset restriction signaling, the signaling jointly restricting the precoders in a group for each of one or more groups of precoders by restricting a common component shared by the precoders in that group, wherein the precoders are Kronecker product-based precoders, and wherein the signaling is rank-agnostic so as to restrict the precoders regardless of the precoder's transport rank; and The received signaling is decoded (410) into precoders in each of the one or more groups of precoders that are jointly restricted. 5. The method of claim 4, wherein one of the components comprises a beam precoder. 6. The method of claim 5, wherein the beam precoder is based on the Kronecker product of Qh and Mh, where _Qh is an integer horizontal oversampling factor and _Mh is a horizontal dimension, and _Qv is an integer vertical oversampling factor and _Mv is an integer vertical dimension. 7. The method of any one of claims 5-6, wherein the beam precoder is a Kronecker product of different beamforming vectors associated with different dimensions of a multidimensional antenna array. 8. The method of claim 7, wherein the different beamforming vectors include discrete Fourier transform (DFT) vectors. 9. The method of any one of claims 5-6, wherein the beam precoder is a beamforming vector for transmission on a specific layer of multilayer transmission, wherein different scaled versions of that beamforming vector are transmitted on different polarizations. 10. The method of any one of claims 5-6, wherein the beam pre-encoder is a beamforming vector for transmission at the following layers: Multiple different layers in multi-layer transmission; Multiple different layers of multilayer transmission, wherein said layers are transmitted in orthogonal polarization; or Specific layers and specific polarizations. 11. The method of claim 5, wherein the beam precoder is a Kronecker product of first and second beamforming vectors with first and second indices, wherein the first and second beamforming vectors are associated with different dimensions of a multidimensional antenna array, and wherein the codebook subset restriction signaling jointly restricts precoders in a group of precoders of the same pair having values for the first and second indices. 12. The method of any one of claims 4-6, wherein the wireless device (14) reports channel state information to the network (10) based on received codebook subset restriction signaling. 13. The method of any one of claims 4-6, wherein the codebook subset restriction signaling includes a bitmap, wherein different bits in the bitmap are dedicated to indicating whether different beam precoders are restricted from use. 14. The method of any one of claims 5-6, wherein the precoder is restricted if at least one of the precoders comprising one or more beam precoders is restricted. 15. A network node (10, 600) for signaling to notify a wireless communication device (14, 800) which precoders in the codebook are restricted from use, the network node (10, 600) being configured to: Generate codebook subset restriction signaling, which, for each of one or more groups of precoders, jointly restricts the precoders in that group by restricting a certain component commonly shared by the precoders in that group, wherein the precoders are Kronecker product-based precoders, and wherein the signaling is rank-agnostic so as to restrict the precoders regardless of the precoder's transport rank; and The generated signaling is transmitted from the network node (10, 600) to the wireless communication device (14, 800). 16. The network node of claim 15, wherein one of the components includes a beam precoder. 17. The network node of claim 16, wherein the beam precoder is based on the Kronecker product of Qh and Mh, where _Qh is an integer horizontal oversampling factor and _Mh is a horizontal dimension, and _Qv is an integer vertical oversampling factor and _Mv is an integer vertical dimension. 18. A wireless communication device (14, 800) for decoding signaling from a codebook indicating which precoders are restricted from use, said wireless communication device (14, 800) being configured to: Receive codebook subset restriction signaling, said signaling jointly restricting the precoders in a group for each of one or more groups of precoders by restricting a certain component commonly shared by the precoders in that group, wherein said precoders are Kronecker product-based precoders, and wherein said signaling is rank-agnostic so as to restrict the precoders regardless of the precoder's transport rank; and The received signaling is decoded into precoders in each of the one or more groups of precoders that are jointly restricted. 19. The wireless communication device of claim 18, wherein one of the components includes a beam precoder. 20. The wireless communication apparatus of claim 19, wherein the beam precoder is based on the Kronecker product of Qh and Mh, where _Qh is an integer horizontal oversampling factor and _Mh is a horizontal dimension, and _Qv is an integer vertical oversampling factor and _Mv is an integer vertical dimension. 21. The wireless communication apparatus of any one of claims 19-20, wherein the beam precoder is a Kronecker product of different beamforming vectors associated with different dimensions of a multidimensional antenna array. 22. The wireless communication apparatus of claim 21, wherein the different beamforming vectors include discrete Fourier transform (DFT) vectors. 23. The wireless communication apparatus of any one of claims 19-20, wherein the beam precoder is a beamforming vector for transmission at a specific layer of multilayer transmission, wherein different scaled versions of that beamforming vector are transmitted at different polarizations. 24. The wireless communication apparatus of any one of claims 19-20, wherein the beam pre-encoder is a beamforming vector for transmission at the following layers: Multiple different layers in multi-layer transmission; Multiple different layers of multilayer transmission, wherein said layers are transmitted in orthogonal polarization; or Specific layers and specific polarizations. 25. The wireless communication apparatus of claim 19, wherein the beam precoder is a Kronecker product of first and second beamforming vectors with first and second indices, wherein the first and second beamforming vectors are associated with different dimensions of a multidimensional antenna array, and wherein the codebook subset restriction signaling jointly restricts precoders in a group of precoders that have values for the same pairs of the first and second indices. 26. The wireless communication apparatus of any one of claims 18-20, wherein the wireless apparatus (14) is further configured to report channel state information to the network (10) based on received codebook subset restriction signaling. 27. The wireless communication apparatus of any one of claims 18-20, wherein the codebook subset restriction signaling includes a bitmap, wherein different bits in the bitmap are dedicated to indicating whether different beam precoders are restricted from use. 28. The wireless communication apparatus of any one of claims 19-20, wherein the precoder is restricted if at least one of the precoders comprising one or more beam precoders is restricted. 29. The wireless communication device according to any one of claims 19-20, wherein the wireless communication device is user equipment. 30. A computer-readable storage medium containing stored instructions that, when executed by at least one processor of a node (10, 14), cause the node (10, 14) to perform the method as described in any one of claims 1-14.