TECHNICAL FIELD

The present disclosure relates generally to the field of communication receivers. More particularly, it relates to receiver processing of a received signal comprising respective signal parts from a plurality of users.

BACKGROUND

Traditionally, signal transmission to, or from, multiple user equipments (UEs) in a cellular network (NW) is preferably done by ensuring, or at least attempting to ensure, orthogonality between the transmitted signals. Such an approach may be denoted conventional orthogonal multiple access (COMA). Typical ways to achieve such orthogonality is via allocation of orthogonal resources, such as resources that are orthogonal in one or more of a time domain, a frequency domain, and a spatial domain.

To mitigate imperfections in such allocations and/or to mitigate imperfections introduces by the propagation channel, a communication receiver typically applies signal processing aiming at restoring orthogonality. Examples of such signal processing include equalizing, interference rejection combining (IRC), and minimum mean square error (MMSE) detection. Application of such signal processing may be relevant for orthogonal frequency division multiplex (OFDM) receivers or multiple-input, multiple-output (MIMO) receivers; but also for non-linear variants of such receivers.

An extension of COMA transmission aims to reuse time-frequency (T/F) resources for serving users that are located in spatially non-overlapping regions of the cell coverage area. For example, in uplink multi-user MIMO (UL MU-MIMO) transmission, a multiple-antenna receiver in the network node (e.g., a gNB) may be used to separate signals from multiple users (UEs) sharing the same T/F resources when the users are spatially separated in a physical sense and/or their effective single-input, multiple-output (SIMO) channel vectors are sufficiently uncorrelated. P75821 W02

WO 2020/064999 PCT/EP2019/076160

2

In some scenarios, it may be desirable to be able to handle of an even larger number of users (using the given communication resources) than would be allowed according to the COMA or MU-MIMO approach. One way of achieving this is to apply non-orthogonal multiple access (NOMA). In a typical NOMA scenario, at least two non-spatially separable users share T/F resources. Then, a communication receiver may apply successive interference cancellation (SIC) to mitigate the resulting multi-user interference.

A SIC receiver operates in a sequential manner, removing (one by one) signal parts associated with each user from the received signal as they are decoded. Each removal step requires regeneration, from the decoded signal, of the signal part associated with the user, a subtraction operation, and re-estimations (e.g., of a signal covariance matrix) as the remaining parts of the received signal changes after each removal step.

This incurs a computational load and necessitates a processing flow in the receiver with timing dependencies between the different processing steps. The related receiver timing budget impact and receiver processing block scheduling impact significantly affect the total receiver complexity and design efforts; as well as receiver processing latency.

Therefore, there is a need for alternative receiver processing approaches for reception of a signal comprising respective signal parts from a plurality of users. Preferably, such approaches are suitable for processing of NOMA signals. Also preferably, such approaches alleviate one or more of the complexity and the latency associated with the interference removal processing. Furthermore, such approaches should preferably not compromise receiver performance; or at least achieve receiver performance that is not severely impaired.

SUMMARY

It should be emphasized that the term "comprises/comprising" when used in this specification is taken to specify the presence of stated features, integers, steps, or components, but does not preclude the presence or addition of one or more other features, integers, steps, components, or groups thereof. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. P75821 W02

WO 2020/064999 PCT/EP2019/076160

3

Generally, when an arrangement is referred to herein, it is to be understood as a physical product; e.g., an apparatus. The physical product may comprise one or more parts, such as controlling circuitry in the form of one or more controllers, one or more processors, or the like.

It is an object of some embodiments to solve or mitigate, alleviate, or eliminate at least some of the above or other disadvantages.

According to a first aspect, this is achieved by a method of a communication receiver.

The method may comprise receiving a signal comprising respective signal parts from a plurality of users.

The method may also comprise grouping the plurality of users into two or more groups. The method may also comprise selecting a first one of the two or more groups, and processing the received signal to extract information of the respective signal part of each user of the first group, wherein the processing for the users of the first group is performed in parallel.

The method may also comprise regenerating the respective signal part of each user of the first group based on the corresponding extracted information, and removing the regenerated respective signal part of each user of the first group from the received signal to provide a first reduced received signal.

The method may be a multi-user detection method and/or an interference mitigation method (such as a successive interference cancellation, SIC, method).

The plurality of users may be two or more users, three or more users, four or more users, etc. Generally, the plurality may comprise any suitable minimum number of users.

The received signal may be a non-orthogonal multiple access (NOMA) signal. The respective signal parts may be superimposed in one or more of a time domain, a frequency domain, and a spatial domain.

The grouping step may be performed before or after the reception step. The regeneration of the respective signal part of each user of the first group may also be performed in parallel for the users of the first group. P75821 W02

WO 2020/064999 PCT/EP2019/076160

4

The removal of the regenerated respective signal part of each user of the first group from the received signal may be performed collectively for the users of the first group.

In some embodiments, the method may further comprise selecting a second one of the two or more groups, and processing the first reduced received signal to extract information of the respective signal part of each user of the second group, wherein the processing for the users of the second group is performed in parallel.

The second group may be a different group than the first group.

In some embodiments, the method may further comprise regenerating the respective signal part of each user of the second group based on the corresponding extracted information. In some embodiments, the method may further comprise removing the regenerated respective signal part of each user of the second group from the first reduced received signal to provide a second reduced received signal.

The regeneration of the respective signal part of each user of the second group may also be performed in parallel for the users of the second group. The removal of the regenerated respective signal part of each user of the second group from the received signal may be performed collectively for the users of the second group.

In some embodiments, the method may further comprise repeating the selection and processing steps for some or all groups of the two or more groups.

In some embodiments, the method may further comprise repeating the regeneration and removal steps for some or all groups of the two or more groups (although typically not for the last selected and processed group).

In some embodiments, grouping the plurality of users into two or more groups may comprise letting two users belong to a same group when their mutual interference falls below an intra group interference threshold. The intra-group interference threshold may be static or dynamic. For example, the intra-group interference threshold may be dynamically configured to ensure a minimum group size (e.g., two, three, or any suitable number) of at least one of the groups. P75821 W02

WO 2020/064999 PCT/EP2019/076160

5

In some embodiments, grouping the plurality of users into two or more groups may comprise letting two users belong to a same group when they are associated with signature sequences that entails a relatively low cross correlation in a set of available signature sequences.

A relatively low cross correlation may be defined as a cross correlation that falls below a cross correlation threshold. The cross correlation threshold may be static or dynamic. For example, the cross correlation threshold may be configured to ensure a minimum group size (e.g., two, three, or any suitable number) of at least one of the groups and/or a minimum number of groups (e.g., two, three, or any suitable number).

In some embodiments, grouping the plurality of users into two or more groups may comprise letting two users belong to a same group when they are associated with different spatial transmission resources.

Being associated with different spatial transmission resources may be defined as being spatially separable in any suitable way (e.g., by means of beam-forming reception, MIMO reception, etc.).

In some embodiments, grouping the plurality of users into two or more groups may comprise letting two users belong to a same group when they are associated with different time resources and/or different frequency resources.

Additionally or alternatively, for some embodiments, grouping the plurality of users into two or more groups may comprise keeping a size of each of the two or more groups below or equal to a maximum allowable group size. The maximum allowable group size may, for example, be associated with (hardware, HW, and/or software, SW) constraints limiting the number of users that can be processed in parallel.

In some embodiments, selecting the first one of the two or more groups may comprise selecting the group with highest (among the groups) collective user received signal strength.

In some embodiments, selecting the second one of the two or more groups may comprise selecting the group with highest (among the groups excluding the first group) collective user received signal strength. P75821 W02

WO 2020/064999 PCT/EP2019/076160

6

In some embodiments, selecting a further one of the two or more groups may comprise selecting the group with highest (among the not yet selected groups) collective user received signal strength.

In some embodiments, selecting the first one of the two or more groups may comprise selecting the group with highest (among the groups) received signal strength of the strongest user of the group.

In some embodiments, selecting the second one of the two or more groups may comprise selecting the group with highest (among the groups excluding the first group) received signal strength of the strongest user of the group.

In some embodiments, selecting a further one of the two or more groups may comprise selecting the group with highest (among the not yet selected groups) received signal strength of the strongest user of the group.

In some embodiments, at least some of the plurality of users may share transmission resources (e.g., in one or more of a time domain, a frequency domain, and a spatial domain) in a non- orthogonal multiple access scenario.

In some embodiments, the method may be dynamically activated based on one or more activation criteria, and dynamically de-activated based on one or more de-activation criteria.

In some embodiments, the method may further comprise determining a minimum possible intra-group interference for each of the two or more groups, and de-activating the method when the largest minimum possible intra-group interference among the groups falls above a de-activation threshold.

In some embodiments, a maximum possible intra-group interference may be determined for each of two or more prospect groups, and the method may be activated when the largest maximum possible intra-group interference among the groups falls below an activation threshold.

A second aspect is a computer program product comprising a non-transitory computer readable medium, having thereon a computer program comprising program instructions. The computer P75821 W02

WO 2020/064999 PCT/EP2019/076160

7

program is loadable into a data processing unit and configured to cause execution of the method according to the first aspect when the computer program is run by the data processing unit.

A third aspect is an apparatus for a communication receiver. The apparatus comprises controlling circuitry. The controlling circuitry may be configured to cause reception of a signal comprising respective signal parts from a plurality of users, grouping of the plurality of users into two or more groups, selection of a first one of the two or more groups, and processing - in parallel for the users of the first group - of the received signal to extract information of the respective signal part of each user of the first group. The controlling circuitry may also be configured to cause regeneration of the respective signal part of each user of the first group based on the corresponding extracted information, and removal of the regenerated respective signal part of each user of the first group from the received signal to provide a first reduced received signal.

A fourth aspect is a communication device comprising the apparatus of the third aspect. The communication device may, for example, be any of a wireless communication device, a (wireless) receiver device, and a radio access node (e.g., a network node).

In some embodiments, any of the above aspects may additionally have features identical with or corresponding to any of the various features as explained above for any of the other aspects.

An advantage of some embodiments is that alternative receiver processing approaches are provided for reception of a signal comprising respective signal parts from a plurality of users.

Another advantage of some embodiments is that approaches are suitable for processing of NOMA signals are provided.

Yet an advantage of some embodiments is that alleviation of one or more of the complexity and the latency associated with the interference removal processing is enabled. Yet another advantage of some embodiments is that receiver performance is not compromised (or at least not severely impaired). P75821 W02

WO 2020/064999 PCT/EP2019/076160

8

According to various embodiments, the number of interference cancellation (IC) stages is reduced and/or the timing budget for the receiver is improved. This may, for example, be due to reduction of the number of dependent operations when scheduling availability of receiver functional blocks in a HW-accelerated implementation.

According to various embodiments, the HW cost of the receiver is reduced and/or receiver performance that can be achieved for the same resources is improved.

BRIEF DESCRIPTION OF THE DRAWINGS

Further objects, features and advantages will appear from the following detailed description of embodiments, with reference being made to the accompanying drawings. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the example embodiments.

Figure 1 is a flowchart illustrating example method steps according to some embodiments;

Figure 2 is a schematic block diagram illustrating example functional and/or structural modules for processing of a received signal comprising respective signal parts from a plurality of users;

Figure 3 is a schematic block diagram illustrating example functional and/or structural modules for processing of a received signal comprising respective signal parts from a plurality of users according to some embodiments;

Figure 4 is a schematic drawing illustrating grouping according to some embodiments;

Figure 5 is a schematic block diagram illustrating an example apparatus according to some embodiments; and

Figure 6 is a schematic drawing illustrating an example computer readable medium according to some embodiments.

DETAILED DESCRIPTION

As already mentioned above, it should be emphasized that the term "comprises/comprising" when used in this specification is taken to specify the presence of stated features, integers, steps, or components, but does not preclude the presence or addition of one or more other features, integers, steps, components, or groups thereof. As used herein, the singular forms "a", P75821 W02

WO 2020/064999 PCT/EP2019/076160

9

"an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

Embodiments of the present disclosure will be described and exemplified more fully hereinafter with reference to the accompanying drawings. The solutions disclosed herein can, however, be realized in many different forms and should not be construed as being limited to the embodiments set forth herein.

As mentioned above, it may be desirable to be able to handle of an even larger number of users (using the given communication resources) than would be allowed according to the COMA or MU-MIMO approach. One example when this can be relevant is when the available degrees of freedom (DoF) regarding communication resources and spatial separation are fewer than the number of users to be served.

One way of achieving accommodation of a large number of users is to apply non-orthogonal multiple access (NOMA). In a typical NOMA approach, multiple UEs may be scheduled in the same T/F/S (timing-frequency-spatial) resources, whereby the signals of at least some of the different UEs will not be substantially orthogonal at the receiver. Thus, there will exist residual inter-user interference that needs to be handled by the receiver.

Thus, by the nature of NOMA transmission, multiple signals are received non-orthogonally and - generally - the overlapping (superpositioned) signals may typically need to be separated by the receiver prior to decoding. To assist such separation, UE-specific signature sequences (SSs) may be imposed on the signal of each user, which may facilitate extraction of the individual user signals at the receiver and/or which may enable construction of an effective end-to-end channel which is closer to diagonal (with less cross correlation between users) than without application of the SSs.

In an example of a NOMA approach utilizing SSs, each UE spreads its (e.g., quadrature amplitude modulation, QAM) information symbols using an N -length spreading sequence (signature sequence, SS, or signature vector) { s _k }. K denotes the number of simultaneously active UEs. For a base station (BS, e.g., a gNB) with single antenna, the received signal vector y £€ ^N can be written as P75821 W02

WO 2020/064999 PCT/EP2019/076160

10 where iV is the number of resources (e.g., resource elements, REs) spanned by the signature vectors and carrying the same information symbols, h _fe is the channel vector between the k- th UE and the gNB; x _k is the information symbol of the /c-th UE, w represents noise, and the operator O stands for the pointwise multiplication of two vectors (Hadamard product). For a base station with multiple receive antennas, the received signal corresponding to a single information symbol may be formed by concatenating the iV-length received vector y from each antenna.

From a system performance point-of-view, it may be considered optimal to jointly choose transmit strategies for all UEs and then employ a joint multi-user detector (MUD). Typically, the signature sequences are designed to have certain desired correlation properties, and the construction of the signature sequences {s _fe } (SS) lead to differences between various transmission schemes.

The SS design may be based on different criteria, e.g. low cross-correlation and/or sparsity. In general, when overloading the system with more UEs than can be supported by time-frequency- spatial resources, some residual interference between users will remain as mentioned above. The design of SSs may typically focus on creating sequence sets that minimize that crosstalk between UEs, e.g. Welch bound sequences.

In this disclosure, the focus of various examples will - without being intended as limiting - be on symbol-level spreading schemes and the term "signature sequence" (SS) will be used to refer to NOMA user-specific signatures that differentiate the signals of the users (and are used to separate the signals at the receiver). Other terms may exist, e.g., the more general term "signature" which may be used to include many types of NOMA schemes.

The NW is typically in control of the UE operating mode - scheduling according to COMA, MU- MIMO, and/or NOMA (e.g., using signature sequences for further separation). In the NOMA mode, both the transmitter and the receiver (e.g., the UE and the network node, respectively, in a cellular NW uplink (UL) use case), are typically aware of the relevant signature sequence(s) used and/or of relevant properties of the signature sequence(s) used. For example, the NW may P75821 W02

WO 2020/064999 PCT/EP2019/076160

11

inform each UE about its allocated SS and/or provide other information which is sufficient for the UE to determine a suitable SS.

One receiver structure category that is attractive for NOMA reception based on overall system considerations (e.g., in a new radio - NR - gNB) is the MMSE-SIC receiver. It represents a compromise between non-linear interference mitigation on one hand and incremental complexity and design effort on another hand; compared to the baseline MU-MIMO MMSE receiver. A typical MMSE-SIC receiver is shown in Figure 2 and will be elaborated on later herein.

A SIC receiver operates in a sequential manner, removing (one by one) signal parts associated with each user from the received signal as they are decoded. Each removal step requires regeneration, from the decoded signal, of the signal part associated with the user, a subtraction operation, and re-estimations (e.g., of a signal covariance matrix) as the remaining parts of the received signal changes after each removal step.

This incurs a computational load and necessitates a processing flow in the receiver with timing dependencies between the different processing steps. The related receiver timing budget impact and receiver processing block scheduling impact significantly affect the total receiver complexity and design efforts; as well as receiver processing latency.

Therefore, as already noted above, there is a need for alternative receiver processing approaches for reception of a signal comprising respective signal parts from a plurality of users.

In the following, embodiments will be described whereby such alternative receiver processing approaches are provided.

Figure 1 illustrates an example method 100 for a communication receiver according to some embodiments. The method may be a multi-user detection method and/or an interference mitigation method (such as a successive interference cancellation, SIC, method).

In step 130 a signal is received, which comprises respective signal parts from a plurality of users. The received signal may be a non-orthogonal multiple access (NOMA) signal, wherein the respective signal parts may be superimposed in one or more (typically all) of a time domain, a frequency domain, and a spatial domain. P75821 W02

WO 2020/064999 PCT/EP2019/076160

12

In step 140 the plurality of users are grouped into two or more groups. Grouping may be based on various criteria, but typically users are sorted into the same group when they have no mutual interference (i.e., when they are orthogonal), or when they have a relatively low mutual interference. For example, grouping may comprise letting two users belong to the same group when their mutual interference (e.g., after initial receiver signal processing such as combining and de spreading) falls below an intra-group interference threshold (which may be static or dynamic) and/or letting two users belong to the same group when they are associated with signature sequences that entails a relatively low cross correlation in a set of available signature sequences. Alternatively or additionally, grouping may comprise letting two users belong to a same group when they are associated with different time/frequency/spatial transmission resources.

The grouping may also be subject to conditions such as a maximum allowable group size and/or a minimum allowable group size for one or more of the groups. A maximum allowable group size may, for example, be associated with (hardware, HW, and/or software, SW) constraints limiting the number of users that can be processed in parallel. A minimum allowable group size may, for example, stipulate that at least one group should comprise more than one users; otherwise there would typically not be any difference compared to a conventional SIC approach.

The grouping step may be performed before or after the reception step.

If the grouping is based on cross correlation in a set of available signature sequences, the grouping can be performed at any time as soon as the users have been associated with respective signature sequences. In such examples, the grouping can typically be kept fixed as long as the association of signature sequences to users does not change.

If the grouping is based on user association with spatial resources, the grouping can be performed at any time as soon as it is determined which users are spatially separable. In such examples, the grouping can typically be kept fixed as long as the spatial conditions does not change substantially.

If the grouping is based on measured mutual interference, the grouping can typically be performed after step 130 (or even after the entire SIC processing; then applied to the next P75821 W02

WO 2020/064999 PCT/EP2019/076160

13

received signal) when it is possible to estimate the mutual interference. In such examples, the grouping may be updated at regular intervals in time (e.g., for each execution of step 130).

In step 150, one of the groups is selected and the received signal is processed (e.g., demodulated and decoded) in parallel for each user of the selected group in step 160 to extract information of the respective signal part the users of the selected group. The parallel processing typically entails reduced latency and/or more relaxed timing dependencies between different processing steps. Furthermore, some processing results (e.g., covariance matrix computation) may be shared by the user of the selected group instead of being performed for each user separately.

In step 170, the respective signal part of each user of the selected group is regenerated (e.g., via soft-symbol estimation, re-encoding, and re-modulation) based on the corresponding extracted information, and the regenerated respective signal part of each user of the selected group is removed from the received signal in step 180 to provide a reduced received signal. Typically, the regeneration in step 170 may also be performed in parallel for the users of the selected group and the removal in step 180 may be performed collectively for the users of the selected group.

After the first execution of step 180, the process returns to step 150 where a new (not yet selected/processed) group is selected. The new selected group is processed in step 160 to extract information of the respective signal part the users of the selected group as described above. As long as there are more groups to process (Y-path out of step 190) the respective signal part of each user of the selected group is regenerated in step 170 based on the corresponding extracted information as described above, the regenerated respective signal part of each user of the selected group is removed from the reduced received signal as described above to provide a further reduced received signal, and the process returns to step 150 where a new (not yet selected/processed) group is selected.

When there are no more groups to process (N-path out of step 190) the process returns to step 130 for processing of a new received signal.

The selection in step 150 may be based on various criteria, but typically groups may be selected in an order of signal strength. For example, step 150 may comprise selecting the not yet selected P75821 W02

WO 2020/064999 PCT/EP2019/076160

14

group that has the highest collective user received signal strength among the not yet selected groups, or selecting the not yet selected group whose strongest user signal strength is highest among the not yet selected groups.

One way of achieving signal strength (e.g., power) estimations for the different groups may be based on demodulation reference signals (DMRS). To this end, DMRS may be transmitted in orthogonal ports and/or may be IC-processed ahead of the data signals.

In some embodiments, the communication receiver may be configured for a non-group processing mode (as illustrated by 110; e.g., an conventional - per user - successive interference cancellation mode) and a group processing mode (as illustrated by 120 and comprising the execution of the method steps 130, 140, 150, 160, 170, 180, 190 as described above).

Then , the method of the group processing mode 120 may be dynamically activated based on one or more activation criteria, and dynamically de-activated based on one or more de activation criteria. The activation and de-activation criteria may be any suitable criteria, and are typically related to possible intra-group interference.

Activation is executed from the non-group processing mode and is illustrated by step 115. For example, step 115 may comprise determining a maximum possible intra-group interference for each of two or more prospect groups, and activating the group processing mode when the largest maximum possible intra-group interference among the groups falls below an activation threshold (Y-path out of step 115) and remaining in the non-group processing mode otherwise.

De-activation is executed from the group processing mode and is illustrated by step 125. For example, step 125 may comprise determining a minimum possible intra-group interference for each of two or more groups, and de-activating the group processing mode when the largest minimum possible intra-group interference among the groups falls above a de-activation threshold (Y-path out of step 125) and remaining in the group processing mode otherwise (N- path out of step 125). Even though the de-activation is illustrated as performed after the reception step 130 in the example of Figure 1, it should be noted that the de-activation may be performed at any suitable position of the flow in the group processing mode.

Figure 2 schematically illustrates an example receiver structure 200 for processing of a received signal comprising respective signal parts from a plurality of users. The example receiver P75821 W02

WO 2020/064999 PCT/EP2019/076160

15

structure 200 exemplifies a typical MMSE-SIC receiver that may be used for NOMA reception on top of a MU-MIMO receiver.

Each user (three in this example) is processed separately and in sequence in respective processing stages as illustrated by the dashed boxes 210 (for User 1), 220 (for User 2) and 230 (for User 3). Thus, single-user detection is performed at each processing stage; comprising combining and modulation (213, 223, 233) and decoding (214, 224, 234) in correspondence with a MU-MIMO MMSE receiver.

In all but the last processing stage, the signal of the processed user is regenerated (217, 227) via soft-symbol estimation (215, 225) and re-encoding (216, 226). The regenerated signal is then removed (e.g., subtracted; 218, 228) from the input signal to that stage (either the originally received signal or the received signal reduced by earlier processing stages) to provide a (further) reduced received signal. This is to support SIC for NOMA reception in addition to MU-MIMO MMSE receiver processing.

In correspondence with a MU-MIMO MMSE receiver, a covariance matrix and its inverse (providing weights to compensate for interference and channel imperfections in the combining and modulation step) is calculated at each stage as illustrated by 212, 222, and 232. The calculation is based on the input signal to that stage and on the channel estimates 211, 221, and 231. The MU-MIMO weight computations may be in terms of per-user combining weight processing, which is computationally equivalent to matrix multiplication in conventional multi user or multi-layer notation.

The MMSE-SIC receiver preferably detects users in the order of strongest received signals, or (more generally) in the order of highest decoding margins. Typically, an MMSE-SIC receiver (e.g., at a gNB) may determine that order autonomously (e.g., without requiring information about dynamic modulation and coding scheme, MCS, adjustment or decoding order assumptions from the individual transmitting UEs). The order may be determined based on, e.g., the qualities of the associated demodulation reference signals (DMRS) for each UEs.

Some benefits of iterative receivers (e.g., parallel interference cancellation (1C), belief propagation, etc.) over single-pass receivers (e.g., SIC) lie in the ability to improve performance when multiple users are received with similar power, or (more generally) with similar limited P75821 W02

WO 2020/064999 PCT/EP2019/076160

16

decoding margins at a given stage. When users with sufficient decoding margins can be identified at each stage (similar to, e.g., identifying power differences for grant-free NOMA UEs scheduled with similar MCSs), a SIC receiver such as that illustrated in Figure 2 performs almost as well.

In contrast to iterative approaches, each user is decoded only once which provides for reduced complexity compared to iterative approaches. However, problems with complexity, latency and/or timing remain (especially when the number of users is relatively high) as already mentioned above.

Complexity issues may originate from re-encoding, signal re-generation, and signal removal for decoded users; and from re-computation of the covariance matrix for each processing stage after signal removal. Furthermore, there may be general architecture impact due to signal flow dependencies, shared memory access, etc.

The MMSE-SIC receiver (e.g., that of Figure 2) operates in a sequential manner, removing users one by one as they are decoded. Each removal step requires regenerating the signal of the decoded user, performing a removal (subtraction) operation, and re-estimations (e.g., the signal covariance matrix). This incurs a computational load and necessitates a processing flow in the receiver with timing dependencies between the signals of different users. The related receiver timing budget impact and receiver processing block scheduling impact significantly affect the total receiver complexity and design efforts.

As mentioned before, there is thus a need for a receiver for NOMA signal reception that can alleviate the processing associated with interference removal after decoding each user, without compromising receiver performance.

Figure 3 schematically illustrates an example receiver structure 300 according to some embodiments for processing of a received signal comprising respective signal parts from a plurality of users. The example receiver structure 300 exemplifies a modification of the example receiver structure 200 of Figure 2.

In the modified example receiver structure 300, all users of a group of users (in this example; two groups having two users each) are processed in parallel in respective group processing stages as illustrated by the dashed boxes 310 (for Users 1 and 2) and 330 (for Users 3 and 4). P75821 W02

WO 2020/064999 PCT/EP2019/076160

17

Thus, plural parallel user detection is performed at each processing stage; comprising combining and modulation (313, 323, 333, 343) and decoding (314, 324, 334, 344) for each user in correspondence with a MU-MIMO MMSE receiver.

In all but the last processing stage, the signals of the users of the processed group are regenerated (317, 327) via soft-symbol estimation (315, 325) and re-encoding (316, 326). The regenerated signals are then removed (e.g., subtracted; 328) from the input signal to that stage (either the originally received signal or the received signal reduced by earlier processing stages) to provide a (further) reduced received signal.

A covariance matrix and its inverse (providing weights to compensate for interference and channel imperfections in the combining and modulation step) is calculated at each stage as illustrated by 312 and 332. Thus, complexity is reduced by requiring less computations of covariance matrices and their inverses. The calculation is based on the input signal to that stage and on the channel estimates 311, 331.

Apart from the complexity reductions, processing latency and the number of timing dependencies between signals is reduced compared to the example receiver structure 200 of Figure 2.

Figure 4 schematically illustrates example grouping according to some embodiments. The communication device (CD) 400 receives signals from eight UEs (UE1-UE8) 411, 412, 413, 414, 421, 422, 423, 424. The eight UEs are grouped into two groups 410, 420 wherein UEs of one group are spatially separable from all users of the other group. Within each group, the users are separated using suitable SSs as will be exemplified later herein. Alternatively, grouping may be based on spatial (or other) separation. For example, each group may comprise two UEs (e.g., UE1 and UE5) that are spatially separable. In any case, the receiver of the communication device 400 may process the users of each group in parallel as described in connection with Figures 1 and 3.

Figure 5 schematically illustrates an example apparatus 510 according to some embodiments. Any of the features and examples described above (e.g., in connection with Figures 1 and/or 3) may be equally applicable to the example apparatus 510. For example, the example apparatus P75821 W02

WO 2020/064999 PCT/EP2019/076160

18

510 may comprise the receiver structure 300 of Figure 3 and/or may be configured to cause execution of the method steps of the method 100 of Figure 1.

The example apparatus 510 comprises controlling circuitry (CNTR) 500 configured to cause reception of a signal comprising respective signal parts from a plurality of users. To this end, the example apparatus may comprise or be otherwise associated with (e.g., may be connectable, or connected, to) receiving circuitry (e.g., a receiver; illustrated in Figure 5 as part of a transceiver, TX/RX, 530) configured to receive the signal.

The controlling circuitry (CNTR) 500 is also configured to cause grouping of the plurality of users into two or more groups. To this end, the controlling circuitry may comprise or be otherwise associated with grouping circuitry (GRO; e.g., a grouper) 501 configured to group the plurality of users into two or more groups. The grouping circuitry may be comprised in or otherwise associated with (e.g., connectable, or connected, to) the example apparatus 510.

The controlling circuitry (CNTR) 500 is also configured to cause selection of one of the two or more groups. To this end, the controlling circuitry may comprise or be otherwise associated with selecting circuitry (SEL; e.g., a selector) 502 configured to select one of the two or more groups. The selecting circuitry may be comprised in or otherwise associated with (e.g., connectable, or connected, to) the example apparatus 510.

The controlling circuitry (CNTR) 500 is also configured to cause processing, in parallel for the users of the selected group, of the received signal to extract information of the respective signal part of each user of the selected group. To this end, the controlling circuitry may comprise or be otherwise associated with processing circuitry (PROC; e.g., a processor) 503 configured to process, in parallel for the users of the selected group, the received signal to extract information of the respective signal part of each user of the selected group. The processing circuitry may be comprised in or otherwise associated with (e.g., connectable, or connected, to) the example apparatus 510. The processing circuitry may, for example, comprise one or more of a combiner, a demodulator, and a decoder.

The controlling circuitry (CNTR) 500 is also configured to cause regeneration of the respective signal part of each user of the selected group based on the corresponding extracted information. To this end, the controlling circuitry may comprise or be otherwise associated with P75821 W02

WO 2020/064999 PCT/EP2019/076160

19

regenerating circuitry (REG; e.g., a regenerator) 504 configured to regenerate the respective signal part of each user of the selected group based on the corresponding extracted information. The regenerating circuitry may be comprised in or otherwise associated with (e.g., connectable, or connected, to) the example apparatus 510. The regenerating circuitry may, for example, comprise one or more of a soft symbol estimator, a re-encoder, and a re-modulator.

The controlling circuitry (CNTR) 500 is also configured to cause removal of the regenerated respective signal part of each user of the selected group from the received signal to provide a reduced received signal. To this end, the controlling circuitry may comprise or be otherwise associated with removal circuitry (REM; e.g., a subtractor) 505 configured to remove the regenerated respective signal part of each user of the selected group from the received signal to provide a reduced received signal. The removal circuitry may be comprised in or otherwise associated with (e.g., connectable, or connected, to) the example apparatus 510.

The example apparatus 510 of Figure 5 may, for example, be comprised in a communication device (e.g., any of a wireless communication device, a (wireless) receiver device, and a radio access node (e.g., a network node)).

Thus, some embodiments provide a group-MMSE-SIC receiver for NOMA as well as a method for a NOMA receiver. According to some embodiments, the complexity of NOMA reception may be reduced. Further exemplifications and embodiments will be described in the following.

In typical embodiments, the received signal parts from multiple UEs that are identified as substantially mutually interfere nee -free after initial receiver processing (combining and de spreading) are demodulated and decoded in parallel at each SIC stage. Due to the lack of cross talk between the received signal parts from the UEs in each group, the received signals from the UEs in each group can be decoded substantially interfere nee -free after subtracting the interference from the previous group. Proper design and allocation of signature sequences SS may particularly facilitate this approach. The grouping is then known ahead of time and if the intra-group SSs are orthogonal, interference within the group is fully removed (after de-spreading). P75821 W02

WO 2020/064999 PCT/EP2019/076160

20

It should be noted, however, that principles exemplified herein can be generalized to other NOMA schemes or to MU-MIMO transmission where loading exceeds the available spatial degrees of freedom and where SIC is advantageous.

Furthermore, the approach is explained for an uplink multi-user (UL MU) transmission example, but it should be noted that the principles are equally applicable to downlink (DL) transmission.

Certain types of welch-bound spreading multiple access (WSMA) SS design allow defining groups of sequences that are mutually orthogonal such that non-zero cross-correlation is present only between groups. One such approach is described in the following:

A (potentially large) set of SSs may be provided with non-uniform inter-sequence cross correlation properties, wherein some sequence pairs have low cross-correlation whereas other sequence pairs have high cross-correlation. As a general principle, obtaining relatively low cross correlation for some SS pairs leads, for an unchanged total set size, to increased cross correlation for other pairs.

The approach may include implicitly or explicitly grouping the UEs during SS allocation so that intra-group SS cross-correlation is low. Thereby, the SSs may be used for efficient user separation within each group. In contrast, inter-group SS correlation may be allowed to be high since the SSs will not be primarily relied upon to separate inter-group users. For that, other mechanisms (e.g. spatial approaches) may be used.

This is illustrated in Figure 4 depicting UL NOMA where UE1-UE8 are allocated SSs 1-8 respectively. The groups 410, 420 are separated at the receiver of 400 by applying spatial suppression to UEs from the other group(s) when the users of one group are to be detected. Within each group, it is primarily the SS properties that are used to separate the intra-group users.

In this example, the SSs 1-4 allocated to UE1-UE4 have low cross-correlations, as have SSs 5-8 allocated to UE5-UE5. In contrast any SS pair where the SSs belong to different groups (e.g., SSI and SS5) is allowed to have relatively high cross-correlation. Denoting the nth SS by a column vector S _n and forming an SS matrix S = ... 5 ₈ ], S may have a structure that yields: P75821 W02

WO 2020/064999 PCT/EP2019/076160

21

where, H and L denote relatively "high" and relatively "low" cross-correlation values, respectively. The entries marked with L may represent zero or non-zero magnitudes, and they may or may not differ between different positions. The entries marked with H may represent relatively high magnitudes (at least on average) compared to the -marked entries, and they may or may not differ between different positions. Possibly, the entries marked H may also represent relatively small magnitudes compared to potential cross-correlation magnitudes in general, so that the SSs may provide also some extent of inter-group separation.

Typically, users that are not otherwise separable (time-frequency-spatial) may be allocated SSs that have a corresponding intra-group S ^H S block of the form:

1 L L L

L 1 L L

L L 1 L

L L L 1

In some embodiments, the base set for SS generation is, e.g., a Grassmannian set where the S ^H S has the structure shown above.

Application of some embodiments reduce the complexity of NOMA reception by allowing multiple users to be demodulated and decoded at each SIC stage. SS sequence allocation according to the above scheme further facilitates this. Since the intra-group SS set is orthogonal (or very close to orthogonal), interference within the group is fully (or at least almost fully) removed after de-spreading. Due to the lack of cross-talk between UEs in a group, all UEs in a given group can be decoded in parallel after subtracting the contribution of the previous group. This reduces the number of SIC stages and improves the timing budget for the receiver (e.g., due to the reduction of the number of dependent operations when scheduling the availability of receiver functional blocks in a HW-accelerated implementation).

Some embodiments may be described by the following steps: P75821 W02

WO 2020/064999 PCT/EP2019/076160

22

10: Determine two or more groups of users sharing T/F resources, where intra-group interference after an initial receiver processing stage is below a threshold.

20: Determine (select) a group out of the unprocessed user groups that is strongest according to a predetermined metric.

30: For all users in the selected group, demodulate and decode their signals.

40: If further unprocessed groups remain, for all users in the selected group, regenerate their signal estimates and remove them from the current received signal estimate, to create an updated, reduced-interference received signal estimate.

50: If further unprocessed groups remain, repeat steps 20-40.

In step 10, the receiver determines two or more groups of users with zero or low intra-group interference after initial receiver processing (e.g. combining (e.g. IRC) and de-spreading (with the appropriate SS)).

If the block-orthogonal SS design of described above is used, the groups are already pre-defined according to the SS allocation; users with SSs in an orthogonal SS subset constitute a group.

In other cases, the grouping may be based on, e.g., spatial properties of the different UE signals as will be further exemplified below. The intra-group interference levels (cross-correlation) may be estimated, e.g., based on channel properties derived from the respective DMRS. For example, channel estimation vector (for multiple antenna elements, per subband, or preferably over the entire signal bandwidth) cross-correlation may be used as a metric for determining the corresponding data channel cross-talk levels. UE groups may then be formed using the criterion that the worst-case pairwise cross-talk in a group should lie below a predetermined threshold value. The groups may have equal sizes or the groups may contain unequal numbers of UEs.

If multiple groups remain to be decoded, the group with the best signal quality metric to be processed is determined in step 20. The signal quality metric may be, e.g., the total or average signal power of UEs in the group (where best corresponds to highest), minimum signal power of any UE in the group (where best corresponds to highest), the total or average decoding margin of UEs in the group (where best corresponds to highest), minimum decoding margin of any user in the group (where best corresponds to highest), etc. Decoding margin denotes the P75821 W02

WO 2020/064999 PCT/EP2019/076160

23

difference between the SINR of the received signal and the minimum SINR required to decode a user with a predetermined reliability.

In step 30, all users in the group with the best signal quality metric are subjected to processing steps comprising one or more of de-spreading, combining, demodulation (soft value extraction), and decoding. The processing may be performed independently, in parallel threads, since typically no interaction between the signals of the users is required due to the low (possibly zero) cross-correlation. In some embodiments, joint demodulation may be applied to two or more UEs in the group if any residual cross-correlation should be present.

In step 40, for all UEs in the group, their transmit signal estimates are created if further unprocessed user group(s) remain. Generally, any suitable approaches for signal re-generation may be applied.

For code word (CW) level IC, the decoder output may be used as input to re-encode the transport block. The decoder output may be hard if decoding was successful, or soft otherwise. For symbol level IC, the demodulator output may be used to create soft symbol estimates. In some examples, the transmit signal estimates for resource elements (REs) in the frequency domain are then multiplied by corresponding per-UE channel estimates to generate the estimated received signal contributions from each user. The regenerated signal estimates are then subtracted from the received signal (or the reduced received signal, as applicable). The subtraction may take the form of subtracting multiple individual user signals, or of summing the individual signals to create a group received signal estimate and subtracting the sum.

If the most recent processed group was not the last group, and additional users remain to process (see step 50), the processing flow repeats from step 20 where the next group to be processed is selected based on the updated, interference-cancelled signal estimate at the output of step 40.

An example block diagram of the group-MMSE-SIC receiver where users 1 and 2 belong to the first group and users 3 and 4 to the second is given in Figure 3.

The above examples illustrate various embodiments using the NOMA transmission example, assuming the use of a specific SS design. The principles of the solutions presented herein may, however, be extended to other reception scenarios (e.g. other NOMA schemes or non-NOMA P75821 W02

WO 2020/064999 PCT/EP2019/076160

24

MU-MIMO transmission) where certain groups of users may be assumed (or can be identified/verified) to be received with negligible intra-group interference.

For example, a MMSE-SIC receiver may also be used for MU-MIMO-like reception when the T/F resource reuse exceeds the available spatial degrees of freedom and the interference cannot substantially be removed via linear receiver processing alone. This is the case e.g. when the receiver has less antennas than users sharing the T/F resources. In one embodiment, UEs can be grouped into groups of up to a maximum number (corresponding to the number of antennas) of users each, where the users are sufficiently separated so that a linear receiver fully or substantially suppresses the intra-group interference. UEs in each group may then be processed by the group SIC receiver without intra-group dependencies and removed in one stage when all the UEs in the group have been decoded.

In this more general setting, the groups may be formed e.g. by identifying groups of up to the maximum number users with similar power. Their DMRS may be used to evaluate their effective channels and verify that they are sufficiently spatially separated. The group of users with highest receiver power is decoded first, next power level thereafter, etc.

Some embodiments also provide dynamic activation of the group-MMSE-SIC processing mode. This may be related to the group identification task (the grouping step). In these approaches it may be determined whether to invoke the receiver in single-interferer subtraction mode (conventional, non-group processing) or in the group subtraction mode (group processing). The group MMSE-SIC mode may be activated when it is detected that there are groups with substantially orthogonal users (e.g., in connection with allocation of NOMA SS sets with orthogonal SS groups to UEs).

An alternative to applying the non-group processing mode and the group processing mode is to always apply the group processing mode where a group size equal to one is allowed.

An example of SS computation by the NW will now be given to illustrate one approach how SS vectors may be produced to generate codebooks and/or individual UE SS vectors. The obtained SSs may then be signaled or otherwise distributed to UEs.

Assume there are K single antenna transmitters communicating with a single antenna receiver. Each transmitter has a single symbol to transmit. This symbol modulates a temporal P75821 W02

WO 2020/064999 PCT/EP2019/076160

25

codeword (CW) vector, called a signature sequence (SS), before transmitting the vector over N symbols or time slots, i.e., the symbol is spread (or repeated) over N time slots. It should be noted that this representation is a sampled baseband version of the communication process.

Since all the transmitters access the channel over the same N time (and/or frequency) slots, there is interference among them. This interference arising due to multiple access (MA) is called multiple access interference (MAI). The MA communication may be viewed as a network with N degrees of freedom (DoF) trying to serve K users, each with a required quality of service (QoS).

The design of the SSs may aim at placing each CW at an optimal distance (or angle) from other CWs in the vector space.

When K £ N, there can always be a collision free transmission from all the users, since there can be at least one DoF, which is a time (and/or frequency) slot, for each user for its transmission. This leads to an interference free transmission and such a MA transmission scheme is called orthogonal multiple access (OMA). With OMA there is a performance loss, which is quite visible when each user has a QoS. The system capacity (SC) is also not optimal.

OMA is not possible when the system is overloaded, i.e., when K > N. So the SS vectors should be carefully adjusted to allow controlled interference among the users such that the performance indicators are optimized. Since the SS vectors are no longer orthogonal, the MA scheme is known as Non-Orthogonal Multiple Access (NOMA). The SS for each of the K users should preferably be designed in such a manner that the overall mean squared error (MSE) is minimized. Choosing another performance indicator (PI) such as the signal-to-noise-plus- interference ratio (SINR) or the SC is also a possibility while considering the SS design.

It may be desirable to maximize these two Pis in the NW. Fortunately for an overloaded system, under certain design conditions, optimizing one PI leads to optimizing the other two Pis. To understand this, another PI - called total squared correlation (TSC), which is directly related to the previously mentioned three Pis - may be introduced. The SS is designed by optimizing (or minimizing) the TSC since the NW optimality, in the SC sense and also simultaneously in the overall MSE sense, is defined by the achievable lower bounds (LB) on the TSC in an overloaded system. P75821 W02

WO 2020/064999 PCT/EP2019/076160

26

For the transmitter k, let b _k be the transmitted symbol that modulates a unit norm SS vector s _k . The signal model may be given as y = Sb + z, where z is the zero-mean additive white Gaussian noise vector with a covariance matrix I, the overall SS matrix with a CW in each of its columns is S, the transmit symbol vector is b. The transmit power of each transmitter is set to unity, so the power control problem is not addressed here. A unit norm temporal receive filter f _k , such as a matched filter (MF) or a linear minimum mean squared error (MMSE) filter, may be employed by the receiver to obtain an estimate b _k for the transmitted symbol b _k . The post processed SINR of each user is given as: where trace(-) is the trace operator, v _k is the noise component in the SINR y _fe . The trace(-) term in the denominator is the TSC, which also contains the desired unit signal power. So an additional unity term arises in the denominator. If the post processed noise is white, i.e., the noise power of each v _k is the same, then the TSC can directly be used as a PI.

A LB known as Welch Bound (WB) is defined for the TSC. For overloaded systems it is given as

K

< TSC and for the under loaded systems it is K < TSC. To meet the optimality conditions

N

in the mentioned Pis, the bound should be satisfied by equality. In such a case, the obtained SS is called a Welch Bound Equality (WBE) SS. It should be noted that the SC optimality may not be achieved by the binary WBE SS vectors.

For the construction of the WBE SS, interference avoidance (IA) techniques are known. A nice property of IA methods is that the SS can be obtained iteratively in a sequential and distributed manner. It is guaranteed that the iterations converge, since there exists a fixed-point for S that is an optimum. Though verifying if the obtained optimum is local or global is not easy. By optimum it is meant that the entire matrix S converges as an ensemble (as against each CW convergence), with the considered PI reaching the required tolerance. Hence the converged WBE CWs are not unique. At convergence, all the mentioned Pis are optimized.

From the center part of the SINR equation, let R _fe = I) , which is the correlation matrix of the interference plus noise. It can be readily identified that minimizing the denominator (or equivalently maximizing y _fe ) is a well known Rayleigh-Quotient problem. From P75821 W02

WO 2020/064999 PCT/EP2019/076160

27

this, the Eigen vector corresponding to the minimum Eigen value of R _fe may be considered as CW for user k, if it is assumed that f _k is matched to s _fc . The fixed-point iterations start from the users choosing a random CW. In a given sequential user order, each user updates its SS s _k by solving the Eigen value problem while other SS, S _j ,j ¹ k, are kept fixed. After user k, the next user updates it's CW in the same way by assuming the other CWs to be fixed. The iterations progress up to the final user in the order, such that in each iteration there are K updates, one for each CW in S. After the final update in the given iteration, the first user in the order restarts the updates until convergence.

Again, from the center part of the SINR equation, the solution to f _k can also be identified as the well known Generalized Eigen Value Problem (GEVP), i.e., finding a common Eigen value for the matrix pair (I, R _fe ). The solution to which is the linear MMSE vector given as - _1/2 , in

( ^s ¾ ² ¾) its normalized form. Sequential iterations as mentioned before can be used, except that instead of solving the Eigen value problem, the normalized linear MMSE expression is used during updates. For this SINR maximization problem (or TSC minimization), the obtained solution to S from both the MMSE IA iterations and the Eigen vector IA iterations is the same fixed-point.

A Kronecker product based approach may be employed to obtain (or construct) higher dimensional WBE SS, i.e., higher N values, from lower dimensional WBE SS. If the elements of S are binary antipodal, then WBE set is defined when N is a multiple of 4, i.e., mod(iV, 4) = 0. In cases with mod(iV, 4) ¹ 0, where the WB is loose, a LB known as Karystinos-Pados (KP) bound may be used instead.

Generally, when an arrangement is referred to herein, it is to be understood as a physical product; e.g., an apparatus. The physical product may comprise one or more parts, such as controlling circuitry in the form of one or more controllers, one or more processors, or the like.

The described embodiments and their equivalents may be realized in software or hardware or a combination thereof. The embodiments may be performed by general purpose circuitry. Examples of general purpose circuitry include digital signal processors (DSP), central processing units (CPU), co-processor units, field programmable gate arrays (FPGA) and other programmable hardware. Alternatively or additionally, the embodiments may be performed by specialized circuitry, such as application specific integrated circuits (ASIC). The general purpose P75821 W02

WO 2020/064999 PCT/EP2019/076160

28

circuitry and/or the specialized circuitry may, for example, be associated with or comprised in an apparatus such as a (wireless) receiver device, a (wireless) communication device, or a radio access node (e.g., a network node).

Embodiments may appear within an electronic apparatus (such as a (wireless) receiver device, a (wireless) communication device, or a radio access node) comprising arrangements, circuitry, and/or logic according to any of the embodiments described herein. Alternatively or additionally, an electronic apparatus (such as a (wireless) receiver device, a (wireless) communication device, or a radio access node) may be configured to perform methods according to any of the embodiments described herein.

According to some embodiments, a computer program product comprises a computer readable medium such as, for example a universal serial bus (USB) memory, a plug-in card, an embedded drive or a read only memory (ROM). Figure 6 illustrates an example computer readable medium in the form of a compact disc (CD) ROM 600. The computer readable medium has stored thereon a computer program comprising program instructions. The computer program is loadable into a data processor (PROC) 620, which may, for example, be comprised in a (wireless) receiver device, a (wireless) communication device, or a radio access node 610. When loaded into the data processing unit, the computer program may be stored in a memory (MEM) 630 associated with or comprised in the data-processing unit. According to some embodiments, the computer program may, when loaded into and run by the data processing unit, cause execution of method steps according to, for example, any of the methods illustrated in Figure 1 or otherwise described herein.

Generally, all terms used herein are to be interpreted according to their ordinary meaning in the relevant technical field, unless a different meaning is clearly given and/or is implied from the context in which it is used.

Reference has been made herein to various embodiments. However, a person skilled in the art would recognize numerous variations to the described embodiments that would still fall within the scope of the solution.

For example, the method embodiments described herein discloses example methods through steps being performed in a certain order. However, it is recognized that these sequences of P75821 W02

WO 2020/064999 PCT/EP2019/076160

29

events may take place in another order without departing from the scope of the solution. Furthermore, some method steps may be performed in parallel even though they have been described as being performed in sequence. Thus, the steps of any methods disclosed herein do not have to be performed in the exact order disclosed, unless a step is explicitly described as following or preceding another step and/or where it is implicit that a step follows or precedes another step.

In the same manner, it should be noted that in the description of embodiments, the partition of functional blocks into particular units is by no means intended as limiting. Contrarily, these partitions are merely examples. Functional blocks described herein as one unit may be split into two or more units. Furthermore, functional blocks described herein as being implemented as two or more units may be merged into fewer (e.g. a single) unit.

Any feature of any of the embodiments disclosed herein may be applied to any other embodiment, wherever suitable. Likewise, any advantage of any of the embodiments may apply to any other embodiments, and vice versa. Hence, it should be understood that the details of the described embodiments are merely examples brought forward for illustrative purposes, and that all variations that fall within the scope of the solution are intended to be embraced therein.