TECHNICAL FIELD

Variousexamplesofthe disclosure generallyrelateto re-configurable reflectivede- vices.Variousexamplesspecificallyrelatetotechniquesofconfigu ring re-configura- ble reflectivedevicestosupportmultiple spatialdatastreams.

BACKGROUND

Inorderto increase acoverage areaforwirelesscommunication,itisenvisionedto use re-configurable reflective devices(RRD)to communicatewithwirelesscommuni- cationdevices(UEs).

RRDs arealso sometimesreferredtoasreflecting large intelligentsurface (LIS).

See,e.g.,Sha Hu,FredrikRusek,and Ove Edfors. "Beyond massive MIMO:The po- tentialofdatatransmissionwith large intelligentsurfaces."IEEE Transactionson Sig- nalProcessing 66.10 (2018):2746-2758.The RRD canbe implemented byanarray ofantennasthatreflectincidentelectromagneticwaves/signals.Th e arrayofanten- nascanbe passive.I.e.,the arrayofantennasmaynotbe ableto change amplitude letalone provide amplification;butmayprovide avariable phase shift.An inputspa- tialdirectionfrom which incidentsignalsona datacarrierare accepted and anoutput spatialdirection intowhichthe incidentsignalsare reflected can be re-configured,by changing a phase relationshipbetweenthe antennas.

Ithasbeenobservedthatitcan bedifficultto supportmultiple spatialdata streamsby a single RRD,in particularforpassive antenna arrays.

SUMMARY

Accordingly,a need existsforsupporting multiple data streamsbya single RRD.

Thisneed ismetbythefeaturesofthe independentclaims.Thefeaturesofthede- pendentclaimsdefine embodiments. A method of operating an RRD is provided. One or more wireless communication de- vices (UEs) are served by an access node via the RRD using multiple spatial data streams. The RRD includes multiple reflective elements (REs). The multiple REs are grouped into a plurality of subsets. The multiple REs of each subset are re-configura- ble to provide multiple spatial filters. The method includes obtaining, from the access node and for each subset of the plurality of subsets, a respective first indicator and a respective second indicator; the respective first indicator is indicative of a respective entry of a respective first codebook while the respective second indicator is indicative of a respective entry of a respective second codebook. The method also includes re- configuring, for each subset of the plurality of subsets, the respective multiple REs to provide a respective one of the multiple spatial filters. Said re-configuring is based on the respective entry of the respective first codebook and the respective entry of the respective second codebook.

A computer program or a computer-program product or a computer-readable storage medium including program code is provided. At least one processor can load and ex- ecute the program code. Upon loading and executing the program code the at least one processor performs a method of operating an RRD. One or more wireless com- munication devices (UEs) are served by an access node via the RRD using multiple spatial data streams. The RRD includes multiple reflective elements (REs). The multi- ple REs are grouped into a plurality of subsets. The multiple REs of each subset are re-configurable to provide multiple spatial filters. The method includes obtaining, from the access node and for each subset of the plurality of subsets, a respective first indi- cator and a respective second indicator; the respective first indicator is indicative of a respective entry of a respective first codebook while the respective second indicator is indicative of a respective entry of a respective second codebook. The method also includes reconfiguring, for each subset of the plurality of subsets, the respective mul- tiple REs to provide a respective one of the multiple spatial filters. Said re-configuring is based on the respective entry of the respective first codebook and the respective entry of the respective second codebook.

An RRD is provided, wherein one or more wireless communication devices can be served by an access node via the RRD using multiple spatial data streams. The RRD includes multiple reflective elements, the multiple reflective elements being grouped into a plurality of subsets, the multiple reflective elements of each subset being re- configurable to provide multiple spatial filters. The RRD includes a control circuitry configured to obtain, from the access node and for each subset of the plurality of subsets, a respective first indicator indicative of a respective entry of a respective first codebook and a respective second indicator indicative of a respective entry of a re- spective second codebook. The control circuitry is further configured to re-configure, for each subset of the plurality of subsets, the respective multiple reflective elements to provide a respective one of the multiple spatial filters, based on the respective en- try of the respective first codebook and the respective entry of the respective second codebook.

A method of operating an access node is provided. The access node serves one or more wireless communication devices via an RRD, using multiple spatial data streams. The RRD comprises multiple reflective elements, the multiple reflective ele- ments being grouped into a plurality of subsets, the multiple reflective elements of each subset being re-configurable to provide multiple spatial filters. The method in- cludes providing, to the RRD and for each subset of the plurality of subsets, a re- spective first indicator indicative of a respective entry of a respective first codebook and a respective second indicator indicative of a respective entry of a respective sec- ond codebook.

A computer program or a computer-program product or a computer-readable storage medium including program code is provided. At least one processor can load and ex- ecute the program code. Upon loading and executing the program code the at least one processor performs a method of operating an access node. The access node serves one or more wireless communication devices via an RRD, using multiple spa- tial data streams. The RRD comprises multiple reflective elements, the multiple re- flective elements being grouped into a plurality of subsets, the multiple reflective ele- ments of each subset being re-configurable to provide multiple spatial filters. The method includes providing, to the RRD and for each subset of the plurality of sub- sets, a respective first indicator indicative of a respective entry of a respective first codebook and a respective second indicator indicative of a respective entry of a re- spective second codebook.

An access node is provided. The access node is configured to serve one or more wireless communication devices via an RRD using multiple spatial data streams. The RRD comprises multiple reflective elements, the multiple reflective elements being grouped into a plurality of subsets, the multiple reflective elements of each subset be- ing re-configurable to provide multiple spatial filters. The access node comprises a control circuitry configured to provide, to the RRD and for each subset of the plurality of subsets, a respective first indicator indicative of a respective entry of a respective first codebook and a respective second indicator indicative of a respective entry of a respective second codebook.

An RRD includes multiple reflective elements being grouped into a plurality of sub- sets. For each subset of the plurality of subsets, the reflective elements the respec- tive subset are arranged on a respective circle.

A method of operating an RRD device is provided. One or more UEs are being served by an access node via the RRD using multiple data streams. The RRD in- cludes multiple reflective elements. The multiple reflective elements are being grouped into a plurality of subsets. The multiple reflective elements of each subset are being reconfigurable to provide multiple spatial filters. The method includes se- lecting, for each subset of the plurality of subsets, a respective entry of a respective first codebook and a respective entry of a second codebook. At least the entries of the second codebooks are selected based on second indicators indicative of these entries of the second codebooks and obtained from the access node. Optionally, also the entries of the first codebooks can be selected based on first indicators indicative of these entries of the first codebooks obtained from the access node; otherwise the entries of the first codebooks can be selected based on a positioning procedure be- tween the RRD and the one or more UEs. Then, it is possible to reconfigure, for each subset of the plurality of subsets, the respective multiple reflective elements to pro- vide a respective one of multiple spatial filters, based on the respective selected en- try of the respective first codebook and the respective selected entry of the respec- tive second codebook.

A computer program or a computer-program product or a computer-readable storage medium including program code is provided. At least one processor can load and ex- ecute the program code. Upon loading and executing the program code the at least one processor performs a method of operating an RRD. One or more UEs are being served by an access node via the RRD using multiple data streams. The RRD in- cludes multiple reflective elements. The multiple reflective elements are being grouped into a plurality of subsets. The multiple reflective elements of each subset are being reconfigurable to provide multiple spatial filters. The method includes se- lecting, for each subset of the plurality of subsets, a respective entry of a respective first codebook and a respective entry of a second codebook. At least the entries of the second codebooks are selected based on second indicators indicative of these entries of the second codebooks and obtained from the access node. Optionally, also the entries of the first codebooks can be selected based on first indicators indicative of these entries of the first codebooks obtained from the access node; otherwise the entries of the first codebooks can be selected based on a positioning procedure be- tween the RRD and the one or more UEs. Then, it is possible to reconfigure, for each subset of the plurality of subsets, the respective multiple reflective elements to pro- vide a respective one of multiple spatial filters, based on the respective selected en- try of the respective first codebook and the respective selected entry of the respec- tive second codebook.

An RRD is provided. One or more UEs are being served by an access node via the RRD using multiple data streams. The RRD includes multiple reflective elements. The multiple reflective elements are being grouped into a plurality of subsets. The multiple reflective elements of each subset are being reconfigurable to provide multi- ple spatial filters. The RRD comprises a control circuitry configured to select, for each subset of the plurality of subsets, a respective entry of a respective first codebook and a respective entry of a second codebook. At least the entries of the second codebooks are selected based on second indicators indicative of these entries of the second codebooks and obtained from the access node. Optionally, also the entries of the first codebooks can be selected based on first indicators indicative of these en- tries of the first codebooks obtained from the access node; otherwise the entries of the first codebooks can be selected based on a positioning procedure between the RRD and the one or more UEs. Then, it is possible that the control circuitry is config- ured to reconfigure, for each subset of the plurality of subsets, the respective multiple reflective elements to provide a respective one of multiple spatial filters, based on the respective selected entry of the respective first codebook and the respective selected entry of the respective second codebook. It is to be understood that the features mentioned above and those yet to be ex- plained below may be used not only in the respective combinations indicated, but also in other combinations or in isolation without departing from the scope of the in- vention.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 schematically illustrates a communication system including a base station and a UE according to various examples.

FIG. 2 schematically illustrates details of the communication system according to FIG. 2.

FIG. 3 schematically illustrates details of a communication system including a base station, two UEs, as well as an RRD according to various examples.

FIG. 4 schematically illustrates details with respect to the RRD according to various examples.

FIG. 5 is a flowchart of a method according to various examples.

FIG. 6 is a flowchart of a method according to various examples.

FIG. 7 schematically illustrates a system model of a communication system including a base station, multiple UEs, as well as an RRD according to various examples.

FIG. 8 schematically illustrates operation of a subset of reflective elements of the RRD according to various examples.

FIG. 9 schematically illustrates multiple spatial data streams and multiple beams used for communicating data between a base station and the UE via the RRD ac- cording to various examples. FIG. 10 schematically illustrates an example structural implementation of reflective elements of the RRD as well as antenna elements of the base station according to various examples.

FIG. 11 is a flowchart of a method according to various examples.

FIG. 12 is a flowchart of a method according to various examples.

DETAILED DESCRIPTION OF THE DRAWINGS

Some examples of the present disclosure generally provide for a plurality of circuits or other electrical devices. All references to the circuits and other electrical devices and the functionality provided by each are not intended to be limited to encompass- ing only what is illustrated and described herein. While particular labels may be as- signed to the various circuits or other electrical devices disclosed, such labels are not intended to limit the scope of operation for the circuits and the other electrical de- vices. Such circuits and other electrical devices may be combined with each other and/or separated in any manner based on the particular type of electrical implemen- tation that is desired. It is recognized that any circuit or other electrical device dis- closed herein may include any number of microcontrollers, a graphics processor unit (GPU), integrated circuits, memory devices (e.g., FLASH, random access memory (RAM), read only memory (ROM), electrically programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM), or other suitable variants thereof), and software which co-act with one another to per- form operation(s) disclosed herein. In addition, any one or more of the electrical de- vices may be configured to execute a program code that is embodied in a non-transi- tory computer readable medium programmed to perform any number of the functions as disclosed.

In the following, examples of the disclosure will be described in detail with reference to the accompanying drawings. It is to be understood that the following description of examples is not to be taken in a limiting sense. The scope of the disclosure is not in- tended to be limited by the examples described hereinafter or by the drawings, which are taken to be illustrative only. The drawings are to be regarded as being schematic representations and elements illustrated in the drawings are not necessarily shown to scale. Rather, the various el- ements are represented such that their function and general purpose become appar- ent to a person skilled in the art. Any connection or coupling between functional blocks, devices, components, or other physical or functional units shown in the draw- ings or described herein may also be implemented by an indirect connection or cou- pling. A coupling between components may also be established over a wireless con- nection. Functional blocks may be implemented in hardware, firmware, software, or a combination thereof.

Techniques are described that facilitate wireless communication between nodes. A wireless communication system includes a transmitter node and one or more re- ceiver nodes. The nodes communicate on a data carrier. In some examples, the wireless communication system can be implemented by a wireless communication network, e.g., a radio-access network (RAN) of a Third Generation Partnership Pro- ject (3GPP)-specified cellular network (NW). In such case, the transmitter node can be implemented by an access node such as a base station (BS) of the RAN, and the one or more receiver nodes can be implemented by wireless communication devices (also referred to as user equipment, UE). Hereinafter, for sake of simplicity, various examples will be described with respect to an example implementation of the trans- mitter node by a BS and the one or more receiver nodes by UEs - i.e. , to downlink (DL) communication; but the respective techniques can be applied to other scenar- ios, e.g., uplink (UL) communication and/or sidelink communication.

According to various examples, it is possible to use multi-antenna techniques. Multi- antenna techniques are used to enhance reliability and/or throughput of wireless communication. The BS uses multiple antennas. Thereby, a signal can be transmit- ted redundantly (diversity multi-antenna mode) along multiple spatial data streams or multiple signals can be transmitted on multiple spatial data streams (spatial multiplex- ing multi-antenna operational mode). The multiple independent data streams when using spatial multiplexing multi-antenna operational mode are sometimes also re- ferred to as layers.

It is possible also to use beamforming: here, each spatial data stream can be defined by focusing the transmission energy for transmitting (transmit beam, TX beam) and/or the receive sensitivity for receiving (receive beam, RX beam) to a particular spatial direction. For beamforming, the process of identifying the appropriate beams is often referred to beam management.

Various techniques are related to using a spatial multiplexing multi-antenna opera- tional mode including multiple independent spatial data streams that may include multiple signals, i.e. , diversity multi-antenna mode.

According to various examples, an RRD is used to serve one or more UEs. The mul- tiple spatial data streams can be implemented between the BS and the RRD. Each UE of the one or more UEs may be served by one or more of the multiple spatial data streams. Where a given UE is served by two or more spatial data streams, the data rate can be increased (multi-antenna gain).

As a general rule, various options are available for implementing the RRD. The RRD may include an antenna array. Here, antennas form re-configurable reflective ele- ments. The RRD may include a meta-material surface. Here, sub-wavelength ele- ments form the re-configurable reflective elements.

According to various examples described herein, the RRD may not be capable to be re-configured to adjust an amplitude of the reflected signal, e.g.., the RRD may not implement an amplify-and-forward functionality; but be restricted to imposing a varia- ble phase shift. The RRD may include a passive array of reflective elements, not able to amplify or generally change amplitude of the electromagnetic waves.

To forward an incident signal, the RRD may not decode the signal. The RRD may not translate an incident signal into the baseband. As a general rule, the RRD is config- ured to employ multi-antenna techniques. In particular, the RRD is reconfigurable to provide multiple spatial filters. Thereby, a spatial data stream between two nodes - e.g., the BS and the UE - can be diverted. Each one of the multiple spatial filters is associated with a respective input spatial direction from which incident signals on a respective data radio carrier are accepted, as well as with a respective output spatial direction into which incident signals are reflected or amplified by the RRD. Each out- put spatial direction is associated with a respective beam. Each spatial filter is achieved by imposing a certain phase shift onto the incident electromagnetic waves, at each respective re-configurable element. The RRD thereby implements beamform- ing.

There are many schools of thought for how RRDs should be integrated into 3GPP- standardized RANs. In an exemplary case, the NW operator has deployed the RRDs and is therefore in full control of the RRD operations. The UEs, on the other hand, may not be aware of the presence of any RRD, at least initially, i.e. , it is transparent to a UE whether it communicates directly with the BS or via an RRD. The RRD es- sentially functions as a coverage-extender of the BS. The BS may have established a control channel with the RRD. According to another exemplary case, it might be a pri- vate user or some public entity that deploys the RRD. Further, it may be that the UE, in this case, controls RRD operations. The BS, on the other hand, may not be aware of the presence of any RRD and, moreover, may not have control over it/them what- soever. The UE may gain awareness of the presence of RRD by means of some short-range radio technology, such as Bluetooth, wherein Bluetooth may refer to a standard according to IEEE 802.15, or WiFi, wherein WiFi may refer to a standard according to IEEE 802.11 , by virtue of which it may establish the control channel with the RRD. The control channel can thus be on an auxiliary link.

In a further exemplary case, neither the UE nor the BS are aware of the presence of the RRD. The RRD may be transparent with respect to a communication between the UE and the BS on a data carrier. The RRD may gain awareness of the UE and/or the BS and re-configure itself based on information obtained from the UE and/or BS.

The three exemplary cases described above are summarized in TAB. 1 below.

TAB. 1 : Scenarios for RRD integration into cellular NW.

According to various examples disclosed herein, it is possible to implement a sce- nario A according to TAB. 1 ; some scenarios may also employ a LIE-RRD control channel according to scenario B. Based on the BS-RRD control channel, it is possi- ble to perform a channel sounding procedure. The BS and the RRD can participate in the channel sounding procedure. Here, reference signals may be transmitted on the control channel, e.g., from the BS to the UE or vice versa. Thereby, the channel be- tween the RRD and the BS can be sounded, i.e., a channel matrix H can be deter- mined. The particular implementation of such channel sounding procedure is not ger- mane for the techniques described herein. Rather, in this regard, it is possible to use reference techniques generally known to the person skilled in the art. According to various examples, it is possible to use multiple data streams. These multiple data streams can be used to serve one or more UEs. A typical example would be to serve multiple UEs using the multiple data streams.

Next, in TAB. 2, various options are described how it is possible to separate multiple data streams at the RRD.

TAB. 2: Various reference implementations for separating multiple spatial data streams at the RRD. Challenges and limitations of these reference implementations have been explained above.

In view of these challenges and limitations of the reference implementations for sepa- rating multiple spatial data streams at the RRD, the inventors have conceived an- other concept of forwarding multiple spatial data streams via an RRD. Such concept relies on a modified SDMA approach, further developing the SDMA approach of TAB. 2, example V. According to various examples described herein, it is possible to sepa- rate multiple spatial data streams by partitioning the RRD. This means that the re- configurable reflective elements (REs) are grouped into subsets.

Such partitioning - according to the modified SDMA approach - can reduce the array gain if compared to scenario V of TAB. 2, i.e. , if compared to the general per-RE SDMA approach. Such techniques have various advantages when compared to the reference implementations according to TAB. 2. Firstly, a larger number of data streams can be supported without reducing the spectral efficiency of each data stream (i.e., no time or bandwidth sharing); this is beneficial when compared to time or frequency or code division multiplexing (cf. TAB 2, options I and II). Secondly, a codebook-based approach using codebooks of limited size can be used, which is an advantage if compared to, e.g., TAB. 2: option IV. Thirdly, it is possible to determine the codebooks with high accuracy and limited computational efforts, e.g., using an optimization of limited complexity; this can be an advantage when compared to op- tion IV.. Fourthly, it is possible - at least in some scenarios - to maintain UE trans- parency, i.e. , the UE does not need to be aware of it being served via the RRD, cf. TAB. 1 , scenario I. Fifthly, the number of data streams supported is not limited to two; this is beneficial when compared to polarization division multiplexing (cf. TAB 2, op- tion III). It would even be possible to integrate polarization multiplexing in the tech- niques described herein.

Then, to forward multiple spatial data streams using a codebook-based re-configura- tion of the RRD in combination with partitioning of the RRD, the following pre-condi- tions are assumed to be met: firstly, the BS is aware of the beam provides a strong- est signal from the RRD to the UEs; secondly, the BS is aware of the channel be- tween the BS and the RRD; and thirdly, there is a dedicated control channel between the BS and the RRD in which the BS can control the RRD (cf. TAB. 1 , example I). The specific implementation of respective processes to obtain such information are out of scope of the subject disclosure. Very briefly, some examples will be given: for instance, the UE and the RRD could perform a relative positioning on a respective control channel, e.g., using a short-range communication protocol, and respective relative positions may be reported to the BS. It would also be possible that, both, the UE, as well as the RRD report their respective locations in a global coordinate sys- tem so that the BS can conclude on the relative positioning. It would be possible that RRD-beam identification toward the UE is embedded in the communication based on pilot signals - i.e., reference signals can be indicative of the RRD beam identity of the particular beam on which they are sent; this is similar to the BS-beam identifica- tion toward a UE in a non-RRD case. Furthermore, by communicating reference sig- nals on a control channel between the BS and the RRD, it is possible to determine the channel between the BS and the RRD, more specifically it is possible to deter- mine the channels between each one of the subsets of the RRD and the BS. For ex- ample, path loss and angle of arrival measurements could be performed. Still further, the control channel between the BS and the RRD can be established using a short- range communication protocols such as Bluetooth or Wi-Fi.

According to various examples, the RRD is partitioned, i.e., the REs of the RRD are grouped into a plurality of subsets. Each subset can provide multiple spatial filters, depending on the configuration of the respective REs of that subset. According to these techniques, each subset can select I isolate a specific spatial data stream. Non-selected data streams can be scattered or even nullified in certain direc- tions, at the respective subset.

This means that different subsets can apply different spatial filters. Different subsets can reflect the incident electromagnetic waves into different directions. Thereby, spa- tial data streams can be flexibly forwarded into different directions in the surrounding of the RRD.

According to the techniques described herein, the interference between the different spatial data streams can be low. A small codebook can be sufficient. Each subset can have its own codebook or codebooks.

According to examples, two codebooks are used per subset. A combination of the two codebooks then defines the spatial filter to be applied at the respective subset. The BS can then provide to the RRD, for each subset, a respective first indicator and a respective second indicator. The respective first indicators indicative of a respective entry of the respective first codebook of the subset; while the respective second indi- cators indicative of a respective entry of a respective second codebook of that same subset. Then, it is possible to reconfigure, for each subset, the respective multiple re- flective elements to provide a respective spatial filter, based on the respective entries of the respective first codebook and the respective second codebook.

In other words, the BS indicates, to each subset, two codebook entries of two code- books to apply. The two codebooks and their function are summarized in TAB. 3 be- low.

TAB. 3: A first codebook - hereinafter referred to as steering vector codebook - and a second codebook - hereinafter referred to as compensation codebook - are used per subset, to select and implement, for each subset of REs of the RRD, a respective spatial filter. Respective codebook indices can be provided by the BS to the RRD, e.g., using a control channel (cf. TAB. 1 , example I).

At the BS, a specific precoding can be used that correlates with the determined com- pensation codebooks. The precoding depends on the channel between the BS and the RRD.

FIG. 1 schematically illustrates a communication system 100. The communication system includes two nodes 101 , 102 that are configured to communicate with each other via a data carrier 111. In the example of FIG. 1 , the node 101 is implemented by an access node, more specifically a BS, and the node 102 is implemented by a UE. The BS 101 can be part of a cellular NW (not shown in FIG.1 ). As a general rule, the techniques described herein could be used for various types of communication systems, e.g., also for peer-to-peer communication, etc. For the sake of simplicity, however, hereinafter, various techniques will be described in the context of a communication system that is implemented by a BS 101 of a cellular NW and a UE 102.

As illustrated in FIG. 1 , there can be DL communication, as well as UL communica- tion. Various examples described herein particularly focus on the DL communication from the BS 101 to the UE 102. However, similar techniques may be applied to UL communication from the UE 102 to the BS 101.

The UE 102 and the BS 101 can communicate on the data carrier 111. For instance, the data carrier 111 may have a carrier frequency of not less than 20 GHz or even not less than 40 GHz. The data carrier 111 may be via an RRD (not illustrated in FIG. 1 ).

FIG. 2 illustrates details with respect to the BS 101. The BS 101 implements an ac- cess node to a communications network, e.g., a 3GPP-specified cellular network. The BS 101 includes control circuitry that is implemented by a processor 1011 and a non-volatile memory 1015. The processor 1011 can load program code that is stored in the memory 1015. The processor 1011 can then execute the program code. Exe- cuting the program code causes the processor to perform techniques as described herein, e.g.: providing a codebook index to a RRD; participating in a channel sound- ing procedure of one or more channels between the BS 101 and the RRD; transmit- ting and/or receiving (communicating) data on the data carrier 111 ; performing an op- timization to determine a codebook for use at the RRD; providing the codebook to the RRD; etc.

FIG. 2 also illustrates details with respect to the UE 102. The UE 102 includes control circuitry that is implemented by a processor 1021 and a non-volatile memory 1025. The processor 1021 can load program code that is stored in the memory 1025. The processor can execute the program code. Executing the program code causes the processor to perform techniques as described herein, e.g.: transmitting and/or receiv- ing (communicating) payload data on the data carrier 111 , e.g., via an RRD (not shown in FIG. 2); participating in a positioning procedure for determining a relative position of the UE 102 with respect to an RRD, e.g., in the case of line-of-sight signal propagation (which is only one option considered in this disclosure); etc.

FIG. 2 also illustrates details with respect to communication between the BS 101 and the UE 102 on the data carrier 111. The BS 101 includes an interface 1012 that can access and control multiple antennas 1014. Likewise, the UE 102 includes an inter- face 1022 that can access and control multiple antennas 1024.

While the scenario of FIG. 2 illustrates the antennas 1014 being coupled to the BS

101 , as a general rule, it would be possible to employ transmit-receive points (TRPs) that are spaced apart from the BS.

The interfaces 1012, 1022 can each include one or more TX chains and one or more receiver chains. For instance, such RX chains can include low noise amplifiers, ana- logue to digital converters, mixers, etc. Analogue and/or digital beamforming would be possible. Thereby, phase-coherent transmitting and/or receiving (communicating) can be implemented across the multiple antennas 1014, 1024. Multi-antenna tech- niques can be implemented.

By using a TX beam, the direction of signals transmitted by a transmitter of the com- munication system is controlled. Energy is focused into a respective direction or even multiple directions, by phase-coherent superposition of the individual signals originat- ing from each antenna 1014, 1024. Thereby, a spatial data stream can be directed. The spatial data streams transmitted on multiple beams can be independent, result- ing in spatial multiplexing multi-antenna transmission; or dependent on each other, e.g., redundant, resulting in diversity multi-input multi-output (MIMO) transmission.

As a general rule, alternatively or additionally to such TX beams, it is possible to em- ploy receive (RX) beams.

FIG. 3 illustrates aspects with respect to communicating via an RRD 109. Two UEs

102, 103 are served by the BS 101 via an RRD 109. Different steering vector define beams 671 -672 that are directed to the UE 102 and the UE 103, respectively. Each UE 102, 103 can be served by the BS 101 using one or more respective spatial data streams between the BS 101 and the RRD 109 (the spatial data streams are not il- lustrated in FIG. 3 for the sake of simplicity). FIG. 4 illustrates aspects in connection with the RRD 109. The RRD 109 includes an array of REs 1094 - e.g., antennas or meta-material unit cells - each imposing a re- spective configurable phase shift when reflecting incident electromagnetic waves. The array of REs 1094 in the illustrated example is passive; i.e. , the REs 1094 may not be able to change the amplitude of the electromagnetic waves. This array of REs 1094 forms a reflective surface 611 . Typically, antennas can impose gradually vary- ing phase shifts; while meta-material unit cells may, in some examples, be configured to provide only a set of phase shifts with a higher degree of quantization if compared to antennas, e.g., a either one of two phase shifts, e.g., such as +90° or -90° or 0° and +180° (“1 -bit setting”).

Each RE 1094 can locally provide a respective phase shift, i.e., each RE 1094 may be individually configured.

The re-configuration of REs 1094 defines respective spatial filters that are associated with spatial directions into which incoming electromagnetic waves are reflected, i.e., on a macroscopic level. This defines the spatial direction into which an outgoing beam 671 -672 (cf. FIG. 3) is reflected.

As illustrated in FIG. 4, the REs 1094 are grouped into subsets 691 , 692. This group- ing is also referred to as partitioning the REs 1094 of the RRD 109. This grouping can be a logical grouping; i.e., may not be reflected in any hardware implementation. For example, different groupings are possible using the same hardware. It would even be possible that some REs 1094 belong to multiple subsets, i.e., that the sub- sets overlap. The grouping may be defined with respect to codebooks that are pro- vided. Each subset 691-692 can provide a respective spatial filter. Different subsets 691 -692 can provide the same or different spatial filters. The REs 1094 of each sub- set 691 -692 can be coherently addressed to implement a respective spatial filter.

While in the scenario of FIG. 4 each subset 691 , 692 includes a count of nine REs 1094, as a general rule, the count of REs per subset 691 , 692 can be larger or smaller.

While FIG. 4 illustrates two subsets 691 , 692, as a general rule, more than two sub- sets can be defined by the respective partition of the array of REs 1094. For example, the count of subsets may be predefined - e.g., in a calibration phase - depending on a maximum count of spatial data streams to be supported by the RRD 109. It would also be possible that the count of subsets, or generally the partitioning, is dynamically changed when communicating payload data.

The RRD 109 includes an antenna interface 1095 and a processor 1091 that can ac- tivate respective spatial filters one after another, e.g., in accordance with a re-config- uration timing that defines the dwell time per spatial filter.

Further, there is a communication interface 1092 such that communication on an auxiliary link 199 can be established between the RRD 109 and, e.g., the BS 101 and/or the UE 102. Example implementations of the auxiliary link 199 include, e.g., a Wi-Fi protocol or a Bluetooth protocol. Out-of-band signaling may be used with re- spect to the communication of payload data via the REs. The auxiliary link 199 could also be implemented using the same communication protocol used for communica- tion payload data; e.g., a separate bandwidth part may be used for the auxiliary link 199. In-band signaling may be used. Different bandwidth parts may be used for the communication of payload data via the REs and the auxiliary link. A wired connection would be possible. A control channel with, e.g., the BS 101 and/or the UE 102 can be established on the auxiliary link 199. The auxiliary link 199 can be used to assist the beam management procedure at the RRD 109. It would be possible to implement a channel sounding procedure and/or a positioning procedure.

The processor 1091 can load program code from a non-volatile memory 1093 and execute the program code. Executing the program code causes the processor to per- form techniques as described herein, e.g.: re-configuring each one of the REs 1094 to provide a selected one of multiple spatial filters, e.g., on subset-level; determining a configuration of each one of the REs 1094; selecting a spatial filter by accessing one or more codebooks, e.g., the codebooks according to TAB. 3; obtaining indices indicating entries of the one or more codebooks; applying a partition into the subsets 691 -692, participating in a channel sounding procedure with the BS 101 ; participating in a positioning procedure with one or more UEs 102, 103; etc.

FIG. 4 illustrates - as an example - a cartesian array of REs 1094. This is only one option. In other examples, the REs can be arranged in concentric circles (as will be explained in further detail in connection with FIG. 10). Also, the BS 101 can have a circular antenna array, wherein the center axis of the BS and of the RRD 109 are aligned. Also, an angular offset would be possible.

Next, operation of the communication between the BS 101 and the UEs 101-102 us- ing the RRD 109 will be explained.

FIG. 5 is a flowchart illustrating an example method. FIG. 5 is a flowchart illustrating two phases of the operation. The BS 101 and the RRD 109 can operate differently during these two phases.

Box 6005 denotes a calibration phase. The calibration phase of box 6005 precedes a communication phase at box 6010 where payload data is communicated via the RRD 109. For example, the calibration phase of box 6005 can be part of an installation/de- ployment of the RRD 109 and optionally the BS 101 at a given site, e.g. , a VR gam- ing environment.

Details with respect to the calibration phase of box 6005, as well as the communica- tion phase of box 6010 are discussed below in connection with FIG. 6.

FIG. 6 is a flowchart of a method according to various examples. The method of FIG. 6 illustrates details of the method of FIG. 5 according to an example implementation.

At box 6105, the RRD 109 is deployed. Optionally the BS 101 is deployed. This in- cludes setting up, at certain locations, the RRD 109 and optionally the BS 101. In particular, the relative position between the RRD 109 and the BS 101 may be fixed.

At box 6105, a partitioning of the RRD 109 can be defined, i.e., the REs 1094 of the RRD 109 are grouped into a count of P subsets. The grouping of the REs 1094 means that multiple REs 1094 are included in each subset 691 , 692. This can be in accordance with predefined design rules, e.g., a supported maximum count of spatial data streams. Typically, the maximum of count of spatial data streams can depend on a count of REs. Each subset may have a predetermined minimum count of REs, e.g., not less than 10 or 100 REs. In some examples, it would even be possible that the partitioning of the RRD 109 is dynamically changed or selected from multiple candidate partitionings. For example, the particular partitioning - e.g., defined by the count of subsets and/or the grouping of the REs 1094 into the subsets - may even be re-selected during the communica- tion phase of box 6010. Decision criteria could be count of UEs being served, count of spatial data streams used, etc.

At box 6110, a channel sounding procedure between the RRD 109 and the BS 101 occurs. This can include communicating reference signals (RSs). Both, the RRD 109 and the BS 101 participate in the channel sounding procedure, e.g., by transmitting and/or receiving reference signals and/or providing respective feedback. For exam- ple, reference signals can be communicated on the auxiliary link 199. The BS 102 thus obtains the channel, H, more specifically, the channel for each subset, H _p , p=1...P.

The BS 101 can have knowledge regarding the count of reflection directions, e.g., the BS 101 can have knowledge regarding the steering vector codebook, cf. TAB. 3. For example, the BS 101 can inform the RRD 109 regarding the steering vector code- book; or vice versa.

At box 6115, the BS 101 determines a precoder G to use during the communication phase. The role of this precoder is to allow each subset 691 -692 to select one spe- cific layer from the BS signal for further reflection.

Then, based on the channels H _p for each subset and the precoder G, the compensa- tion vectors are determined for all subsets, to populate the compensation codebooks, at box 6120.

The compensation codebooks are then provided to the RRD 109, together with an in- dication of the particular subset associated with each compensation codebook, at box 6125. Using the compensation codebooks, it is possible to isolate different spa- tial data streams during the communication phase of box 6010.

The steering vector codebooks are predefined and remain fixed during the calibration phase of box 6005. It would be possible that the RRD 109 provides the steering vec- tor codebooks to the BS 101 ; or vice versa. Up to now, the calibration phase of box 6005 has been described. Next, the commu- nication phase of box 6010 will be described. This corresponds to the operations dur- ing payload data communication.

At box 6130, the relative position of the UEs being served by the BS 101 with respect to the RRD 109 is being determined. For example, it is possible to acquire infor- mation about which reflection directions - outgoing beam indices of the RRD 109, cf. beams 671 , 672 in FIG. 3 - that should be used to reach the UEs. Conventional beam management procedures can be used.

At box 6135, it is possible to allocate one or more subsets 691 -692 to each spatial data stream being used by the BS 1012 to serve the one or more UEs 102-103.

In particular, it is possible to flexibly map data streams to subsets, as well as to flexi- bly map UEs to spatial data streams. For instance, each spatial data stream may be mapped to a single subset or to multiple subsets. A UE may be served by one or more data streams.

To give a concrete example, the allocation could be as follows: For a first subset, the respective steering vector codebook indicator defines a reflection direction towards a given UE and the respective compensation codebook indicator defines selection of a first spatial data stream (the first spatial data stream is then associated with the given UE, i.e. , carries signals encoding payload data directed to the given UE); for a differ- ent, second subset, the respective steering vector codebook indicator defines a sec- ond reflection direction towards the same given UE, but the respective compensation codebook indicator selects a different spatial data stream. Thereby, this given UE can be served via two data streams that can be spatially separated at the UE, e.g., because they have different angles-of-arrival. For instance, one of the two data streams may propagate non-line-of-sight, the other one may propagate line-of-sight. Thereby, the overall data rate can be increased. Alternatively, two data streams can be used to transmit payload data redundantly, i.e., a diversity scheme. The net result is an increased array gain and, consequently, larger user data rates.

Each data stream can, in some scenarios, have two sub-streams, associated with dif- ferent polarizations. There can be various decision criteria taken into consideration when allocating sub- sets to spatial data streams; and some of these decision criteria are summarized be- low in TAB. 4.

TAB. 4: Various options for allocating subsets to data streams and/or subsets to UEs. Decision criteria described in TAB. 4 can be combined with each other or with further decision criteria, to thereby form cumulative decision criteria.

The allocation of box 6135 can be implemented at the BS 101 ; the RRD 109 is not required to be involved.

Then, at box 6140, the codebook indices of the compensation codebook are provided to the RRD 109. The compensation vector singles out the transmission layer in- tended for user k.

In typical implementations, the steering vector indices of the steering vector code- book are also provided, by the BS 101 , to the RRD 109. The beam index, which cor- responds to a steering vector, controls the reflection direction. There can be other scenarios in which the RRD 109 selects the appropriate steering vectors for each subset autonomously, e.g., when it has knowledge of the relative positioning of the UEs in its surrounding. At, box 6145, the respective spatial filters are applied at each subset. This can in- clude multiplying the steering vector - as indicated by the steering vector codebook - and the compensation vector - as indicated by the compensation codebook - for each RE of a respective subset and re-configuring each RE accordingly. Then, payload data can be transmitted by the BS 101 via the RRD 109 using the multiple spatial data streams to the one or more UEs 102-103, at box 6150.

As illustrated in FIG. 6, boxes 6130-6150 can be reiterated. In particular, it is possible to repeatedly determine the relative positioning of the one or more wireless communi- cation devices with respect to the RRD 109, e.g., at the BS 101 or the RRD 109.

Also, the allocation of box 6135 can be dynamically updated from time to time. For instance, it would be possible that, at a first point in time, the codebook indicators of the steering vector codebook and the codebook indicators of the compensation code- book of two or more of the subsets define different spatial filters (i.e. , select different spatial data streams, possibly for different UEs); while, at a later, second point in time, the steering vector codebook indicators and the compensation codebook indi- cators of the same two or more subsets define the same spatial filter (i.e., select the same spatial data stream for the same UE).

By such techniques, various effects can be achieved. All UEs (1...K) can receive in- terference free signals simultaneously. The amount of required signaling is kept very low, as the BS needs merely to indicate two codebook indices (cf. TAB. 2); for exam- ple, the number of codebook indices to signal can be as low as the count of UEs. The MIMO multiplexing gain is maintained, since all UEs are simultaneously served.

There is a, potential, power loss of K, since the antenna array is partitioned; in some implementations, this power loss can be substantially smaller.

Next, an example mathematical description of the techniques above, is provided. This will be discussed in connection with FIG. 7 and FIG. 8 and FIG. 9.

Assume an M antenna-port BS (i.e., having M antenna elements 1014, cf. FIG. 2; FIG. 7), K single antenna-port UEs 102-104, and an RRD 109 with N reflective ele- ments 1094. Assume a grouping of the N reflective elements into P subsets, each containing Q = N/P reflective elements 691 -693. In the absence of noise, the signal from the BS 101 to subset p can be expressed as y _p = H _p Gx where x is the transmitted Kx1 data symbol vector, G is an MxK BS precoder matrix, and H _p is the channel having a dimensionality of QxM. The channel is determined based on a channel sounding procedure, cf. FIG. 6: box 6110. The index p counts the subsets 1 ...P formed by the REs of the RRD. Likewise, the output signal at each of the K UEs 102-104 due to subset p can be expressed by a vector r: where a _p (θ _k ) is a steering vector at subset p, of dimension 1xQ, in the direction (solid angle) θ _k towards user k, and z _p is the reflected signal at subset p.

Note that here it is assumed for sake of simplicity that the number of spatial data streams K equals the number of UEs. As a general rule, there can be more spatial data streams than UEs. But it would even be possible that a spatial data stream is shared between two UEs, e.g., for broadcasted or point-to-multipoint payload data.

The reflected signal z _p can, in turn, be expressed as z _p = diag(w _p ) y _p where w _p is a Qx1 vector describing the phase rotations - i.e. , defining the spatial fil- ters provided by each subset - and diag(-) defines a diagonal matrix with its argu- ment along the main diagonal. This defines the operation at each subset, as illus- trated in FIG. 8. The input signal y _p , denoted 4016, and the output signal z _p , denoted 4017, are defined by the spatial filter w _p .

Based on the equations above, one arrives at the following input-output relation

The overall system model is illustrated in FIG. 7. The precoding at 4010 using matrix G applied to the input signal vector x - at 4005, e.g., encoding payload data, cf. box 6150 of FIG. 6 - is shown. Eq. (1 ) describes the input-output relationship for downlink data from the BS to the UE. I.e., r = r _DL .

The RRD 109 includes p=1...P subsets; here three subsets 691-693 are shown. Only the channel H _p to subset 692 is illustrated, 4015. The steering vectors are illustrated at 4020.

RRDs 109 are typically passive, i.e., do not amplify; accordingly, a constraint is that all elements of the vectors w _p have, without loss of generality, unit magnitude; that is, they are complex exponentials (phasors).

Based on the above input-output relation of Eq. (1 ), the objective can be formulated as follows: Find {G, w ₁ , ... , w _P } such that r = diag(f)x. In other words, find a BS pre- coder matrix G and spatial filters for each subset of REs 1094 at the RRD 109, so that interference between the various UEs 102-104 is reduced or zero. Here, f is a Kx1 vector representing signal strengths for the K users; the presence of “diag(f)” eliminates inter-user interference.

However, subject to the constraint that all the elements of the vectors w _p shall have unit magnitude, to find good (in the sense of capacity, SNR, etc.) spatial filters w _p , i.e., to find a good configuration of the RRD 109, is a non-deterministic polynomial- time hard problem, since it entails optimization over complex exponentials which is not a compact set.

There are many possibilities to solve this problem. One example solution of sufficient accuracy for selecting {w _p } will be described below.

To reduce complexity, two codebooks that define together the possible vectors w _p are used (cf. TAB. 2). The first codebook includes entries defining first contributions to the spatial filters w _p defining reflection directions aligned with positions of the mul- tiple UEs in line-of-sight case or generally aligning beams so that communication with the UEs is good; this is the steering vector codebook. The second codebook includes entries defining second contributions to the phase rotations w _p associated with se- lecting different spatial data streams associated with the multiple UEs; this is the compensation codebook. Next, it is described how to construct/populate the codebooks at subset p. The steer- ing vector codebooks for each subset 691-693 are related only to the reflection direc- tions at the RRD, and can described as

A _p = {a _p (φ _m ), 1 ≤ m ≤S,} (2) where S is a design parameter, {φ _m } is a set of beam directions from the RRD, and a _p (φ _m ) is a vector of size 1 x Q. In what follows, it is assumed that the beam direc- tions {φ _m } are selected so densely that one can assume that θ _k = φ _m , for some value m'. This is illustrated at 4020 in FIG. 7.

As a general rule, the steering vector codebook can be fixed and predefined.

As a further general rule, the selection of entries from the steering vector codebook can occur at the BS 101 , e.g., when the BS 101 tracks the position of each UE 102- 104 with respect to the RRD 109. It would also be possible that the selection of en- tries from the steering vector codebooks is made at the RRD 109.

The compensation codebook is of the form

C _p = {c _pq , 1 ≤ q ≤ K,} (3) where each vector in the compensation codebook has, at least in some examples, unit magnitude elements.

To specify the entries in these codebooks that should be applied at subset p, the BS 101 provides, to the RRD 109, the codebook indices (m ^p , q ^p ), one for each of the two codebooks. The operation at a given subset 692 is illustrated in FIG. 8. Then, the spatial filter at each subset is obtained as wherein is Hadamard product, i.e. , element-wise product.

Substituting Eq. (4) into Eq. (1 ) yields

Under the assumption that the steering vectors a _p (θ _k ) are mutually orthogonal over the index k, Eq. (5) becomes where R _{Q k} is a KxQ matrix where row k is all ones, and the other rows have the property that they have mean 0 as Q grows large. The indicator variable /(p) is de- fined as the targeted user index of the p-th subset. Next, the precoder matrix G is dis- cussed; a discussion of how to generate c _pqP follows afterwards.

The precoder matrix G can be selected as the solution to the following optimization problem where P _T is a measure of transmit power, and |Z| is a matrix of the same size as Z, however having elements defined by the magnitudes of the elements of Z . 1 is ma- trix having all elements set to value “1”.

Further, y _min is a parameter used for tradeoff purposes and is related to gain. The parameter may be predefined or may be set differently in different implementations. With a large y _min , one obtains a large array gain, at the expense of possibly larger in- terference. y _min > 0 prevents trivial solution X=0.

To put Eq. (7) into words: G is determined such that the matrices H _p G have elements that have as equal magnitudes as possible. In other words, the goal function of the optimization penalizes differences between elements of each matrix H _p G (i.e. , the matrix obtained from multiplying the precoder matrix defining precoding at the BS with the channel matrix of each subset) for all subsets p; the respective reference value is defined by y _p and can, in general, vary from subset to subset. It would also be possible that y _p is the same for all p, i.e., y _p = y ₀ , ∀p. Other formulations of the goal function optimizations are also possible that penalize, but they should somehow take into consideration that the matrices |H _p G| should have as equal magnitudes across all their elements as possible.

The goal function of Eq. (7) can be solved as an iterative numerical optimization algo- rithm. Various optimization algorithms can be used, e.g., descent methods, evolution- ary methods, pattern search methods, to name just a few.

Based on the precoder matrix, or more generally based on the channels H _p , it is then possible to determine the compensation codebook C _p . Let b _pk denote the kth column of H _p G. The kth compensation vector in C _p is . With that, we have for some vector z that can be interpreted as a gain.

Inserting this into Eq. (5), and assuming that the minimum cost of the optimization problem stated above is 0, yields

Invoking the properties of RQJ( _P ) , one obtains where the matrix E _p,I(p) has the following properties (under assumption that the mini- mum cost of the optimal solution to the optimization problem stated above is 0):

• Element (l(p),l(p)) is Q

• All other elements at row l(p) and column l(p) are zero (under assumption that the steering vectors to different users are mutually orthogonal).

• All other elements are small, in the sense that after division with Q, they tend to 0 as Q grows large.

From this, it follows that UE k receives that is, interference free transmission, with an array gain proportional to the number of antennas at each subset devoted to user k. ‘Interference free’ holds true as long as the UEs are spatially separated; UEs spatially close to each other need to be sep- arated via FDD, TDD or CDD, which are efficient only for this case, i.e. when sharing a beam.

For example, FIG. 9 illustrates the selection of the spatial data stream 4103 from a total of four spatial data streams 4104 to be forwarded to the UE 104 using the output beam 672 - there is a total of seven beams 671-677, corresponding to seven entries in the steering vector codebook. In FIG. 9, this means: m ^p = 2, i.e., use the second beam from codebook A _p to steer the reflection only towards UE 104; as well as q ^p = 3, i.e., use the third compensation vector from codebook C _p to only reflect the third layer 4103. As a further example: If m ^p = 6 and q ^p = 1, then UE 102 would be served by subset p.

Above, various functional features have been explained in connection with operating the RRD to support multiple spatial data streams. The techniques described herein are flexible in the sense that they support various geometries of reflective surfaces of the RRD, e.g., different arrangements of REs such as antennas or meta-material unit cells. As a general rule, the particular geometry of the REs of the RRD and their grouping into subsets is not germane for the functioning of the techniques. This also applies to the arrangement of the antennas of the BS. For instance, various geome- tries can be considered in the optimization according to Eq. (7); here, the channel matrices H _p will generally depend on the arrangement of the REs of the RRD, as well as the arrangements of the antennas of the BS.

Hereinafter, an example geometry of the RRD will be described that has certain ben- efits, in particular, an analytical solution to the optimization according to Eq. (7) can be possible, in particular when combined with a specific configuration of antenna ele- ments of the BS. FIG. 10 schematically illustrates an arrangement of the REs 1094 of the RRD 109, as well as of the antenna elements 1014 of the BS 101 .

As illustrated in FIG. 10, the RRD 109 again includes multiple REs 1094 that are grouped into a plurality of subsets 691 -697. For each one of the subsets 691 -697, the REs of the respective subset 691-697 are arranged on a respective circle 701 -703, having different diameters. For instance, in the illustrated scenario of FIG. 10, the REs 1094 of the subset 691 are arranged on an inner ring/circle 701 , the REs 1094 of the subsets 692-693 are arranged on a middle circle 702, and the REs 1094 of the subsets 694-697 are arranged on the outer circle 703.

As illustrated, the circles 702 and 703 thus accommodate reflective elements 1094 grouped into multiple respective subsets 692-693, as well as 694-697. The reflective elements 1094 of each one of the subsets 692-693 are alternatingly arranged on the circle 702; likewise, the reflective elements 1094 of the subsets 694-697 are alternat- ingly arranged on the circle 703. More specifically, the reflective elements of each subset 691-697 are evenly spaced on the respective circle 701 -703. Thereby, for the REs 1094 of each one of the subsets 691 -697, a b-fold rotational symmetry with re- spect to the center point of each circle 701 -703 of the respective subset 691 -697 is obtained, where b is the count of REs 1094 per respective subset 691-697.

For example, it would be possible that the count of REs per circle 701-703 is a multi- ple of some base integer number Q. The partitioning of the RRD 109 into the subsets 691 -697 is to let Q antennas from each circle 701 -703 constitute one subset 691 - 697. All subsets 701 -703 thus include the same count of REs 1094. This is generally optional; it would also be possible that different subsets 701 -703 include different counts of Res 1094.

The circles 701 -703 are concentrically arranged; a center axis 710 is illustrated, as illustrated in the scenario of FIG. 10, also the antenna elements 1014 of the BS 101 are arranged on a respective circle 704 that is concentrically arranged with respect to the center axis 710. By such a configuration, orthogonal angular momentum electromagnetic modes can be selected for the propagation of the electromagnetic waves. It is possible to analyti- cally solve the optimization according to Equation (7). Very briefly, this is because the channel matrices of the subsets 692-693, as well as 694-697 of each circle 702-703 including more than a single subset are identical due to the rotational symmetry. De- tails will be described below.

The effect of such layout is that the channel matrix H thereby has the form where F _Q denotes the DFT matrix of size Q, and * is complex conjugation. The matrix ∑ has all elements equal to zero, except along its 0, K, 2K, ... ,(Q/K-1 )K lower diago- nals. This matrix allows one to trivially solve the optimization problem of Eq. (7) by selecting G as any K columns from F _M . The compensation vectors can be taken as K columns from . The gains, z are related to the entries of ∑. To maximize those, the K columns from F _M should be selected as the K columns in which the largest ele- ments of E are found. The columns of are the rows of ∑ in which these largest ele- ments are to be found. The gain z _k is the largest element of column k in ∑. In case this is a complex number, the compensation vector should be phase rotated to make it real, so that coherent combining at the receiver side is obtained.

In cases where the BS 101 and the RRD 109 are not lined up concentrically with re- spect to a common center axis 710, some minor residual inter-user interference will still be present. This can be straightforwardly handled by incorporating a zero-forcing precoder into G.

FIG. 11 is a flowchart of a method according to various examples. For example, the method of FIG. 11 could be implemented by an RRD such as the RRD 109 dis- cussed above. It would be possible that the method of FIG. 11 is implemented by the processor 1091 upon loading respective program code from the memory 1093. The RRD can communicate with an access node and optionally with one or more UEs, e.g., via auxiliary radio carriers and respective control channels. Optional boxes are marked with dashed lines in FIG. 11 . At optional box 8005, the RRD participates in the channel sounding procedure with the access node. Here, reference signals can be transmitted and/or received on an auxiliary radio carrier (such as the auxiliary radio carrier 199 as discussed in connec- tion with FIG. 4). Upon participating in the channel sounding procedure, the method proceeds to box 8010.

At optional box 8010, compensation codebooks are obtained for each one of multiple subsets. Each subset includes multiple REs. The compensation codebooks specify settings for each respective RE. Details of the compensation codebook have been discussed above in connection with TAB. 3. The compensation codebook is also dis- cussed above in connection with Equation (3). At box 8010, the set of indices c _pq can be obtained.

Box 8005 and box 8010 can be part of the calibration phase as discussed in connec- tion with box 6005 of FIG. 5.

At box 8015, the RRD may participate in the positioning procedure with one or more UEs that are arranged in the surrounding area of the RRD and served or potentially being served via the RRD. In other examples, such positioning may be implemented at the access node, transparent to the RRD.

Next, at box 8020, the RRD obtains, from an access node, as well as for each subset of multiple subsets, a respective compensation codebook indicator indicative of a re- spective entry of the compensation codebook.

It is optionally possible that the RRD also obtains, from the access node, a respective steering vector codebook index for each subset. In other examples, the selection of the entries of the steering vector codebook can be made autonomously at the RRD device based on the relative positioning between the RRD and the one or more UEs in its surrounding, based on the positioning procedure of box 8015. For instance, the access node may indicate which specific UEs are being served via the RRD and then the RRD can check whether the respective positioning procedure has been per- formed.

Based on the selected entries of the respective steering vector codebook and the re- spective compensation codebook, the REs of each subset are reconfigured at box 8025. Then, payload data can be communicated between the access node and one or more UEs.

FIG. 12 is a flowchart of a method according to various examples. For example, the method of FIG. 12 could be implemented by an access node, e.g., the BS 101 dis- cussed above. It would be possible that the method of FIG. 12 is implemented by the processor 1011 upon loading respective program code from the memory 1015. The access node can communicate with one or more UEs via an RRD. The access node can also communicate with the RRD, e.g., via an auxiliary radio carrier in a respec- tive control channel. Optional boxes are marked with dashed lines in FIG. 12.

At optional box 8105, the access node participates in a channel sounding procedure with the RRD. Box 8105 is interrelated with box 8005 of FIG. 11 .

At optional box 8110, the precoder matrix is determined, and based on the precoder matrix, the compensation codebooks for each subset are determined. Respective techniques have been described in connection with Eq. (7).

Then, the compensation codebooks are provided to the RRD, at optional box 8115. Box 8115 is thus interrelated with box 8010 of FIG. 11 .

At optional box 8120, the positions of one or more UEs that are being served via the RRD are determined, with respect to the RRD. This allows one to select the appropri- ate entries of the steering vector codebooks, for each subset. The outgoing beams can be selected, cf. FIG. 9, beams 671 -677.

For this, a mapping between data streams - subsets - UEs can be used and the re- spective mapping can be determined at box 8125, by allocating, for each one of one or more UEs being served via the RRD, one or more subsets to at least one spatial data stream associated with the respective UE.

Details with respective decision criteria to be considered at box 8125 have been dis- cussed above with respect to TAB. 4.

Next, at box 8130, the compensation codebook indices and optionally steering vector indices are provided to the RRD. Box 8130 thus is inter-related with box 8020. Summarizing, an SDMA technique to separate multiple spatial data streams at the RRD has been described. Here, the REs of the RRD are not modifying the amplitude of the electromagnetic waves. The REs can be passive elements that only impose a variable phase shift. The RRD is divided into subsets. This generally can result in a tendency to reduce the array gain; in particular if compared to a full superposition of the RRD precoders without partitioning (which, ideally would require amplitude and phase shifts to each RE; for passive REs only capable of providing a phase shift, it is a mathematically complex challenge).

The techniques described herein can be transparent to the UE. The techniques de- scribed herein can use the full bandwidth and the RDD does not need to switch beams with high rate, allowing for low power and no need of synchronizing the RRD with a timing reference of a communication protocol used by the BS and the UE to communicate payload data via the RRD.

Although the invention has been shown and described with respect to certain pre- ferred embodiments, equivalents and modifications will occur to others skilled in the art upon the reading and understanding of the specification. The present invention in- cludes all such equivalents and modifications and is limited only by the scope of the appended claims.

For illustration, above, various scenarios have been discussed according to which the BS provides the steering vector codebook indices to the RRD. For instance, such a scenario does not require that the UE is aware of the presence of the RRD. How- ever, other scenarios are conceivable and according to such scenarios, it would be possible that a positioning procedure is performed between the RRD and each UE of the one or more UEs. Then, the RRD can be aware of the relative positioning of the one or more UEs and its surrounding and may, autonomously, without involvement of the BS, select the appropriate steering vector codebook index for each one of its sub- sets. Still, the BS can inform the RRD regarding the compensation codebook index for each subset. For further illustration, above various scenarios have been described in connection with downlink data transmitted from the BS via the RRD towards the UE. For recipro- cal spatial streams, it is possible to use respective precoding and RRD beams also for uplink data from the UE via the RRD to the BS. For downlink data x _DL , Eq. 1 provides:

This can be used to determine the input-output relationship for uplink data as:

Note that with this approach, there are no additional constraints imposed on the UEs.

For still further illustration, examples have been described in which subsets of REs are disjointed, i.e. , each RE is grouped into a single subset. As a general rule, it would be possible that subsets overlap. Thereby, the array gain can be potentially in- creased.