Method for determining processes for generating precoding parameters and merging parameters for rate splitting multiple access in MU-MIMO communication systems, and transmitter and receiver

Abstract

A method of determining a procedure for generating a precoding parameter and a merge parameter in an MU-MIMO RSMA communication system is presented. The determination process uses selectable target attributes of the communication connections, selectable processes for processing respective errors associated with estimated channel coefficient matrices for all communication channels, and selectable design techniques as inputs. Further, a method of generating precoding parameters and merge parameters for wireless interfaces of a first communication device and a second communication device, respectively, according to a previously determined procedure is presented. The generation process provides a joint determination of precoding parameters and merge parameters in an MU-MIMO RSMA communication system in which the CSI is only imperfectly known. Yet further, a method for operating a first wireless communication device and a second wireless communication device in the MU-MIMO RSMA communication system using precoding parameters and merge parameters determined according to the process is presented.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the field of wireless communications, and more particularly, to wireless communications using rate-splitting multiple access (RSMA) in multi-user multiple-input multiple-output (MU-MIMO) communication systems.

[0002] Notation

[0003] Scalar values are denoted in italic lower case letters as x, while complex vectors and matrices are denoted in bold lower case letters and upper case letters as x and X, respectively. Complex tensors are denoted in calligraphic bold upper case letters as H. T and (·) * denote the transpose operator and the complex conjugate operator, respectively, and diag(·) denotes the diagonalization operator. |·| denotes the absolute value operator, while l denotes the l-norm. and Var x (x) denote the expectation operator and the variance operator of x with respect to the x distribution given by and denote the real and complex number fields, respectively, and XN(μ,ν) denotes the real and complex Gaussian distributions with mean μ and variance ν. BACKGROUND

[0004] The current fifth generation (5G) and upcoming sixth generation (6G) wireless communications and beyond are designed to serve a large number of high-mobility users, e.g., vehicles, subways, highways, trains, drones, low earth orbit (LEO) satellites, etc.

[0005] The core requirements of 5G communications include serving data-driven use cases with data rate requirements up to 20 Gbps in the downlink (DL), i.e., enhanced mobile broadband (eMBB), providing ultra-reliable low-latency communications (URLLC) with 10 -5 or less block error rate (BLER) and 1 ms or less latency, and providing unlicensed access in the uplink (UL) to a large number of low-complexity and low-power devices, especially for implementing massive machine type communications (mMTC). These requirements can not necessarily be satisfied simultaneously. The core requirements of 6G communications go beyond 5G, including simultaneously satisfying eMBB and URLLC, simultaneously satisfying enhanced eMBB and mMTC, enhanced URLLC and mMTC, and simultaneously satisfying enhanced eMBB, URLLC, and mMTC, albeit based on trade-offs, i.e., accepting compromises in any one or more of these three.

[0006] ​Various methods are known to ensure correct access of a plurality of user equipment (UE) units to a base station (BS) using shared radio resources. Initially deployed communication systems typically use so-called orthogonal multiple access (OMA) schemes, which can be considered to serve a single user per resource. Recent developments have led to the emergence of non-orthogonal multiple access (NOMA) methods, which can be considered to serve multiple users per resource. This simple distinction does not fully reflect modern communication design, i.e. OMA-based communication networks actually serve multiple users on orthogonal resources using time division multiple access (TDMA), frequency division multiple access (FDMA), code division multiple access (CDMA) or orthogonal frequency division multiple access (OFDMA). In addition, these modern communication systems are typically equipped with multiple antennas and can further extend multiple user access through spatial domain processing in the form of multi-user linear precoding (MU-LP), spatial division multiple access (SDMA), multi-user multiple-input multiple-output (MU-MIMO) and massive MIMO. MU-LP, SDMA, MU-MIMO serve users in a non-orthogonal way, as multiple users are assigned different precoders, resulting in different “beams” pointing to respective different users in the same time-frequency grid and interfering with each other in the same cell. All these multiple user access schemes require proper interference management at the transmit side or receive side to enable proper interference cancellation (IC).

[0007] Existing 5G communications have faced many challenges, such as multi-user interference due to imperfect channel state information (CSIT) at the transmitter when performing MIMO beamforming. In particular, CSIT can become outdated due to high mobility (where the channel changes during the processing time required to determine the CSI) and channel blockage (due to objects appearing in the wireless communication path while processing the CSI).

[0008] Conventional multi-user multi-antenna methods, such as SDMA, MU-MIMO, rely heavily on timely and highly accurate CSIT or receiver-side CSI (CSIR). In practice, CSIT / R is never perfect, especially due to pilot reuse, channel estimation (CE) errors, pilot contamination, limited and quantized feedback accuracy, delays and latencies, mobility (increasing speeds of vehicles, trains, satellites, flying objects, and emerging applications such as vehicle-to-everything), radio frequency (RF) impairments (e.g., phase noise, inaccurate calibration of RF chains, sub-band level estimation, etc.).

[0009] Rate-splitting multiple access (RSMA) has recently emerged as a powerful multiple access, interference management, and multi-user strategy in next-generation communication systems. RSMA refers to a broad class of multi-user schemes whose versatility relies on the principle of rate-splitting (RS). RS consists of splitting a message into a respective common and private part, distributing the common part to be encoded and precoded into a common stream, and distributing the private part to be encoded and precoded into a private stream, and superimposing the common stream on top of all private streams in a non-orthogonal fashion, i.e., transmitting the common and private streams simultaneously.

[0010] In the downlink, RSMA uses linear or non-linear precoded RS at the transmitter, i.e., at the base station (300), to split each user message into one or more common and private messages. The common messages are combined and encoded into a common stream for the intended user. The common stream is decodable by all receivers, while the private stream is only decodable by its corresponding receiver. The receiver has to retrieve each part to reconstruct the original message. After decoding the common stream from the received signal, the receiver applies successive interference cancellation (SIC) or any other form of joint decoding on the common stream to enable correct decoding of the private stream. The decoded common and private streams are combined to retrieve the originally transmitted message.

[0011] One major advantage of RS and its message-splitting functionality is the flexible management of inter-user interference. Indeed, RS can be seen as a combination of transmit-side interference cancellation and receive-side interference cancellation, where the contribution of the common stream can be adjusted according to the level of interference the receiver needs to cancel. This deviates from the only transmit-side interference cancellation strategy of SDMA and the only receive-side interference cancellation strategy of NOMA, respectively.

[0012] Using RSMA in MU-MIMO systems, i.e., systems where the BS and the UEs have multiple antennas configured for beamforming, also known as spatial multiplexing, requires proper precoding in the transmitter for proper beamforming and proper combining in the receiver to best exploit the signals of all antennas. Spatial multiplexing introduces additional multi-user interference, especially due to imperfect beamforming that inevitably “leaks” a part of the signal to other UEs that are not the target of the beam, which needs to be handled in the receiver. While the common channel part of RSMA can still provide useful information for those UEs that are not the target of the beam for performing CE and IC, there is currently no joint precoder and combiner at the transmit side that takes into account CSI uncertainty and the resulting imperfect SIC in the receiver of a MU-MIMO RSMA system, thus making the MU-MIMO RSMA system prone to performance degradation. This challenge is particularly difficult to solve in heterogeneous systems, where different UEs have different numbers of antennas, and known methods cannot be used or perform severely degraded in this case. SUMMARY

[0013] It is therefore desirable to provide an improved method for determining precoding parameters and combining parameters for a wireless device of a MU-MIMO RSMA communication system, and to provide a corresponding receiver and transmitter (to accommodate for the case of imperfectly known CSI at the transmitter and / or receiver), as well as methods of operating the receiver and transmitter, respectively. It is further desirable to provide methods and apparatuses that can be used in a communication system with UEs having different numbers of antennas without suffering from a severe performance degradation.

[0014] This need is solved by the method of determining a process for generating precoding parameters and combining parameters as presented in claim 1, the method of generating precoding parameters and combining parameters as presented in claim 5, the method of operating a first wireless communication device as presented in claim 11, the method of operating a second wireless communication device as presented in claim 14, the wireless communication device as claimed in claim 18, and the computer program product as claimed in claim 19. A corresponding computer readable storage medium is presented in claim 20. Embodiments and developments of the methods and apparatuses are provided in the respective dependent claims.

[0015] In particular, the methods described hereinafter take into account the problem in the downlink direction of such MU-MIMO RSMA systems, i.e. the imperfectly known CSI at the receiver severely hinders the decoding process (e.g. the SIC process), and thus the detection of the transmitted symbols in the receiver.

[0016] The present application will be described hereinafter assuming an exemplary MU-MIMO RSMA communication system comprising a first wireless communication device (e.g. a base station (300) having N t ≥ 1 transmit antennas) and K second wireless communication devices (e.g. each having M k ≥ 1 antennas) user equipments (400). In such a system, the RSMA transmitted signal is given by

[0017]

[0018] where s c ~ XN(0, IN Lc ) and are the common signal and the precoder matrix for the common signal of length L c , respectively, and s k ~ XN(0, IN Lk ) and are the private signal of the k-th UE and the precoder matrix for the private signal of length L kthe precoder matrix of the private signal of the kth UE. K is the index set of all receivers or UEs.

[0019] The received signal y k at the kth UE is denoted as

[0020]

[0021] where, Hkis the actual channel matrix between the base station (300) and the kth UE, n k is the received additive white Gaussian noise (AWGN) vector,

[0022] At the kth receiving UE side, the message of interest initially is the common signal s c which is the component s k carried in the received signal y c directly detected, and the kth private signal s k which is obtained by applying serial interference cancellation (SIC) to the received signal with the known information of the estimated common signal s c .

[0023] The received common signal y c,k can be written as

[0024]

[0025] where, U c,k denotes the combiner matrix of the common message at the kth receiver or UE, and is the additive white Gaussian noise (AWGN) at the kth receiver or UE.

[0026] In the ideal case assumed above, the actual channel coefficient matrix H k is perfectly known by each UE, which allows perfect SIC to be performed at the receiver, resulting in soft replicas

[0027]

[0028] where, U k is the combiner matrix of the private signal of the kth receiver or UE.

[0029] Based on the above system description, the achievable total rate R total of the RSMA transmission from the BS to the kth receiver or UE, using the corresponding rates R c,k and R k of the common signal and the private signal of the kth receiver, respectively, is obtained as

[0030]

[0031] The SINRs of the common and private messages are given by

[0032]

[0033] and

[0034]

[0035] where U c,k and U k are the combiner matrices of the common and private signals at the receiver, respectively.

[0036] In the MU-MIMO case discussed herein, the estimated recovered common signal is expressed as

[0037]

[0038] where y k is the received signal, V c is the beamformer matrix at the transmitter, U c,k is the beamformer matrix at the receiver, and H k is the channel coefficient matrix, which is assumed to be ideal in this case.

[0039] The estimated recovered private signal is expressed as

[0040]

[0041] The preceding discussion assumes perfect knowledge of the CSI at all receivers and that all UEs are configured identically, e.g., all receivers have the same number of antennas, i.e.,

[0042] However, in a real scenario, the UEs in the system will have different antenna configurations (i.e., the system is heterogeneous), and for some or all k can have different number of antennas M k This will significantly degrade the robustness and performance of the communication.

[0043] Further, in a real scenario, the actual channel coefficient matrix H k at the receiver is unknown, making the SIC imperfect, resulting in a residual interference term due to the CSI error, which leads to a severe degradation in the receiver performance. This interference is represented by the term in the following equation, which represents the soft copy of the received signal under this imperfect condition

[0044]

[0045] wherein,

[0046]

[0047] wherein, is an estimated channel coefficient matrix shared between the BS and the kth UE with imperfect knowledge, and is the corresponding estimation of the imperfectly known CSI of the communication channel between the BS and the kth UE and the shared "error" part. Note that in this specification, the expression "estimated and shared" means that the estimated information is communicated (i.e., shared) by the estimating entity to one or more other entities in the system.

[0048] In the case of imperfectly known CSI, the estimated recovered common signal and the private signal can be reformulated as

[0049]

[0050] wherein the estimated channel coefficient matrix takes into account the imperfectly known CSI. The precoder matrix and the combiner matrix V c , V k , U c,k , U k are designed to incorporate the heterogeneity in the number of antennas of the multiple receivers, as will be further discussed below.

[0051] In the case of imperfect SIC, the SINR of the private message is given by

[0052]

[0053] wherein, represents the residual interference due to imperfect SIC caused by the imperfect CSI.

[0054] It is apparent that a real RSMA system exhibits a rate loss from this residual interference on top of the multi-user interference, which is ultimately caused by the imperfectly known CSI at the receiver.

[0055] The present invention solves this problem by jointly determining the precoding parameters or matrix V and the combining parameters or matrix U at the transmitter or BS and providing them to the receivers or UEs. Depending on the respective communication protocol used in the communication system, the CSI can be determined in the BS or in the UEs and provided to the BS to determine the precoding parameters and the combining parameters.

[0056] The UE uses the precoding parameters or matrix and the combining parameters or matrix to improve the signal estimation and recovery, and thus the detection of the transmitted symbols. To this end, it is assumed that the UE has access to the precoding parameters or matrix V and the combining parameters or matrix U via ideal feedback, as well as the estimated CSI used in the transmitter, if not previously determined in the UE.

[0057] Figure 1 The main components of, for example, a corresponding transmitter in a base station 300 and, for example, a receiver in a UE 400 are shown, respectively. Note that when the estimated CSI is determined in the UE, the UE provides the CSI to the BS via the same ideal feedback.

[0058] In the base station 300, after the signal to be transmitted to a plurality of UEs is split into a common part and a plurality of corresponding private parts, and after the common signal and the private signals are encoded, the resulting signal s is provided to a precoder 302. A beamformer (BF) 304 provides the precoding matrix V to the precoder 302, which outputs a signal x that is finally transmitted via a plurality of antennas 306 of the base station 300. Transmitting the respective precoded signals through the plurality of antennas effectively results in an electronic beamforming of the private parts of the transmission towards the respective receivers. The beamformer 304 jointly determines the precoding matrix V and the combiner matrix U from the estimated channel coefficients provided in the channel coefficient matrix The precoding matrix V and the combiner matrix U are transmitted to the UE 400 via an ideal feedback link 399, i.e., it can be assumed that they are fully available at the UE 400 when decoding the transmitted signal. The precoding matrix V and the combiner matrix U are transmitted to the UE 400 via an ideal feedback link 399, i.e., it can be assumed that they are fully available at the UE 400 when decoding the transmitted signal.

[0059] At the UE 400, the transmitted signal is received via a plurality of antennas 402, and the received signal y is provided to combiners 404a, 404b. The combiner 404a combines the common message part of y c,k using the combiner matrix U k and outputs the combined received signal y c,k to a decoder 408 configured to decode the common signal. The combiner 404b combines the private message part of y k using the combiner matrix U k and outputs the combined received signal U k y k to an interference cancellation (IC) unit 410 of a detector 406. The combiner uses the previously received combiner matrix U to electronically beamform towards the transmitter. Based on the common signal output from the decoder 408 and the combined received signal U k y k, the IC unit 410 determines a version of the received signal with greatly reduced interference, which is provided to a decoder 412 configured to decode the private signal. The detector 406 outputs an estimated signal representing the transmitted common and private signals

[0060] Before describing embodiments of the proposed application in further more detail, in Figure 2 and Figure 3 the signal or message flow in an exemplary assumed communication protocol in a downlink RSMA system, time division duplex (TDD) and frequency division duplex (FDD) is shown, respectively.

[0061] Figure 2 A swim-lane diagram showing the messages exchanged between a BS and a target UE in a DL direction in a TDD communication system and the corresponding processing invoked at the respective ends. First, the target UE transmits a pilot signal to the BS. The pilot signal can be part of a regular communication transmission from the target UE to the BS. The BS uses the pilot signal to perform a CE, i.e. estimate a channel coefficient matrix and determine a precoder matrix V for electronically beamforming the appropriate signals transmitted by the N t ≥1 transmit antennas towards the target UE, and a combiner matrix U. The estimated channel coefficient matrix The precoder matrix V and the combiner matrix U output from the BF are fed back to the UE, which stores the matrices for decoding the received messages. Next, the BS splits a message to be sent into a common part intended for multiple UEs and a private part intended only for the target UE, i.e. produces a RSMA signal. The RSMA signal is then precoded using the precoder matrix V, resulting in N t ≥1 transmit signals for each of the N k ≥1 antennas of the target UE, and uses the precoder matrix V and the combiner matrix U previously received from the BS to combine the signals. After decoding the common and private messages, they can be combined into the originally sent message. t

[0062] Figure 3 A swim-lane diagram showing the messages exchanged between a BS and a target UE in a DL direction in a FDD communication system and the corresponding processing invoked at the respective ends. Here, the BS first transmits a pilot signal to the target UE. Similar to the reference Figure 2 ​In the previously discussed protocols, pilot signals can be part of the regular communication transmission from the BS to the target UE. The target UE uses pilot signals to perform CE, i.e., to estimate the channel coefficient matrix. And estimate the channel coefficient matrix Transmitted to the BS. The BS uses the channel coefficient matrix. In the beamformer, the method for passing through N is determined accordingly. t At least one transmit antenna transmits appropriate signals toward the target UE to perform electronic beamforming. A precoder matrix V and a combiner matrix U are also used. The precoder matrix V and combiner matrix U output from the BF are fed back to the UE, which stores the matrices used to decode the received messages. Next, the BS splits the message to be transmitted into a common part intended for use by multiple UEs and a private part intended only for the target UE, i.e., generating an RSMA signal. The RSMA signal is then precoded using the precoder matrix V to generate the BS's N... t The transmission signals of each of ≥1 transmitting antennas are transmitted, and these transmission signals are transmitted, which will effectively result in electronic beamforming toward the target UE. The target UE is in its M k Received by BS at ≥1 antenna location N t Signals transmitted by ≥1 transmit antenna are combined using a precoding matrix V and a combiner matrix U previously received from the BS. After decoding public and private messages, they can be merged into the originally transmitted message.

[0063] The two exemplary communication protocols briefly discussed above ensure that the BS has all the information needed in the BF to determine the precoder matrix V and combiner matrix U for transmission to the UE. Providing the UE with information about the precoder matrix V and combiner matrix U output from the BS enables improved signal recovery in the UE.

[0064] from Figures 1 to 3 As can be seen from the discussion, determining the precoder matrix and combiner matrix in the beamformer is a crucial element in performing communication between the BS and the UE. According to the present invention, the beamformer can be adaptable or configurable, thereby enabling the provision of the most suitable precoder matrix and combiner matrix for changing communication requirements and environments.

[0065] Figure 4 An exemplary simplified block diagram of such an adaptable and configurable block 304 in the BS is shown, which handles BF design and configuration to determine beamforming in the BS and merging in the UE. k The precoding matrix V and combiner matrix U required for signals received at ≥1 antenna. In other words, the adaptable and configurable block 304 adaptably designs the BF before determining the precoding matrix and combiner matrix suitable for the corresponding communication requirements and environment.

[0066] The inputs of the BF design block 304 are the estimated channel coefficient matrix and the corresponding matrix representing the error statistics of The output is the precoder matrix V and the combiner matrix U optimized for one of the various specific objectives discussed below. The actual block 304 generating the output from the input signals is shown as a "black box", whose exemplary embodiments will be discussed in more detail below.

[0067] The BF design considers three main elements: CSI imperfection incorporation, beamforming objective, and design technique. Each main element considered in the BF design can have at least two options or embodiments, as exemplarily shown in the following list:

[0068] CSI imperfection incorporation can be done by

[0069] a) averaging, or

[0070] b) estimating the worst-case channel

[0071] The objective of the optimization can be

[0072] a) sum-rate maximization,

[0073] b) minimum-rate maximization, or

[0074] c) power consumption minimization under rate guarantee

[0075] The actual optimization procedure can invoke one of the following design techniques

[0076] a) convex optimization, or

[0077] b) tensor decomposition

[0078] Figure 4 The simplified block diagram shown is presented in more detail in Figure 5 , showing possible combinations of the main elements. As in Figure 4 , the inputs of the BF design block 304 are the estimated channel coefficient matrix and the corresponding matrix representing the error statistics of They are provided to block 304a. In block 304a, the error value of the estimated channel coefficient matrix is determined according to the previous selection of the method to be applied. According to the above list, a choice can be made between averaging the CSI error (block 304a-i) or assuming a worst-case CSI error (block 304a-ii).

[0079] In block 304b, the optimization objective is selected among maximizing the sum rate (block 304b-i), maximizing the minimum rate (block 304b-ii), and minimizing the transmit power while achieving a guaranteed rate (block 304b-iii). Finally, in block 304c, it is determined whether to use iterative convex optimization (block 304c-i) or tensor decomposition (block 304c-ii) to determine the precoding matrix V and the combiner matrix U. Based on the above exemplary list, possible combinations of one of the two alternative options for obtaining estimates of the CSI error statistics, one of the three alternative objectives for precoding, and one of the two design techniques that can be used to determine the precoding matrix V and the combiner matrix U are indicated by the lines connecting the various blocks.

[0080] Based on the options or embodiments in the above exemplary list, twelve different BF designs for determining the precoding and combining parameters can be obtained:

[0081] 1. RSMA precoder and combiner matrices determined by maximizing the sum rate with iterative convex optimization under average CSI error conditions

[0082] 2. RSMA precoder and combiner matrices determined by maximizing the sum rate with iterative convex optimization under worst-case CSI error conditions

[0083] 3. RSMA precoder and combiner matrices determined by maximizing the minimum rate with iterative convex optimization under average CSI error conditions

[0084] 4. RSMA precoder and combiner matrices determined by maximizing the minimum rate with iterative convex optimization under worst-case CSI error conditions

[0085] 5. RSMA precoder and combiner matrices determined by minimizing the power (under rate guarantee) with iterative convex optimization under average CSI error conditions

[0086] 6. RSMA precoder and combiner matrices determined by minimizing the power (under rate guarantee) with iterative convex optimization under worst-case CSI error conditions

[0087] 7. RSMA precoder and combiner matrices determined by maximizing the optimized sum rate with tensor decomposition under average CSI error conditions

[0088] 8. RSMA precoder and combiner matrices determined by maximizing the optimized sum rate with tensor decomposition under worst-case CSI error conditions

[0089] 9. Determining RSMA precoder matrix and combiner matrix by maximizing minimum rate under average CSI error condition by applying tensor decomposition

[0090] 10. Determining RSMA precoder matrix and combiner matrix by maximizing minimum rate under worst case CSI error condition by applying tensor decomposition

[0091] 11. Determining RSMA precoder matrix and combiner matrix by minimizing power (under rate guarantee) under average CSI error condition by applying tensor decomposition

[0092] 12. Determining RSMA precoder matrix and combiner matrix by minimizing power (under rate guarantee) under worst case CSI error condition by applying tensor decomposition

[0093] According to a first aspect of the present application, there is provided a method of determining a process for generating precoding parameters and / or combining parameters for a wireless interface of a first communication device and a second communication device, respectively. The first communication device is configured for wireless communication with a plurality of second communication devices in a MU-MIMO communication system, i.e. each of the first and second wireless communication devices has a plurality of antennas. The method comprises receiving, for all communication channels with all of the plurality of second communication devices, selection inputs for selecting, respectively, a target property of a communication connection, a process for processing an estimated channel coefficient matrix associated respective errors, and design techniques. The selection inputs can be provided by a general communication device configuration, by a preset configuration for a specific message or data type, etc. The method further comprises selecting, from a plurality of target properties of a communication connection, one target property according to the corresponding selection input, which target properties include, inter alia, maximization of a sum rate (i.e. a sum of rates of all connections between the first communication device and the plurality of second communication devices at any given time), maximization of a minimum rate (i.e. maximization of a lowest or worst case rate of each of the second communication devices), or minimization of transmitter power consumption while enabling a guaranteed rate. The latter target property can result in a guaranteed rate of each of the second communication devices at a lowest transmission power, or a guaranteed sum rate over all second communication devices at a lowest transmission power, depending on system requirements. The target properties can also be referred to as targets in the present specification, and can be selected to be valid for all connections initiated or terminated at the BS. The method yet further comprises selecting, according to the corresponding selection input, a process for processing an estimated channel coefficient matrix one of a plurality of processes associated with a respective error. The processes for error handling can comprise, inter alia, averaging over statistical errors or estimating worst-case errors. The averaging can use as input the expected value of the channel estimation error for each connection between the first communication device and one or more second communication devices, which depends on the respective SNR of the pilot signal communication and a channel model, e.g., The method further comprises selecting one of a plurality of design techniques for determining the precoding parameters and / or combining parameters V c , V k , U c,k , U k in dependence on the corresponding selection input. The design techniques can comprise, inter alia, iterative convex optimization or tensor decomposition. Further, the method comprises implementing and configuring the process for generating the precoding parameters and / or combining parameters V c , V k , U c,k , U k using the selected target properties of the communication connection, the selected error handling and the selected design technique. The process for generating is configured to use at least the estimated channel coefficient matrix and the output from the error handling as input. The precoding parameters and / or combining parameters V c , V k , U c,k , U k may comprise scalar values or can be arranged as vectors or matrices.

[0094] In one or more embodiments, the method according to the first aspect of the application is invoked at least in one of the following instances:

[0095] - at predetermined intervals,

[0096] - when a new second wireless communication device (400) joins the plurality of second wireless communication devices (400) connected to the first communication device (300),

[0097] - when one or more of the second wireless communication devices (400) leave the plurality of second wireless communication devices (400) connected to the first communication device (300),

[0098] - when the channel coefficients of at least one of the plurality of second wireless communication devices (400) connected to the first communication device (300) change,

[0099] - and / or when the data message content and / or type to be transmitted to one or more of the plurality of second wireless communication devices (400) connected to the first communication device (300) changes.

[0100] Any of the first or second wireless devices can also require or initiate such a call. This ensures that the target properties can dynamically adapt to changing requirements. The embodiment can further comprise negotiating or selecting a new target property, a process for handling the respective error associated with estimating the channel coefficient matrix , and / or a design technique for determining the precoding parameters and / or the combining parameters V c , V k , U c,k , U k . The embodiment can further comprise negotiating or setting a time when the new parameter set is to be used. Obviously, the earliest dynamic adaptation of the generation process is only possible within the next transmission interval.

[0101] Implementing the process can comprise providing computer program instructions and / or data from a non-volatile memory, which computer program instructions and / or data represent a target property set of the communication connection, a process for handling the error associated with estimating the channel coefficient matrix , and a computer-implemented algorithm for determining the precoding parameters and / or the combining parameters V c , V k , U c,k , U k .

[0102] Figure 6 An exemplary flow chart of a method 100 according to the first aspect of the application is shown. In step 102, it is checked whether the method is called. In the affirmative (the "yes" branch of step 102), a selection input is received in step 110, which selection input is used to select, respectively, a target property of the communication connection, a process for handling the respective error associated with estimating the channel coefficient matrix of all communication channels, and a design technique. In step 120, one target property set of the plurality of target property sets of the communication connection is selected according to the selection input. In step 130, one process of the plurality of processes for handling the respective error associated with estimating the channel coefficient matrix is selected according to the selection input. In step 140, one of the plurality of design techniques for determining the precoding parameters and / or the combining parameters V c , V k , U c,k , U k is selected according to the selection input. Finally, in step 150, a process for determining the precoding parameters and / or the combining parameters V c , V k , U c,k , U k is implemented and configured according to the selected target property of the communication connection, the selected error handling, and the selected design technique.

[0103] According to a second aspect of the present invention, a method is provided for generating precoding parameters and / or merging parameters V for the wireless interfaces of a first communication device and a second communication device, respectively. c V k U c,k U k The method is implemented and configured according to the first aspect described above. A first communication device is configured to wirelessly communicate with multiple second communication devices, each also having multiple antennas, via multiple antennas in a MU-MIMO RSMA communication system. The implemented and configured process applies a selected error processing design technique to iteratively optimize the precoding parameters and combining parameters V based on selected target attributes. c V k U c,k U k When executing the implemented and configured process, the method includes receiving an estimated channel coefficient matrix for all communication channels between the first communication device and each of a plurality of second communication devices. and its error statistics (e.g., as the corresponding error matrix) As input to the implemented and configured process, it may also include receiving information about the noise power at the corresponding k-th receiver. The method further includes processing the information with the corresponding estimated channel coefficient matrix according to the implemented and configured procedures. The method also includes determining and / or optimizing precoding parameters and merging parameters V based on the implemented and configured procedures and the selected set of target attributes of the communication channel. c V k U c,k U k And when the iteration termination criterion is met, the optimized precoding parameters and merging parameters V are output. c V k U c,k U k .

[0104] In one or more embodiments, the precoding parameters and merging parameters V are iteratively determined and optimized. c V k U c,k U k Including precoding parameters and merging parameters V c V k U c,k U k Perform iterative convex optimization, or determine the precoding parameters and merging parameters V. c V k U c,k Uk Previously, the estimated channel coefficient matrix The tensor decomposition is performed and resource allocation is performed based on the result thereof. Any iteration steps that can be present can be repeated until a corresponding termination criterion is met.

[0105] In one or more embodiments of performing the tensor decomposition, at least one of the decomposition factors is a set of diagonal matrices, wherein the common signal is identical in all matrices along at least one spatial position of the diagonal, and wherein the plurality of second wireless devices each have an aggregated overlap along the spatial position of the diagonal that is below a predetermined value or are mutually exclusive.

[0106] In one or more embodiments, the error processing comprises averaging the errors of the estimated channel coefficient matrix or estimating a worst-case error of the estimated channel coefficient matrix The estimating the error can further comprise taking into account a noise power

[0107] In one or more embodiments of determining the precoding parameters and the merger parameters V c , V k , U c,k , U k , a worst-case error is determined for each iteration. The worst-case error determined for each iteration is fed back to the iterative convex optimization or the tensor decomposition, respectively, as an input signal for the next iteration.

[0108] The termination criterion can in particular comprise the condition that each of the precoding parameters V c and V k and the merger parameters U c,k and U k determined in the current iteration is sufficiently close to the respective precoding matrix and merger matrix determined in the previous iteration or iterations. The condition of being sufficiently close can be met, for example, when the normalized change of the values in the matrices between the current iteration and the previous iteration is smaller than a predefined threshold, e.g. smaller than 10 -6 . When comparing the change of the matrices between more than one consecutive iteration, a trend can be determined, the extrapolation of which can be used to set the threshold.

[0109] Alternatively, the termination criterion can comprise that a worst-case error estimated using the precoding parameters V c and V k and the merger parameters U c,k and U k determined in the current iteration is smaller than a worst-case error ​is no longer significantly improved, e.g. the improvement of the worst case error relative to one or more previous iterations is less than a predetermined threshold. The condition of no longer being significantly improved can be verified based on a normalized change of the worst case error. When comparing the improvement of the worst case error to the improvement of more than one previous iteration, the worst case error of more than one previous iteration can be averaged or a trend thereof can be determined, an extrapolation of which can set a reference value for comparison.

[0110] The termination criterion of the iteration process, in particular of the channel tensor decomposition, can further comprise obtaining a set of diagonal matrices for at least one of the decomposition factors, wherein at least one spatial position of the common signal along the diagonal is identical in all matrices, and wherein the spatial positions along the diagonal of the plurality of second wireless devices are each or all of them have an aggregated overlap below a predetermined value or are mutually exclusive. The criterion of the aggregated overlap being below a predetermined value can comprise that none of the overlapping spatial positions has a value exceeding the predetermined value.

[0111] Figure 7 An exemplary basic flow chart of a method 200 of generating precoding parameters and combining parameters V c , V k , U c,k , U k according to the second aspect of the application is shown, which is implemented and configured according to the method of the first aspect. In step 202, the estimated channel coefficient matrices and their error statistics are received as input of the implemented and configured process. As further input of the method, the respective noise power at the k-th UE can be received. In step 210, which is conditionally invoked according to the implemented and configured process, the precoding parameters and combining parameters V c , V k , U c,k , U k are initialized, and in step 220, the errors of the respective estimated channel coefficient matrices are processed according to the implemented process. In steps [230...280], using the previously received estimated channel coefficient matrices and their error statistics as input, the parameters V c , V k , U c,k , U k are determined and optimized iteratively according to the implemented process and the set of target properties selected for the communication channel. When the termination criterion is met, check step 290, the iteration is terminated, and in step 292, the optimized precoding parameters and combining parameters V c , V k , Uc,k , U k The dashed connection from step 290 to step 220 indicates an iteration loop for those cases where the error is determined for each iteration. Exemplary embodiments of the method in which the error is determined for each iteration will be further discussed below.

[0112] In the following sections, various specific embodiments of the method according to the second aspect of the application will be presented.

[0113] In a first specific embodiment, the method according to the first aspect of the application has led to the implementation and configuration of the method according to the second aspect of the application, wherein assuming an average CSI error, the precoding parameters V c and V k and the combiner parameters U c,k and U k are iteratively optimized by convex optimization. The optimization can be implemented as a block coordinate descent procedure, among others, but other optimization methods can also be used.

[0114] Figure 8 A block diagram of a corresponding first specific exemplary precoder and combiner matrix BF design block implemented according to the method 100 of the first aspect of the application is shown. The BF design block applies iterative convex optimization and assumes an average CSI error. The input matrices and are provided to blocks 304a-i. The blocks 304a-i initialize the precoding parameters V and the combiner parameters U, respectively, and provide these parameters as well as the estimated channel coefficient matrices and the corresponding average error to the optimizer blocks 304c-i. In the optimizer blocks 304c-i, the precoding parameters V and the combiner parameters U are iteratively optimized, as indicated by the arrows back and forth between the update blocks for V and U, and the respective updated parameters or parameter sets are provided to the respective other update block. The optimization is performed according to the objective selected in block 304b, i.e. maximizing the sum transmission rate, maximizing the minimum transmission rate, or minimizing the transmit power while achieving a guaranteed transmission rate. The optimization can implement a convex optimization algorithm, for example. Prior to the optimization, a block coordinate descent algorithm can be applied to decouple the optimization variables V c , V k of the precoder and the optimization variables U c,k , U k of the combiner. Further, if necessary, a proper convexification of the function algebraically describing the optimization object can be performed prior to the actual optimization. The optimization is terminated when a termination criterion is met, and the optimized precoding parameters V and combiner parameters U are output. Note that information about the noise power at the respective k-th receiver can also be provided as input to the BF design block (not shown in the figure).

[0115] Figure 9 According to a second aspect of the invention, precoding parameters and combining parameters V are generated for the wireless interfaces of a first communication device (300) and a second communication device (400), respectively. c V k U c,k U k The flowchart of the first specific embodiment of method 200 is shown. The method of the first specific embodiment aims to optimize the precoding parameters and merging parameters V of the k-th UE while maximizing the total rate, assuming the average CSI error. c V k U c,k U k This embodiment of method 200 may include receiving, in step 202, the... The estimated and shared CSI representation The corresponding matrix of error statistics and variance Then perform the following steps:

[0116] 210-Use Initialize precoding parameters V c and V k and the merger matrix U c,k and U k ,

[0117] 212- Total speed R total Approximately because The function representing the estimate and sharing of CSI.

[0118] 214 - Using auxiliary variables and slack variables to convexize the approximate function.

[0119] 220-Error Handling

[0120] 232-Optimize precoding parameters V c and V k :

[0121] 234 - Solving for V with fixed auxiliary variables c and V k The convexity approximation problem, and

[0122] 236- at a fixed V c and V k Update auxiliary variables in the following cases.

[0123] 238 - Check precoding parameter V c and V kCheck if it has converged. If it has not converged, repeat steps 234 and 236.

[0124] otherwise

[0125] 240-Optimize merger parameters U c,k and U k :

[0126] 242- Solving U with fixed auxiliary variables c,k and U k The convexity approximation problem, and

[0127] 244-in fixed U c,k and U k Update auxiliary variables in the following cases.

[0128] 246 - Check merger parameters U c,k and U k Check if it has converged. If it has not converged, repeat steps 242 and 246.

[0129] otherwise

[0130] 280 - Check if the first loop termination criterion is met. If so, output V in step 292. c V k U c,k and U k ,

[0131] Otherwise, repeat steps [230…246 and 280].

[0132] Exemplary termination criteria may include the following condition: converged precoding parameters V c and V k and the convergent merger parameter U c,k and U k Each of these parameters is sufficiently close to the corresponding previously converged precoding parameter and merger parameter.

[0133] In a second specific embodiment, the method according to the first aspect of the invention has led to the implementation and configuration of the method according to the second aspect of the invention, wherein, assuming the worst-case CSI error, the precoding parameter V is iteratively optimized through convex optimization. c and V k and merger parameter U c,k and U k The optimization can be implemented again as a block coordinate descent process, but as in the first specific embodiment, other optimization methods can also be used.

[0134] Figure 10A corresponding second specific exemplary precoder and combiner matrix BF design block applying iterative convex optimization under worst case CSI error conditions is shown. In this scenario, the input matrices and are provided to an initialization block 303 which uses the input (i.e. the CSI and the CSI error statistics) to determine initial precoding parameters V and combiner parameters U, respectively, which are provided to blocks 304a-ii. The blocks 304a-ii estimate the channel coefficient matrix and the corresponding worst case channel coefficients (i.e. the most harmful CSI imperfection) and provide these to an optimizer block 304c-i. The most harmful CSI imperfection can be expressed by minimizing the achievable rate. The worst case channel coefficients of the channel coefficient matrix are determined as

[0135]

[0136] with the constraint

[0137] In the optimizer block 304c-i, the precoding parameters V and the combiner parameters U are iteratively optimized, indicated by the arrows back and forth between the update blocks for V and U, and the respective updated parameters or parameter sets are provided to the respective other update block. The optimization is performed according to the objective selected in block 304b, i.e. maximizing the sum transmission rate, maximizing the minimum transmission rate, or minimizing the transmit power while achieving a guaranteed transmission rate. The optimization can for example implement a convex optimization algorithm. Prior to the optimization, a block coordinate descent algorithm can be applied to decouple the optimization variables V c , V k of the precoder c,k , U k of the combiner. Further, if necessary, prior to the actual optimization, a proper convexification of the function algebraically describing the optimization object can be performed. The optimization is terminated when a termination criterion is met. As in the first specific exemplary precoder and combiner matrix BF design block discussed before, information about the noise power at the respective k-th receiver can also be provided as input to the BF design block (not shown in the figure).

[0138] Figure 11 A corresponding second specific exemplary precoder and combiner matrix BF design block applying iterative convex optimization under worst case CSI error conditions is shown. In this scenario, the input matrices c , V k , U c,k , U kA flowchart of a second specific embodiment of method 200. Method 200 may include receiving, in step 202, a signal from... The estimated and shared CSI representation The corresponding matrix of error statistics and variance Then perform the following steps:

[0139] 210-Use H = {0} Initialize the precoding parameter V c and V k and merger parameter U c,k and U k ,

[0140] 254 - Through iterative and alternating optimization, targeting For all cases of H, optimize V. c V k U c,k and U k ,

[0141] 256- at a fixed V c V k U c,k and U k In this case, the worst-case CSI error is estimated by minimizing the total rate.

[0142] 258 - Update the set of possible worst-case CSI errors

[0143] 280 - Check if the first loop termination criterion is met. If so, output V in step 292. c V k U c,k and U k ,

[0144] Otherwise, repeat steps [254…258 and 280].

[0145] Optimization step 254 can be implemented by optimizing the precoding parameters and combining parameters V of the k-th UE based on any of the selected target attributes of the communication connection (including maximizing the total rate, maximizing the minimum rate, or minimizing power under rate guarantee). c V k U c,k U k Similar to the first specific embodiment described above, optimization can, for example, be implemented using a convex optimization algorithm. Prior to optimization, a block coordinate descent algorithm can be applied to decouple the pre-encoder's optimization variable V. c V k And the optimization variable U of the mergerc,k U k Furthermore, if necessary, appropriate convexification of the function algebraically describing the optimization object can be performed before actual optimization. Optimization terminates when the termination criterion is met, and the optimized precoding matrix V and merger matrix U are output.

[0146] Use the latest precoding and merging parameters V c V k U c,k U k Determining the worst-case CSI error in each iteration ensures that the best available precoding and merging parameters V are eventually found. c V k U c,k U k .

[0147] An exemplary first loop termination criterion may include the following condition: precoding parameter V c and V k and merger parameter U c,k and U k Each of these parameters is sufficiently close to the corresponding previously converged precoding parameter and merger parameter. Alternative exemplary termination criteria may include the following condition: using the precoding parameter V determined in the current iteration. c V k and merger parameter U c U k worst-case error in estimation The worst-case error relative to the one or more previous iterations The improvement is less than the predetermined threshold.

[0148] In a third specific embodiment, the method according to the first aspect of the invention has led to the implementation and configuration of the method according to the second aspect of the invention, wherein, assuming the average CSI error, the precoding parameters V are iteratively and jointly optimized through channel tensor decomposition. c and V k and merger parameter U c,k and U k .

[0149] This specific embodiment of determining the precoding parameter V and the combiner parameter U, respectively, is derived from the idea of a multilinear generalized singular value decomposition (ML-GSVD), e.g., as proposed by L. Khamidullina, A. L. F. de Almeida, and M. Haardt in “Multilinear generalized singular value decomposition (ML-GSVD) with application to coordinated beamforming in multi-user MIMO systems”, Proc. IEEE ICASSP, Barcelona, Spain, 2020, pp. 4587-4591.

[0150] Applying the general principle of ML-GSVD, the matrix Hkof channel coefficients of the k-th UE k can be decomposed into the product of matrices B k , C k , and A T , as shown in Figure 12 . Figure 12 It is shown that there are multiple so-called “slices”, each of which dimensionally exemplarily represents the channel coefficient matrix of one of the k UEs and the decomposition factors. A T is a square matrix similar to the right singular vectors of a conventional singular value decomposition (SVD), except for the fundamental difference that it is common to all slices of H k . Therefore, it is only represented once. B k is a matrix of matrix units for each individual slice, similar to the left singular vectors in a conventional SVD. C k is a diagonal matrix for each individual slice, representing the channel space occupied by each UE. The channel space represented by the diagonal matrix C k may include information about the size of the antenna.

[0151] The most important aspect in designing BF in a spatial division multiple access (SDMA) system is the structure of the decomposed channel, especially the diagonal matrix C k , and the exclusive dimension allocation (i.e., spatial allocation). However, the known ML-GSVD approach is not designed to facilitate a common interface matrix A TThe separation of the subspaces of A designed this drawback obviously cannot facilitate the construction of the TX beamformer. This problem has been solved by K. Ando, H. Iimori, G. T. F. de Abreu and K. Ishibashi in “User-Heterogeneous Cell-Free Massive MIMO Downlink and Uplink Beamforming via Tensor Decomposition”, IEEE Communications Society Open Journal, vol. 3, pp. 740-758, 2022, where a new tensor decomposition is proposed. The new tensor decomposition facilitates the orthogonalization of the subspaces in A by imposing sparsity in the process in matrix C, such that all subspaces in A are fully separated, thus enabling interference-free BF in underloaded scenarios. Figure 13 The corresponding decomposition is exemplarily shown in Fig. 2, where for each of the k channels, the non-zero positions along the diagonal of C k are ordered or grouped such that the superposition of C k does not lead to an overlap of these non-zero positions.

[0152] While the orthogonalized subspaces proposed in the state-of-the-art approaches are generally beneficial for beamforming in MU-MIMO environments, RSMA has specific requirements, in particular due to the common message part, which have not been properly addressed.

[0153] Therefore, the present invention proposes also a new tensor decomposition, which divides the channel space into respective spaces for the “common message” and the “private message”. With this new decomposition, the channel structure can be exhibited as shown in Fig. 3. Figure 14 and Figure 15 In the figure, the channel space of the common message is identified by the positions in the matrix filled with the diagonal hash pattern, while the channel space of the private message is identified by the solid black fill. Figure 15 An enlarged representation of the matrix C k is shown, where the channel space of the common message and the private message is indicated. The diagonal matrix C k now has an overlapping common space for the common message, which is possible since the common signal is common to all UEs. However, the private message has mutually exclusive spaces to reduce interference to the private messages of the respective other UEs.

[0154] This further separation finally takes into account the specific requirements of the RSMA system, i.e. the separation of the common message part and the private message part, and allows to jointly determine the precoding parameters and the combiner parameters V used for beamforming at the transmitter and the receiver, respectively c k c,k k such that interference-free RSMA communication is achieved.

[0155] Figure 16 A block diagram of a corresponding third specific exemplary precoder and combiner BF design block applying channel tensor decomposition and assuming an average CSI error is shown. The input matrices and are provided to block 304a-i. Block 304a-i determines the average CSI error and provides the estimated channel coefficient matrix to the tensor decomposition block 304c-ii. The tensor decomposition block 304c-ii performs the tensor decomposition by iteratively updating each factor B, C, A in the respective blocks 264, 266 and 268 after the initialization in block 260. The iteration of the channel tensor decomposition terminates when the channel tensor does not converge anymore and B, C, A are output. Using the results of the decomposition, the target selected in block 304b previously, i.e. maximizing the sum transmission rate, maximizing the minimum transmission rate or minimizing the transmit power while achieving a guaranteed transmission rate, is implemented with resource allocation and actual BF design. Like in the first and second specific exemplary precoder and combiner matrix BF design blocks discussed before, information about the noise power at the respective k-th receiver can also be provided as input to the BF design block (not shown in the figure).

[0156] Figure 17 A flow chart of a corresponding third concrete embodiment of a method 200 for generating precoding parameters and combiner parameters V c , V k , U c,k , U k for the wireless interfaces of the first communication device (300) and the second communication device (US), respectively, according to the second aspect of the present application is shown. This embodiment of the method 200 can comprise the following steps after receiving in step 202 the estimated and shared CSI represented by , the corresponding matrix representing the error statistics of and the variance :

[0157] 210 - initialization of variables for decomposing the channel tensor H = [H1,..., H k ]​​​

[0158] 262 - Decompose the channel tensor into three factors A, C and B k ,

[0159] 264 - Update B with A and C fixed k ,

[0160] 266 - Update C with A and B fixed k ,

[0161] 268 - Update A with C and B fixed k ,

[0162] 280 - Check if a first loop termination criterion is met, otherwise repeat steps [264...268],

[0163] 270 - Perform resource allocation by computing the transmit power and stream allocation for all UEs, 272 - Compute V c , V k , U c,k and U k ,

[0164] 292 - Output V c , V k , U c,k and U k .

[0165] An exemplary first loop termination criterion for the channel tensor decomposition can comprise obtaining a set of diagonal matrices for at least one of the decomposition factors, wherein the common signal along at least one spatial location of the diagonal is identical in all slices, and wherein the private signals of the plurality of second wireless devices along the spatial locations of the diagonal each or all have an aggregated overlap below a predetermined value or are mutually exclusive.

[0166] The selected error processing (here averaging the errors of the estimated channel coefficient matrix ) is performed prior to the channel tensor decomposition and is part of the initialization step 210. The channel tensor decomposition takes into account the result of the processing of the errors of the estimated channel coefficient matrix and can also take into account the selected target properties of the communication connections with the second communication devices.

[0167] The resource allocation step 270 can be implemented to optimize the precoding parameters and the combining parameters V c , V k , U c,k , U k. To this end, the resource allocation can comprise allocating resources to the plurality of antennas of the first communication device according to at least one of the factors of decomposition (e.g., the diagonal matrix C k .

[0168] In a fourth particular embodiment, the method according to the first aspect of the application has led to the implementation and configuration of a method according to the second aspect of the application, wherein assuming worst-case CSI error estimates, the precoding parameters V c and V k as well as the combiner parameters U c,k and U k are iteratively and jointly optimized by channel tensor decomposition. Considering the estimation of worst-case CSI errors, using the same concept of tensor decomposition presented in the third particular embodiment, more robust precoding and combining parameters V c , V k , U c,k , U k can be obtained.

[0169] Figure 18 A block diagram of a corresponding fourth particular exemplary precoder and combiner BF design block applying tensor decomposition under worst-case CSI error conditions is shown. The estimated channel coefficient matrix is provided to a tensor decomposition block 304c-ii, which performs tensor decomposition by iteratively updating each factor B, C, A in respective blocks 264, 266 and 268 after initialization in block 260. When the channel tensor no longer converges, the iterative channel tensor decomposition terminates and B, C, A are output. Using the results of the decomposition, the resource allocation and the BF design are adopted according to the target selected in block 304b previously, i.e., maximizing the sum transmission rate, maximizing the minimum transmission rate, or minimizing the transmit power while achieving a guaranteed transmission rate. The results of the BF design, i.e., the precoding parameters V and the combiner parameters U, are provided together with the CSI error matrix to block 304a-ii for estimating the corresponding worst-case channel coefficients (i.e., the most harmful CSI imperfection), which are fed back to the iterative tensor decomposition. The most harmful CSI imperfection can be represented by minimizing the achievable rate, as further discussed above in connection with the second particular embodiment. When a termination criterion is met, the iterative optimization is terminated. Like in the first to third particular exemplary precoder and combiner matrix BF design blocks discussed earlier, information about the noise power at the respective k-th receiver can also be provided as input to the BF design block (not shown in the figure).

[0170] Initial decomposition and worst-case error estimation can be based on the input signal, which estimates the CSI and its error statistics. Once the initial decomposition is complete, resource allocation can be performed, and the precoding parameters V and merger parameters U can be calculated separately, taking into account the objectives of resource allocation.

[0171] Exemplary termination criteria may include the following conditions: precoding parameter V c and V k and merger parameter U c,k and U k Each of these is sufficiently close to the corresponding previously converged precoding matrix and merger matrix. Alternative exemplary termination criteria may include the following condition: using the precoding parameters V determined in the current iteration. c V k and merger parameter U c,k U k worst-case error in estimation The worst-case error relative to the one or more previous iterations The improvement is less than the predetermined threshold.

[0172] Figure 19 According to a second aspect of the invention, precoding parameters and combining parameters V are generated for the wireless interfaces of a first communication device (300) and a second communication device (US), respectively. c V k U c,k U k A flowchart of a fourth specific embodiment of method 200. This embodiment of method 200 may include receiving, in step 202, a signal from... The estimated and shared CSI representation The corresponding matrix of error statistics and variance Then perform the following steps:

[0173] 210 - Initialization for decomposing the channel tensor H = [H1, ..., H k The variable ] is used H = {0} Initialize the precoding parameter V c and V k and merger parameter U c,k and U k And determine TM H's initial worst-case scenario,

[0174] 262-Targeting TM In the worst case, H decomposes the channel tensor into three factors A, C, and B. k ,

[0175] 264 - update B with A and C fixed k ,

[0176] 266 - update C with A and B k fixed,

[0177] 268 - update A with C and B k fixed,

[0178] 280 - check if first loop termination criterion is met, otherwise repeat steps [264...268, 280],

[0179] 270 - perform resource allocation by computing the transmit power and flow allocation for all UEs, 272 - compute V c , V k , U c,k and U k ,

[0180] 274 - estimate worst-case CSI error matrix c , V k , U c,k and U k by total rate minimization with V c , V k , U c,k and U k fixed,

[0181] 276 - update the set of possible worst-case CSI errors

[0182] 290 - check if second loop termination criterion is met, otherwise update worst-case CSI errors in step 258 and repeat steps [264...268, 280, 270, 272, 256 and 290],

[0183] 292 - output V c , V k , U c,k and U k .

[0184] As in the third specific embodiment discussed above, an exemplary first loop termination criterion for the channel tensor decomposition can comprise obtaining a set of diagonal matrices for at least one of the decomposition factors, wherein the common signal along at least one spatial location of the diagonal is identical in all slices, and wherein the private signals of the plurality of second wireless devices along the spatial locations of the diagonal each or all have an aggregated overlap below a predetermined value or are mutually exclusive.

[0185] An exemplary second loop termination criterion can include the condition that each of the precoding parameters V c and V k and the combiner parameter U c,k and U k is sufficiently close to the respective previously converged precoding and combiner parameters.

[0186] According to a third aspect of the present application, a method of operating a first wireless communication device is presented. The first communication device (e.g. the base station 300) is configured for wireless communication with a plurality of second communication devices 400 in a MU-MIMO RSMA communication system, i.e. each of the first and second wireless communication devices 300, 400 has a plurality of antennas 306, 402. Figure 21 The method exemplarily shown in Fig. 5 comprises, in step 504a, performing the method according to the first aspect of the present application and the generating procedure according to the second aspect of the present application. Alternatively, in step 504b, the method can comprise receiving the precoding parameters and the combiner parameter V c , V k , U c,k , U k determined according to the second aspect of the present application. The alternative step is indicated by the dashed outline. The method further comprises, in step 506, providing at least the precoding parameters and the combiner parameter V c , V k , U c,k , U k to a precoder 302 of the first communication device 300 and to at least each of the plurality of second wireless devices 400 to which a message is to be transmitted. Further, the method comprises, in the first communication device 300, in step 508, splitting a message to be transmitted to one or more of the plurality of second wireless communication devices 400 into a respective common part and a private part, and providing the split message to the precoder 302, and in step 510, precoding each of the private part and the common part to obtain a transmit signal for each of the plurality of antennas of the first wireless communication device BS. Finally, the method comprises, in step 512, transmitting the precoded transmit signals.

[0187] In one or more embodiments, the method according to the third aspect of the present application further comprises, in step 502, receiving, as input for performing step 504a, an estimated channel coefficient matrix and error statistics thereof for all communication channels between the first communication device 300 and the second communication devices UEs. The receiving can comprise determining the estimated channel coefficient matrix And its error statistics, or receive the information from the corresponding second wireless communication device 400.

[0188] According to a fourth aspect of the invention, a method of operating a second wireless communication device is presented. The second communication device (e.g., user equipment 400) is configured to wirelessly communicate with a first communication device (e.g., base station 300) in a MU-MIMO RSMA communication system, wherein each of the first wireless communication device 300 and the second wireless communication device 400 has a plurality of antennas 306, 402. Figure 22 The method illustrated in the example includes, in step 606, receiving at least precoding parameters and merging parameters V from the first communication device 300. c V k U c,k U k ; and in step 608, a signal is received from the first communication device 300 at a plurality of antennas 402, the signal including a common signal portion s. c and private signal part s k And based on the same precoding parameters and merging parameters V previously received from the first communication device 300 c V k U c,k U k Precoding is performed. The method further includes, in step 610, using previously received precoding parameters and merging parameters (V...). c V k U c,k U k ), for the corresponding common signal portion s received at multiple antennas 402 c and private signal part s k Perform the merging. Merging step 610 generates the merged common signal part s. c and the merged private signal part s k These are provided to detector 406 in step 612. In step 614, detector 406 estimates the transmitted common signal portion s, respectively. c and private signal part s k And in step 622, an estimated signal is provided at the output.

[0189] In one or more embodiments of the method according to the fourth aspect of the invention, wherein the channel coefficient matrix The method further includes: estimating and transmitting precoding parameters and combining parameters V from the second wireless communication device to the first wireless communication device in step 606. c V k U c,k Uk Previously, in step 602, a channel coefficient matrix of a communication channel between the second wireless communication device and the first wireless communication device is estimated Then, in step 604, the estimated channel coefficient matrix is transmitted to the first wireless communication device.

[0190] In step 614, the transmitted common signal s c and the private signal s k may include detecting the common signal part s k from the received signal y c and obtaining the private signal part using the known information of the common signal part s c .

[0191] In one or more embodiments, according to the method of the fourth aspect of the application, the estimating step 614 in the detector 406 includes decoding the common signal part s c in step 616 in the first decoder 408. In step 618, interference cancellation is performed using the decoded common signal part s c and the combined common signal part s c and the combined private signal part s k obtained from the combining step 610 as inputs. The output signal from the interference cancellation step 618 is provided to the second decoder 410 which decodes the private signal part s k from the signal obtained by interference cancellation in step 620.

[0192] In the various embodiments presented above, estimating the channel coefficient matrix may include any known channel estimation method, including but not limited to channel estimation based on basis expansion modeling, etc.

[0193] According to a fifth aspect of the application, a wireless communication device (e.g., a base station or a user equipment) comprises one or more microprocessors, volatile and non-volatile memory, and wireless interface circuitry configured for transmitting and / or receiving electromagnetic signals via a plurality of antennas. The various elements are communicatively connected via one or more data or signal lines or buses. The non-volatile memory stores computer program instructions which, when executed by the microprocessor, configure the wireless device to perform one or more of the methods according to the first, second, third, or fourth aspects of the application as presented above.

[0194] The method described above can be represented by computer program instructions. Thus, a computer program product comprises computer program instructions which, when executed by a microprocessor of a transmitter, cause the microprocessor to perform the method according to the first, second or third aspect of the application as presented above and to control the hardware components of the transmitter of the RSMA MU-MIMO communication system accordingly. These computer program instructions, when executed by a microprocessor of a receiver, cause the microprocessor to perform the method according to the fourth aspect of the application as presented above and to control the hardware components of the receiver of the RSMA MU-MIMO communication system accordingly.

[0195] These computer program instructions can be stored or retrievably transmitted on a computer readable medium or data carrier. The medium or data carrier can be embodied physically, for example in the form of a hard disk, solid state disk, flash memory device, etc. However, the medium or data carrier can also comprise a modulated electromagnetic, electrical or optical signal which is received by a computer via a corresponding receiver and transferred to and stored in the memory of the computer.

[0196] The present application advantageously allows jointly determining the precoding matrix in the transmitter and the combiner matrix, taking into account the non-perfectly known CSI and the specific requirements of RSMA. This leads to an enhanced robustness of the communication, an improved IC at the receiver and ultimately an improved symbol detection, without fundamentally changing the structure of the communication system. The adaptability of the BF design provides multiple ways to ensure resilient, robust and reliable communication.

[0197] The proposed method can advantageously be used in a general wireless communication system using RSMA in the downlink, in particular in a system with heterogeneous UEs with different numbers of antennas, and generally in any such system where the UEs do not have perfect SIC. However, since the RSMA model reconciles known traditional OMA and NOMA access methods, the proposed method is applicable to any traditional downlink wireless communication system including OMA or NOMA.

[0198] The proposed method can advantageously be used for highly mobile devices, such as vehicles, trains, airplanes, etc. BRIEF DESCRIPTION OF DRAWINGS

[0199] The figures of the accompanying drawings serve to illustrate in detail various aspects of the present application. In the drawings:

[0200] Figure 1 The main components of a transmitter (for example in a base station 300) and a receiver (for example in a UE) configured to perform the method according to the present application are shown respectively,

[0201] Figure 2 a channel diagram showing the messages exchanged in the DL direction between a BS and a target UE in a TDD communication system and the corresponding processing invoked at the respective ends,

[0202] Figure 3 a channel diagram showing the messages exchanged in the DL direction between a BS and a target UE in a FDD communication system and the corresponding processing invoked at the respective ends,

[0203] Figure 4 an exemplary simplified block diagram of a block in a BS that processes the BF design and outputs the precoding matrix V and the combiner matrix U needed for beamforming in the BS and combining the signals received at the M≥1 antennas in the UE,

[0204] Figure 5 a more detailed view of the block of the BF design shown, Figure 4

[0205] Figure 6 an exemplary flowchart of the method according to the first aspect of the application,

[0206] Figure 7 an exemplary basic flowchart of the method according to the second aspect of the application for generating the precoding parameters and the combining parameters V c , V k , U c,k , U k ,

[0207] Figure 8 a block diagram of a first specific exemplary precoder and combiner BF design block that applies convex optimization and assumes average CSI error,

[0208] Figure 9 a flowchart of the corresponding first concrete embodiment of the method of the application for optimizing the precoder V for the kth UE, aiming at maximizing the sum rate,

[0209] Figure 10 a block diagram of a second specific exemplary precoder and combiner BF design block that applies convex optimization under worst-case CSI error conditions,

[0210] Figure 11 a flowchart of the corresponding second concrete embodiment of the method of the application for optimizing the precoder V for the kth UE, aiming at maximizing the sum rate,

[0211] Figure 12 a general representation of the matrices and their factors after being submitted to the ML-GSVD operation,

[0212] ​Figure 13 representations of the matrices and their factors after being submitted to an ML-GSVD operation tailored for separate subspaces,

[0213] Figure 14 representations of the matrices and their factors after being submitted to an ML-GSVD operation tailored for RSMA with separate subspaces,

[0214] Figure 15 shows a block diagram of a third specific exemplary precoder and combiner BF design block applying tensor decomposition and assuming average CSI error, Figure 14 a magnified representation of the matrix C k a magnified representation of the matrix C

[0215] Figure 16 shows a block diagram of a third specific exemplary precoder and combiner BF design block applying tensor decomposition and assuming average CSI error,

[0216] Figure 17 shows a flowchart of a corresponding third particular embodiment of the inventive method for optimizing the precoder V and the combiner U for the k-th UE aiming at maximizing the total rate,

[0217] Figure 18 shows a block diagram of a fourth specific exemplary precoder and combiner BF design block applying tensor decomposition under worst-case CSI error conditions,

[0218] Figure 19 shows a flowchart of a corresponding fourth particular embodiment of the inventive method for optimizing the precoder V and the combiner U for the k-th UE aiming at maximizing the total rate,

[0219] Figure 20 shows a block diagram of a transmitter or receiver configured for performing the method according to the invention,

[0220] Figure 21 shows an exemplary flowchart of a method of operating a first wireless communication device wirelessly connected to a plurality of second wireless communication devices in a MU-MIMO RSMA communication system, and

[0221] Figure 22 shows an exemplary flowchart of a method of operating a second wireless communication device wirelessly connected to a first wireless communication device in a MU-MIMO RSMA communication system.

[0222] In the drawings, identical or similar elements can be denoted using the same reference numerals. DETAILED DESCRIPTION

[0223] Figures 1 to 19 have been further described hereinabove and will not be discussed again.

[0224] Figure 20 An exemplary block diagram of a transmitter 300 or receiver 400 according to embodiments of the fifth aspect of the application is shown, respectively. The transmitter 300 or receiver 400 comprises a microprocessor 350, a volatile memory 352, a non-volatile memory 354, a wireless interface circuitry 356 configured for communication with a receiver or transmitter, respectively, by transmitting and / or receiving electromagnetic signals via a plurality of antennas 306, 402. The above-mentioned elements are communicatively connected via one or more signal or data connections or buses 358. The non-volatile memory 354 stores computer program instructions which, when executed by the microprocessor 350, cause the transmitter 300 or receiver 400 to perform the method according to the first, second or third aspect of the application as presented herein.

[0225] Figure 21 An exemplary flow chart of a method of operating a first wireless communication device 300 according to the fifth aspect of the application according to the third aspect of the application is shown. The first wireless communication device 300 is wirelessly connected to a plurality of second wireless communication devices 400 in a MU-MIMO RSMA communication system. In step 502, the estimated channel coefficient matrix H and its error statistics of all communication channels between the first communication device 300 and the second communication devices 400 are received and provided as input to step 504a. In step 504a, the method of determining the process for generating precoding parameters and combining parameters according to the first aspect of the application and the generation method according to the second aspect are performed. Alternatively, the precoding parameters and combining parameters determined according to the method according to the second aspect of the application are received in step 504b. In step 506, at least the precoding parameters and combining parameters are provided to the precoder 302 of the first communication device 300 and to at least each of the plurality of second wireless devices 400 to which a message is to be transmitted. In step 508, a message to be transmitted to one or more of the plurality of second wireless communication devices 400 is split into a respective common signal part s c and a private signal part s k and the split message is provided to the precoder 302. In step 510, the precoder 302 precodes each of the private signal part s k and the common signal part s c to obtain a transmit signal for each of the plurality of antennas of the first wireless communication device 300. Finally, the precoded transmit signals are transmitted in step 512.

[0226] Figure 22A flow chart illustrating a method 500 of operating a second wireless communication device 400 wirelessly connected to a first wireless communication device 300 in a MU-MIMO RSMA communication system is shown. Depending on which wireless device estimates the channel coefficient matrix H Steps 602 and 604 can optionally be performed, wherein the channel coefficient matrix H is estimated and transmitted to the first wireless communication device 300. The method comprises receiving in step 606 at least the precoding parameters and combining parameters V c , V k , U c,k , U k from the first communication device 300. The method further comprises receiving in step 608 a signal y k from the first communication device 300. The signal comprises a common signal part s c and a private signal part s k and precodes it according to the same precoding parameters and combining parameters V c , V k , U c,k , U k previously received from the first communication device 300. The method still further comprises combining in step 610 the respective common signal parts and private signal parts received at the plurality of antennas 402 using the precoding parameters and combining parameters V c , V k , U c,k , U k previously received for precoding to obtain a combined common signal part s c and a combined private signal part s k . In step 612, the combined common signal part s c and the combined private signal part s k are provided to the detector 406 for estimating in step 614 the transmitted common signal part s c and the private signal part s k , which estimated signals are provided at the output in step 622.

[0227] The estimating step 614 can comprise decoding in step 616 the common signal part s c in the first decoder 408, using the decoded common signal part s c in step 618 and the combined common signal part s c and the combined private signal part s kperforming interference cancellation as input, and decoding the private signal part s from the signal obtained by interference cancellation 618 in step 620 k .

[0228] List of reference signs (part of the description)

[0229]

[0230]

Claims

1. A method for determining precoding parameters and combining parameters (V) for generating wireless interfaces for a first communication device (300) and a second communication device (400), respectively. c V k U c,k U k A method (100) for the process of wirelessly communicating with a plurality of second communication devices (400) in a MU-MIMO RSMA communication system, the method comprising: for all communication channels with all of the plurality of second communication devices (400): Receive (110) selection inputs, which are used to: select the target attributes of these communication connections, and the estimated channel coefficient matrix for processing all communication channels. The associated error processes and design techniques, Based on the selection input, one target attribute set is selected from multiple target attribute sets of these communication connections (120). Based on this selection, the input is used to process these estimated channel coefficient matrices. Select one process (130) from the multiple processes of the associated corresponding error. Based on this selection input, the parameters used to determine these precoding parameters and / or merging parameters (V) are obtained. c V k U c,k U k Choose one design technology from a plurality of design technologies, and Based on the selected target attributes, selected error handling procedures, and selected design techniques of these communication connections, implement and configure (150) the methods for generating these precoded parameters and / or merged parameters (V). c V k U c,k U k The process of generating the estimated channel coefficient matrices is configured to use at least these matrices. The output from this error processing is used as input.

2. The method (100) of claim 1, further comprising calling the method at least in one of the following instances, including but not limited to: at predetermined intervals, when a new second wireless communication device (400) joins a plurality of second wireless communication devices (400) connected to the first communication device (300), when one or more second wireless communication devices (400) leave a plurality of second wireless communication devices (400) connected to the first communication device (300), when the channel coefficient of at least one of the plurality of second wireless communication devices (400) connected to the first communication device (300) changes, and / or when the content and / or type of a data message to be transmitted to one or more of the plurality of second wireless communication devices (400) connected to the first communication device (300) changes.

3. The method (100) as claimed in claim 1 or 2, wherein, The optional objective properties of these communication connections include maximizing total rate, maximizing minimum rate, or minimizing power while ensuring rate. Optional procedures for error handling include averaging the CSI error or estimating the worst-case CSI error, and Optional design techniques include iterative convex optimization or tensor decomposition.

4. The method (100) as claimed in claim 1, 2, or 3, wherein, Implementing this process includes providing computer program instructions and / or data representing a target set of attributes of the communication connection, used to process these estimated channel coefficient matrices. The process of determining the error, and the methods used to determine these precoding parameters and / or merging parameters (V c V k U c,k U k The algorithm is implemented by a computer.

5. A method for generating precoding parameters and combining parameters (V) for the wireless interfaces of a first communication device (300) and a second communication device (400), respectively. c V k U c,k U k The method (200) of the first wireless communication device (300) configured to wirelessly communicate with a plurality of second communication devices (400) in a MU-MIMORSMA communication system, the method being implemented and configured according to any one of claims 1 to 4, and comprising: Receive (202) the estimated channel coefficient matrix of all communication channels between the first communication device (300) and these second communication devices (400). Its error statistics serve as input to the implemented and configured processes. Based on the implemented process (220), the corresponding estimated channel coefficient matrix is ​​obtained. The error, Using the previously received estimated channel coefficient matrix Using its error statistics as input, the precoding parameters and combining parameters (V) are determined and / or optimized based on the implemented process and the set of target attributes selected for these communication channels. c V k U c,k U k ),as well as The precoding parameters and merging parameters (V) determined and / or optimized by output (292) c V k U c,k U k ).

6. The method (200) as claimed in claim 5, wherein, Determine and / or optimize (230) precoding parameters and merging parameters (V c V k U c,k U k )include For precoding parameters and merging parameters (V c V k U c,k U k Perform iterative convex optimization, or In determining these precoding parameters and merging parameters (V c V k U c,k U k Before that, the estimated channel coefficient matrix was... Perform tensor decomposition to decompose into factors (A, C, B) k And perform resource allocation based on the results, and Repeat the corresponding iterative or decomposition steps until the termination criterion is met.

7. The method (200) as claimed in claim 6, wherein, When performing tensor decomposition, at least one of the decomposition factors is a set of diagonal matrices (C0). k ), wherein at least one spatial location along the diagonal for the common signal is identical in all matrices, and wherein the spatial locations along the diagonal for the plurality of second wireless devices each have aggregate overlap or mutual exclusion below a predetermined value.

8. The method (200) according to any one of claims 5 to 7, wherein, Error processing (206) includes processing the estimated channel coefficient matrix. The error is averaged, or the estimated channel coefficient matrix is ​​estimated. Worst case error 9. The method (200) as claimed in claim 8, wherein, When these precoding parameters and merging parameters (V) are determined c V k U c,k U k Worst-case error At that time, the worst-case error determined for each iteration These signals are fed back to either iterative convex optimization or tensor decomposition as input signals for the next iteration.

10. The method (200) according to any one of claims 5 to 9, wherein, The termination criteria include the following conditions: these precoding parameters and merging parameters (V c V k U c,k U k The value of ) changes less than a predefined threshold between the current iteration and the previous iteration, or uses the precoding parameters (V) determined in the current iteration. c V k ) and merger parameters (U c,k U k The worst-case error in the estimate The worst-case error relative to the one or more previous iterations The improvement is less than the predetermined threshold.

11. A method (500) for operating a first wireless communication device (300) wirelessly connected to a plurality of second wireless communication devices (400) in a MU-MIMO RSMA communication system, the method comprising: Perform (504a) the method (100) according to one or more of claims 1 to 4 and the generation process (200) according to one or more of claims 5 to 10, or receive (504b) the precoding parameters and merging parameters (V) determined according to one or more of claims 5 to 10. c V k U c,k U k ), At least these precoding parameters and merging parameters (V) c V k U c,k U k The precoder (302) is provided (506) to the first communication device (300) and is provided at least to each of the plurality of second wireless devices (400) to which a message is to be transmitted. The message to be transmitted to one or more of the plurality of second wireless communication devices (400) is split (508) into corresponding public and private parts, and the split message is provided to the precoder (302). Each of these private portions and these public portions is pre-coded (510) to obtain the transmitted signal of each of the plurality of antennas of the first wireless communication device (300), and Transmit (512) pre-coded transmit signal.

12. The method (500) of claim 11, further comprising: Receive (502) the estimated channel coefficient matrix of all communication channels between the first communication device (300) and these second communication devices (400). The error statistics are used as input for this execution step (504a).

13. The method (500) as claimed in claim 12, wherein, Receiving (502) includes determining these estimated channel coefficient matrices at the first wireless communication device (300). And its error statistics, or receive the information from the corresponding second wireless communication device (400).

14. A method (600) for operating a second wireless communication device (400) wirelessly connected to a first wireless communication device (300) in a MU-MIMO RSMA communication system, the method comprising: Receive at least (606) precoding parameters and combining parameters (V) from the first communication device (300). c V k U c,k U k ), Receive (608) signal (y) from the first communication device (300) k This signal includes a common signal portion (s) c ) and private signal section (s k And based on the same precoding parameters and combining parameters (V) previously received from the first communication device (300). c V k U c,k U k Precoding is performed. Using the previously received precoding parameters and merging parameters (V) for precoding. c V k U c,k U k The common signal portion and private signal portion received at multiple antennas (402) are combined (610) to obtain the combined common signal portion and the combined private signal portion. The combined common signal portion and the combined private signal portion are provided (612) to the detector (406) to estimate (614) the transmitted common signal portion and private signal portion, respectively. The signal estimated by (622) is provided at the output.

15. The method (600) of claim 14, further comprising: Receiving (606) precoding parameters and combining parameters (V) from the first communication device (300). c V k U c,k U k )Before: Estimate (602) at least one channel coefficient matrix of the communication channel between the second wireless communication device (400) and the first wireless communication device (300). as well as The estimated channel coefficient matrix Transmit (604) to the first wireless communication device (300).

16. The method (600) as claimed in claim 14 or 15, wherein, The public and private signals transmitted by (614) are estimated to include: From the received signal (y) k The detection of common signals (S) in the ) c ),as well as Using the detected common signal component (s) c The known information of ) is used to obtain the private signal part (s) k ).

17. The method (600) of claim 14, 15, or 16, further comprising, in the detector: In the first decoder (408), the common signal part (s) c Decode (616) Use the decoded common signal portion (s) c ) and the merged common signal portion (s) obtained from the merging step (610). c ) and the merged private signal section (s k ) is used as input to perform interference cancellation (618), Decode (620) the private signal portion (s) from the signal obtained through interference cancellation (618). k ).

18. A wireless communication device (300, 400) comprising one or more microprocessors (350), volatile memory (352) and non-volatile memory (354), and a wireless interface circuitry (356) configured to transmit and / or receive electromagnetic signals via a plurality of antennas (306, 402), wherein, The non-volatile memory (354) stores computer program instructions that, when executed by the microprocessor (352), configure the wireless device (300, 400) to perform the methods described in one or more of claims 1 to 4, 5 to 10, 11 to 13 and / or 14 to 17.

19. A computer program product comprising computer program instructions, which... When executed by a microprocessor (352) of a wireless communication device configured as a transmitter (300), the microprocessor (352) is caused to perform the method according to one or more of claims 1 to 4, 5 to 10 and / or 11 to 13 and accordingly control the hardware components (356) of the transmitter (300) of the RSMA MU-MIMO communication system, or, When executed by a microprocessor (352) of a wireless communication device configured as a receiver (400), the microprocessor (352) is made to perform the method according to claims 14 to 17 and accordingly control the hardware components (356) of the receiver (400) of the RSMAMU-MIMO communication system.

20. A computer-readable medium or data carrier that retrievably transmits or stores the computer program product as claimed in claim 19.

Images