MU-MIMO communication method distinguishing between data and reference symbols for power allocation and corresponding devices

Abstract

MU-MIMO Communication Method Distinguishing Between Data and Reference Symbols for Power Allocation and Corresponding Devices. The invention relates to a communication method (1) comprising: transmission to user equipment, using a dedicated control channel (PDCCH) per user equipment, of control information indicating, for a user equipment, at least one power ratio () per resource element (RE), time frequency between reference symbols (DMRS), and data transmitted to that user equipment for a given antenna port (). Figure for the abstract: Fig. 5

Description

Title of the invention: MU-MIMO communication method distinguishing between data and reference symbols for power allocation and corresponding devices Scope of the invention

[0001] The present invention relates to the field of telecommunications. Within this field, the invention relates more particularly to digital communications implemented by a communication system with a transmitter M transmitting antennas and receivers with N receiving antennas, known as MIMO (Multiple Input Multiple Output), which allows a power level to be allocated per spatial layer, i.e., per transmitting antenna port while maximizing a transmission rate.

[0002] For downlink communication, the transmitter can be a base station, for example, an eNodeB (evolved Node B) or gNodeB for networks based on LTE, LTE Advanced and subsequent technologies (5G, etc.), or a Wi-Fi access point, etc. A receiver can be a terminal such as a smartphone, tablet, connected device, etc. For uplink communication, there are as many transmitters as terminals, and the receiver is a base station.

[0003] The invention finds applications in any beamforming-based system, in particular in radio communication networks according to the 4G, 5G and subsequent standards defined by 3GPP, WiFi communication networks according to the various IEEE 802.11 standards, etc. Previous art

[0004] Figure 1 illustrates a multi-user MIMO communication system with M transmit antennas ANT_E and U receivers with receive antennas ANT_R each with < M- The system includes at least one access point, called base station SB in cellular systems, and U terminals UE, those of the users.

[0005] In the LTE (3 / 4G) standard specified by 3GPP, the number of receiving antennas for a so-called "smartphone" terminal is at least two, while for the NR (5G) standard of 3GPP, the number of receiving antennas for this type of terminal is specified at least four antennas for certain bands (3.5GHz n77 / 78). For base stations, the number of transmitting antennas has steadily increased, reaching up to 64, or even 128 antennas in massive MIMO.

[0006] Note that an antenna port within the framework of the 3GPP mobile standards and in the context of the request is associated with a "logical" antenna which can group one or more RF (Radio Frequency) channels, each RF channel including a conversion digital to analog and / or including an analog to digital conversion. Mapping a logic antenna to N > 1 RF channels is done via a complex vector of dimension N: the signal sent to a logic antenna is multiplied by this complex vector to generate the N inputs of the N RF channels.

[0007] Each antenna comprises one or more radiating elements.

[0008] While the number of transmit and receive RF channels is the same at the base station, this is not the case for mobile devices, which may have fewer transmit RF channels than receive RF channels, typically two transmit RF channels for every four receive RF channels. With regard to a base station, a transmit / receive RF channel is referred to as a "TXRU, Transceiver Unit" in the 5G standard.

[0009] Figure 2, which corresponds to Figure 6.3-1 of the TS36.211 specification, schematically illustrates part of the baseband processing of the data to be transmitted. According to this standard, the data can be encoded using two streams with different yields. The encoded data, in the form of Cw codewords, is scrambled with an SCR scrambler and then mapped to a constellation with a C_MAP mapper. The mapper outputs are distributed into spatial layers by an L_MAP layer mapper. The different Lay spatial layers are then pre-coded by a PRE_COD pre-coder. The pre-coded data from the different spatial layers are mapped by resource element by a RE_MAP resource mapper. From the resource mapper output, the data from each spatial layer is injected into an OFDM multicarrier modulator to generate an OFDM symbol. These OFDM symbols feed the antenna ports.Each spatial layer corresponding to an antenna port feeds a single user (or recipient) at any given time. A user can benefit from several spatial layers simultaneously. In the uplink, the processing specified by the normative document TS36.211 and corresponding to [Fig.5].3-1 of this document is very similar to that of [Fig.2].

[0010] In a conventional manner for telecommunication systems with a transmission channel CH between a transmitting part EM and a receiving part RX1, RXu, RXU, the transmission resources are represented according to one or more time-frequency grids, a representation of which is given by [Fig.3] which corresponds to [Fig.6].2.2-1 of the TS36.211 specification of the LTE (4G) standard of the 3GPP.

[0011] In the context of the 4G or 5G standard, a time-frequency radio resource is a resource element (Resource Element, RE). A resource element, in this context, designates the smallest time-frequency granularity that is identified by the indices k and l (i.e., a subcarrier and a symbol time corresponding to the duration of a multicarrier symbol, typically OFDM in the case of an OFDM multicarrier transmission) on the time-frequency grid of [Fig. 3]. On In this grid, frequencies follow the vertical axis and time follows the horizontal axis.

[0012] For the 5G standard, there is a time-frequency grid per antenna port for each uplink and downlink direction. According to this standard, a frame is divided into subframes. Since the subcarrier spacing of an OFDM symbol is configurable, each subframe contains a number of OFDM symbols that varies according to the subcarrier spacing, as the duration of a symbol is the inverse of the subcarrier spacing. The OFDM symbols are grouped into slots, each slot corresponding to the granularity of a known scheduling decision per elementary time interval (TTI).

[0013] With reference to the radio communication networks specified within 3GPP, mechanisms are specified for a terminal to be connected to a base station. The Physical Downlink Shared Channel (PDSCH) is the main channel used for unicast data transmission and for system information transmission, hereafter referred to as the data channel.

[0014] The receiver determines the time-frequency resource elements it must decode by exploiting control information defined semi-statically at the upper layer (RRC) and dynamically transmitted control information. A dedicated physical control channel per terminal, called PDCCH (Physical Downlink Control Channel), is used for the dynamic transmission of control information for the downlink and uplink, and for scheduling the PDSCH and PUSCH transmission physical channels.

[0015] The PDCCH channel is organized according to several possible formats, called DCI (Downlink Control Information) formats. A DCI format comprises several fields, each field carrying specific information (e.g., allocated transport blocks or PRBs (Physical Resource Blocks) (one or two transport blocks may be allocated), modulation and coding schemes or MCS (for "Modulation and Coding Scheme") allocated for each transport block, identification of allocated antenna ports corresponding to the configuration by the RRC (semi-static) layer of reference symbols, DMRS, number of spatial layers (rank), etc.). The information bits of a PDCCH channel (i.e., the bits of the DCI format) are then associated with a CRC ("Cyclic Redundancy Check") code to enable error detection. The distinctive feature of this CRC code is that it is scrambled with the RNTI identifier of the terminal to which the PDCCH channel is dedicated.This allows the terminal to validate that the PDCCH channel it is decoding is indeed intended for it. If another terminal (with a different RNTI identifier) ​​attempts to verify the validity of the PDCCH channel using this other RNTI identifier, the CRC code verification will return an error.

[0016] As is known, the capacity of mobile telecommunications cellular networks, And in particular, the capacity of networks specified in 3GPP (LTE / 3G-4G, 5G, etc.) is limited by interference. This interference can be of various kinds. Among the most damaging in terms of cellular network capacity, we can distinguish in particular - SU-MIMO interference (for Single User - Multiple Input Multiple Output) related to the use of multiple antennas in transmission and reception, and which corresponds to the interference generated between MIMO data streams allocated to the same terminal; - MU-MIMO (Multiple User - Multiple Input Multiple Output) interference, related to the use of multiple antennas for transmission and reception, which corresponds to the interference generated between MIMO data streams allocated to different terminals; and - Intercellular interference, generated between signals emitted by different cells and destined for different terminals.

[0017] Various methods for reducing the effect of these interferences on network performance are known from the prior art.

[0018] Thus, MU-MIMO type interference can be dealt with in particular by means of beamforming.

[0019] Beamforming is a signal processing technique used in MIMO communication systems comprising arrays of antennas or sensors for the directional transmission or reception of signals. This method consists of applying complex coefficients to the data streams transmitted or received by the antennas in order to assign them particular spatial properties (for example, a preferred direction).

[0020] Thus, the complex coefficient applied to each antenna or antenna port is called the pre-coding coefficient, and the set of these coefficients forms the pre-coding matrix W implemented by the PRE_COD pre-coder. This pre-coding allows for the spatial separation of streams destined for different terminals, enabling receiving terminals to detect a stream with reduced interference from other streams.

[0021] Beamforming is thus achieved by combining the elements of a phase- and amplitude-controlled antenna array in such a way that:

[0022] • the signals interfere constructively in particular directions,

[0023] • the interferences are destructive in other directions.

[0024] Beamforming can be used on the emitting side but also on the receiving side to obtain spatial selectivity.

[0025] By applying appropriate W pre-coding, the transmitter can transmit data on the same time-frequency radio resources to several receivers with several

[0026]

[0027]

[0028]

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035] spatial layers, known as MU-MIMO mode. The selection of the transmitter side of the appropriate VF pre-coding that maximizes spectral efficiency, evaluated for example by the sum of the MIMO communication rates, is subject to a transmission power constraint. Most prior art precoding algorithms are based on optimizing the performance of a MIMO transmission under a single total transmission power constraint. For example, water-filling power allocation, which allocates power based on channel conditions, maximizes SU-MIMO capacity under a single total transmission power constraint per resource element. Let x be the data vector to be transmitted, x → ... coding, yy = 1 yy yy W ri G with M the number of transmitting antennas L 1 2 * v J and Q the diagonal matrix of power transmitted per antenna port, Q = dia^Q(Q? The VF pre-coding matrix virtualizes the M transmitting antennas into v antenna ports. The antenna concept itself can be based on a virtualization of radiating elements. For example, two radiating elements can have the same input signal and therefore share the same virtualized antenna port. The VF pre-coding is described by a matrix of size M x v, that is, of M rows, the number of transmitting antennas, and v columns, the number of antenna ports, which can be written: = with m = 1, ..., M and l = 1, ..., v .Wi The total transmission power constraint for a given resource element can then be expressed in the following form: IrfWQWjsP-,® where tr( ... ) is the trace, yy 1 is the transconjugate matrix, called "daguer", that is to say the transpose matrix of the conjugate matrix of the W7 matrix, P max is the maximum energy per resource element (EPRE). In the TS38.214 standard published by 3GPP (5G), under clause 4.1 "Power allocation for downlink", the ratio P between the energy per resource element (EPRE) of the DMRS references and the energy per resource element (EPRE) of the PDSCH data channel is tabulated according to the DMRS configuration. The drawback of this The approach is that a spatial layer with very low power allocation will have a very noisy channel estimate because the associated DMRS will have very low power. For simplicity in the request, power per resource element refers to energy per resource element.

[0036] The above-mentioned approach is not well suited to real MIMO systems since these are characterized by a limitation of transmission power by power amplifier, i.e., by antenna (TXRU).

[0037] Pre-coding algorithms must therefore satisfy these power constraints per antenna while maximizing MIMO communication performance. The power constraint transmitted by an antenna m must therefore be given by the following expression:

[0038] |WOW11 <PcmtY

[0039] Which can also be written in the form: 100401

[0041] where P aiit is the maximum power per antenna per resource element and Ql the power per resource element for layer / .

[0042] The user u's pre-coded channel corresponds to: ... J with the channel corresponding to user u.

[0043] At the receiver M, the pre-coded channel, i.e., which includes the pre-coding Wf, is estimated from the mutually orthogonal (in frequency and / or time and / or code) demodulation reference signals (DMRS for LTE or NR (3GPP 5G) systems) that are transmitted on the same antenna ports as the data destined for this receiver u and have therefore undergone the same pre-coding W. This estimation of the pre-coded channel is used in reception during the demodulation of the received signals and the decoding of the data. The DMRS symbols carried by a spatial layer corresponding to an antenna port are orthogonal to the DMRS symbols carried by another spatial layer, i.e., another antenna port, and are not interfered with by useful symbols. A DMRS symbol is user-specific.

[0044] The quality of the pre-coded channel estimation performed at the receiver depends on the reception power of the DMRS signals. However, these DMRS signals, as well as the data, are transmitted with a single maximum total power constraint Pmax per resource element, given that the different antenna ports share the same time-frequency resource elements.

[0045] The disadvantage of this approach is that when the data and symbols of DMRS reference transmitted through an antenna port benefits from very low allocated power, so channel estimation is very noisy. Main features of the invention

[0046] The invention relates to a communication method intended to be implemented by a base station of a MU-MIMO system with U > 2 user devices, the system comprising M > 2 transmit antennas, at least 7 antenna ports and N ≤ 2 receive antennas distributed as > 1 receive antenna per user device, Nu > using time-frequency radio resource elements of transmission allocated to at least one antenna port assigned to user equipment u, among the 7 antenna ports, 1 u < U, and dedicated to communication between the base station and user equipment w, such as:

[0047] Communication control information between the base station and the U user equipment, issued by the base station, indicates, for a user equipment u, at least one power ratio per radio time frequency resource element between reference symbols and data transmitted to that user equipment 11 for a given antenna port.

[0048] The invention further relates to a communication method intended to be implemented by a user device taken from among U user devices of a MU-MIMO system with U ≤ 2 user devices and a base station, the system comprising M > 2 transmit antennas, 7 antenna ports and N > 2 receive antennas distributed as > 1 receive antenna per user device, _ Ca*' using time-frequency radio transmission resource elements allocated to at least one antenna port assigned to user equipment u, among the 7 antenna ports, 1 u < U, and dedicated to communication between the base station and user equipment u, such that: - communication control information between the base station and the user equipment, issued by the base station, indicates at least one power ratio per time-frequency radio resource element between reference symbols and data transmitted to that user equipment u for a given antenna port.

[0049] The invention further relates to a communication method for a MU-MIMO system with U >2 user devices, the system comprising M >2 transmitting antennas, at least 7 antenna ports and N >2 receiving antennas distributed as >1 receiving antennas per user device, such that Power allocation distinguishes between power allocated to data and power allocated to reference symbols intended for one of the user devices and takes taking into account a maximum power per transmitting antenna.

[0050] The invention further relates to an access point comprising M transmitting antennas, at least v antenna ports, a transmitter, a receiver, a computer, such that the transmitter is capable of transmitting: data to U >2 user equipment, control information indicating radio time frequency transmission resource elements allocated to at least one antenna port, among the v antenna ports, assigned to a user equipment M,, 1 S u < U, via a control channel dedicated to the user equipment 11, at least one power ratio per time frequency resource element between reference symbols and data transmitted to that user equipment u for a given antenna port.

[0051] The invention further relates to a telecommunications terminal comprising NFT > 1 receiving antenna, a receiver, such that the receiver is capable of receiving: reference symbols emitted by an access point having at least v transmit antenna ports, control information indicating radio time frequency transmission resource elements allocated to at least one antenna port assigned to the terminal from among the v antenna ports, issued by the access point, via a control channel dedicated to that terminal, at least one power ratio per time frequency resource element, between reference symbols and data transmitted to that terminal for a given antenna port, issued by the access point.

[0052] The invention further relates to a computer program on an information medium, said program comprising program instructions adapted to the implementation of a method according to the invention when said program is loaded and executed in a terminal or access point.

[0053] The invention further relates to an information carrier comprising program instructions adapted to the implementation of a method according to the invention, when said program is loaded and executed in a terminal or access point.

[0054] The invention further relates to a digital signal received by a terminal, comprising reference symbols emitted by an access point having at least v antenna ports, control information indicating at least a power ratio per radio resource element time transmission frequency allocated to at least one antenna port assigned to the terminal among the r antenna ports, between reference symbols and data transmitted to that terminal for a given antenna port.

[0055] In the case of a MU-MIMO system, there is an overlap between users between the allocated time-frequency resource elements even if the assigned ports are

[0056]

[0057]

[0058] different between users. User L1 benefits from at least one antenna port. Each user u among the U simultaneously scheduled users can benefit from one or more antenna ports. The Vu > 1 transmit antenna ports assigned to a user are distinct from the transmit antenna ports assigned to other users. The _ yu antenna ports correspond to inputs of a pre-coder whose M outputs feed the M transmit antennas of the MU-MIMO system. And the pre-coding applied to the reference symbols received by a user device is the same as that applied to the data destined for that same user device. According to the process, whether or not the data benefits from a different power than that of the reference symbols for the same antenna port, the user equipment u uses the received power ratio for an antenna port to correct the pre-coded channel estimation made from the reference symbols by weighting this estimation with the received ratio. A user device can thus estimate the pre-coded transmission channel, i.e., the channel that includes pre-coding, by exploiting the received DMRS, under good conditions, without needing to be provided with the pre-coder or information about its selection. Indeed, transmitting the power ratio per resource element between reference symbols and data for the same antenna port allows for channel estimation at the receiver to be performed independently of the power allocated to the data transmission channel.

[0059] Since the pre-coding used in transmission can vary dynamically, i.e. at the rate of sending information on the control channel, the sending of power reports follows the same dynamic, i.e., it occurs at each transmission time interval.

[0060] In particular, the invention relates to the optimization of power allocated per antenna port with a power constraint per antenna given the IL pre-coding by maximizing the dual function of the Lagrangian with an iterative gradient descending algorithm. This approach significantly simplifies the calculations to be performed iteratively compared to the prior art.

[0061] According to one embodiment, the user equipment u receives control information via a control channel dedicated to that equipment. The control channel is, for example, the PDCCH channel according to the 5G standard of 3GPP.

[0062] According to one embodiment of the invention, the control information further indicates a configuration of reference symbols associated with the antenna ports.

[0063] According to one embodiment of the invention, the control information indicates as many power ratios as there are antenna ports assigned to the equipment used. liseur u.

[0064] According to one embodiment of the invention, the configuration is such that the antenna ports are grouped to form NCDM groups such that the reference symbols are spatially code-multiplexed between the antenna ports of the same group and according to which the control channel dedicated to the user equipment u indicates a unique power ratio for each set of antenna ports belonging to the same group and assigned to the user equipment u.

[0065] According to one embodiment of the invention, the method further comprises the use of the same pre-coding matrix to pre-code the data and reference symbols to be transmitted to the user equipment11 and whose composition depends on an estimation of the channels between the base station and each of the U user equipment, the set of channels forming a global channel called MU-MIMO.

[0066] According to one embodiment of the invention, the pre-coding matrix is ​​based on a pseudo-inversion of the MU-MIMO channel.

[0067] According to one embodiment of the invention, the method further comprises a singular value decomposition of the channel matrix Hu between the base station and each user device u such that, ff =j] V' e with UG 'a matrix containing the set of so-called output eigenvectors, ye qMxN“r the matrix containing the set of so-called input eigenvectors and X. = diag^, .... ^7) e a square matrix c<>n>™an> positive real singularities arranged on the diagonal coefficients such that 4, < 2^2 - ' ' ■ ^uJ^R ■> the pseudo-inversion being performed on the matrix r. 1 ir T / tl 7' ^VxA / , called the matrix containing the set of vectors Z = î Vj V2 ... and proper inputs of the MU-MIMO channel.

[0068] According to one embodiment of the invention, the pre-coding matrix is ​​based on a regularized version of the MU-MIMO channel inversion.

[0069] According to an embodiment of the invention, further comprising determining the power allocated to the data allocated to antenna ports and intended for the user u to maximize a sum of the user data rates, subject to a maximum power constraint per transmitting antenna, by alternating between maximizing a Lagrangian L with respect to the power matrix Q for a given value of l1 and then, after updating the value of the vector I1 by taking into account its gradient, maximizing the Lagrangian L with respect to P, with „ _ L. .. 1 P p* * * * ' J vector of Lagrangian multipliers corresponding to the power constraints via antenna.

[0070] According to one embodiment of the invention, the determination of the power allocated to the data also takes into account an EMF constraint at at least one point in space defined by a distance from the base station. This method makes it possible to comply with a dual constraint of electromagnetic exposure and power per antenna.

[0071] According to one embodiment of the invention, the data is transmitted via a PDSCH downlink data channel of a 5G access network. List of figures

[0072] Other features and advantages of the invention will become more apparent upon reading the following description of embodiments, given by way of simple illustrative and non-limiting examples, and the accompanying drawings, among which:

[0073] [Fig-1] [Fig.1] is a diagram of a telecommunications system,

[0074] [Fig.2] [Fig.2] is a very generic diagram of a baseband architecture from an issuer described in relation to prior art,

[0075] [Fig.3] [Fig.3] is a diagram of a time-frequency grid of resource elements described in relation to the prior art,

[0076] [Fig.4] [Fig.4] is a diagram illustrating different configurations of the grids associated with the antenna ports,

[0077] [Fig.5] [Fig.5] is a flowchart of an embodiment of a process according to the invention,

[0078] [Fig.6] [Fig.6] is a diagram of an embodiment of an access point according to the invention,

[0079] [Fig. 7] [Fig. 7] is a diagram illustrating a terminal according to the invention. Description of particular embodiments

[0080] The invention relates to a multi-user MIMO communication system illustrated by Figure 1, already described in relation to the prior art. The MIMO system under consideration comprises at least one access point, called a base station in cellular systems, and U > 2 terminals, those of the users.

[0081] Such a transmission system SYS comprises a transmitter EM and U receivers ÆX1 ... RXu ... RXU, communicating with the transmitter via a global channel CH between the transmitting and receiving antennas. U is the number of receivers (also called "user equipment, terminals, or users") and u is an integer between 1 and U used to identify a receiver. Hereafter, a given receiver is designated by its index (1, ..., ..., U), or by the reference RXi (example: RX\ ...,RXu, ...,RXU where 1... u... U is the index of the receptor).

[0082] The EM transmitter comprises M transmitting antennas ANT_E (M > 1), which may interfere with each other. Similarly, each receiver w comprises NUR receiving antennas ANT_R (A^ > 1), which may interfere with each other.

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089] The global CH channel can be described by an NxM dimension matrix H with By applying appropriate VF precoding, the transmitter can send data to multiple receivers on the same time-frequency resources while minimizing interference at the receiver. In other words, the SYS transmission system is a multi-user MIMO type. The VF precoding matrix virtualizes the M transmit antennas into v antenna ports. The data vector intended for user u before precoding is of size h; that is, the data before user precoding is distributed across spatial layers. User precoding is described by a matrix of size MX vH, i.e., M rows and L columns. The number of spatial layers that a receiver u can receive cannot exceed minQV^, M). In the case of a single receiving antenna at receiver u, = 1, it can only receive one spatial layer. The VF precoding, for all users, is described by a matrix of size M x v. The transmitter sends reference symbols to each of the receivers using the or the same antenna ports as those used for data destined for this receiver, i.e., these reference symbols are therefore pre-coded like the data at using the same pre-coding method. The VF pre-coding is described by a matrix of size M x v, that is, of M rows and v columns, which can be written: w,„j] with m = 1, ..., M and l = 1, ..., v wMa> The reference symbols transmitted to the various receivers are distributed across the time-frequency planes (grids) associated with the antenna ports according to a specific configuration. The reference to this configuration is transmitted to the receivers via a protocol associated with the RRC layer and in a semi-static manner, meaning on a timescale much larger than the scheduling of a receiving device, which is performed on an elementary timescale, known as TTI. It is possible to distinguish different configurations of DMRS symbols that define CDM (Code Division Multiplexing) groups associated with a number of antenna ports. Figure 4 illustrates the CDM group configuration concept for the 5G standard. Each time-frequency grid whose structure matches that of Figure 3 corresponds to a spatial layer or antenna port. Regardless of the group configuration, when a resource element of a grid bears a symbol If there is no reference symbol, then no data transmission for that same resource element can occur on any of the grids (the power for data is therefore zero for that resource element on each grid), and thus there is no spatial interference between DMRS reference symbols and data. However, for the same resource element (RE), there may be a DMRS reference symbol on one or more other grids. In this case, to avoid interference between DMRS reference symbols across multiple grids, the DMRS reference symbols carried by the same resource element are encoded with spreading codes of a so-called spatial code, which are different between grids. Since the DMRS reference symbols of a grid can be distributed solely by frequency, solely by time, or by both time and frequency, the spreading code has a frequency dimension, a time dimension, or both a frequency and a time dimension.

[0090] The upper part of [Fig. 4] represents a first example of a DMRS reference symbol configuration. This configuration defines the CDM groups 0, 1, and 2 for six antenna ports, 1000-1005, when the DMRS symbols are simple (distributed over a single OFDM symbol time) and such that, for a grid, the DMRS symbols are grouped in pairs and occupy two RE resource elements along the frequency axis. Between two grids, i.e., two ports, of the same group, the DMRS symbols occupy the same RE resource elements and are therefore encoded differently with a spatial code so that the receiver can distinguish the different ports. Thus, within the same group and between two ports, the DMRS symbols are orthogonal to each other due to the spatial code. Between the grids of two successive groups, for example, between CDM groups 0 and 1, the DMRS symbols are offset by two RE resource elements along the frequency axis.Therefore, between grids of different groups, the DMRS symbols are orthogonal to each other in frequency.

[0091] The lower part of [Fig. 4] shows a second example of a DMRS reference symbol configuration. This configuration defines CDM groups 0, 1, and 2 for twelve antenna ports, 1000-1011, where the DMRS symbols are doubled (distributed over two OFDM symbol times) and such that, for a grid, the DMRS symbols are grouped in fours and occupy two RE resource elements along the frequency axis and two RE resource elements along the time axis. Between two grids of the same group, the DMRS symbols occupy the same RE resource elements and are therefore encoded differently so that the receiver can distinguish the different ports. Thus, within the same group and between two ports, the DMRS symbols are orthogonal to each other due to the spatial encoding. Between the grids of two successive groups, for example, between CDM groups 0 and 1, the DMRS symbols are offset by two RE resource elements along the frequency axis.Therefore, between grids of different groups, the DMRS symbols are orthogonal to each other in .

[0092]

[0093]

[0094]

[0095]

[0096]

[0097]

[0098]

[0099] frequency. As indicated above, receiving equipment knows the configuration of reference symbols, i.e. the maximum number of groups and the structure of these groups (reference symbol spread over one or more symbol times, over one or more frequencies, spatial code used, ports per group). According to the invention, the power allocation distinguishes between data and reference symbols. Power allocation to reference symbols Let ^^DMRS) be the power ^PMRS) DMRS reference symbols for the antenna port / and therefore for estimating the corresponding precoded channel A / , with e G {1, ...,£} the index of the resource elements bearing these symbols, and E the number of reference symbols per antenna port. Depending on the DMRS configuration, the power per resource element (EPRE) per antenna, denoted Pant, is shared or not between several DMRS symbols. Indeed, several cases must be considered depending on the DMRS configuration: - 1. Case where the DMRS symbols between the different antenna ports are only multiplexed in code (in the case of the examples in [Fig.4], they belong to the same CDM group, for example CDM group 0, that is to say they are spatially multiplexed within the "CDM group", i.e., the reference symbols occupy the same resource elements between the ports and the reference symbols of the same port are coded with a spreading code specific to the port and different from that of another port, the reference symbols of one port are therefore orthogonal in frequency and / or time with the reference symbols of another port), it comes by taking up relation (3): • a. And in the case where the power between the antenna ports is identical, q{dmrs) _ q is constant, V / g{1, ... , v], we have: 11 w,rd !2 Part V mg {1, ..., M} In this case, according to the method of the invention, the power allocated per resource element to the DMRS reference symbols for each antenna port l that best utilizes the available power per Pant antenna is: qID^S) = qv / s{lv| In this case, the powers of the DMRS symbols belonging to the same "CDM" "e" • b. Or in the case where a signaling constraint requires the transmission of a single ratio between the energy per resource element (EPRE) of the DMRS reference symbols and the energy per resource element (EPRE) of the PDSCH data channel:

[0100]

[0101]

[0102]

[0103]

[0104]

[0105] , then _ Bang! In this case, according to the method of the invention, the power allocated per resource element to the DMRS reference symbols for each antenna port / is: / DMRS) 2. Case where the DMRS symbols between the different antenna ports are not multiplexed on the same resource element, i.e. the vector has EC which is transmitted for the estimation of the pre-coded channel has a single non-zero element ^^DMRS) & corresponding to layer / , it comes by taking relation (3): / DMRS) ant In this case, according to the method of the invention, the power allocated per resource element to the DMRS symbols for each antenna port l that best utilizes the available power per antenna Pml is: / .DMRS) _ 7 united maxlw' J m I i 3. General case of several "CDM groups", for example, the first vi DMRS symbols belong to one "CDM group" and the remaining v2 to another "CDM group", • a. and in the case of an equally distributed allocation, the power allocated per resource element to the DMRS symbols for each antenna port / which best exploits the available power per antenna Pai!t is, according to the method according to the invention:

[0106]

[0107]

[0108] / DMRS) _ Panel —----~- VZE 1, maxL, And / DMRS) _ PrMt 7 y-'V:-' maxL,_ œ '1+1 , V1 + v2 • b. or in the case where a signaling constraint imposes the transmission of a single P ratio per "CDM group", the power allocated per resource element to the DMRS symbols for each antenna port l that best exploits the available power per Pant antenna is, according to the method according to the invention:

[0109] p 2 Q , G 1 KCDM_l ​​l ' d

[0110] and

[0111] ^> = pcm^^Qi v / c(,|tl.

[0112] With o _ ! the ratio corresponding to the first “CDM group” "cdm_1 “ \ v । (2,, and "_ I the report corresponding to the second "CDM group". Pcdm_2~\ .2„ yj<

[0113] From case 3), the allocated power can be determined when there are more than two "CDM groups" associated with a number of Vma\ ports.

[0114] In the case of an equally distributed allocation, the power allocated per resource element to the DMRS symbols for each antenna port / is, according to the method according to the invention:

[0115] -.(DMRS) _ 4 ~ V / e 1, ..

[0116] And

[0117] P <mf maxE^H^2 (DM RS)

[0118] And for vj e {v3, ..., :

[0119] APMRS) Pnl V / E Vj+V2+ • • • "b Vy.] + 1, ..., Vj + Vo + ... + Vj

[0120] In the case where a signaling constraint imposes the transmission of only one / 5 ratio per "CDM group", the power allocated per resource element to the DMRS symbols for each antenna port l is, according to the method according to the invention:

[0121] ^=pcm^Ql v / e (1....,,]

[0122] and

[0123] v / g|,.1+1, ,4

[0124] And for Vj e {v3, ..., vmax}: ^DMRS)_XQ; V l E + V2 + ... + Vy.] + 1, ... , V{ + V2 + ... + VJ

[0125] o _ I the report corresponding to the first “CDM group”, P CDMA ~ \ v:' i iV> R — I Rm' 'c raPP°rt corresponding to the second "CDM group" and PCDM_2 ~ U v1'"' । ,2^ 1| M=„(+iPvm, / | Qj

[0126] o _ I the report corresponding to the jth “CDM group”. PCDMJ ~ U y'W , .2

[0127] The pre-coded channel can be expressed in the form of a matrix:

[0128] And the / th column of the matrix is ​​noted: ^MRS.

[0129] The received signal corresponding to the transmission of a DMRS symbol carried by a resource element e associated with the spatial layer / has the expression: 101301 y^ = ^DmS> V / e {l □

[0131] With nl,e the noise at reception associated with the spatial layer l and the resource element e

[0132] The estimation of the pre-coded channel hs can be obtained by multiplying by and by "by means / interpolating" on e knowing that the other resource elements ee {1, ..., E} of this layer / which do not carry DMRS symbols carry data symbols which benefit from a Qf power. It then follows that: 101331

[0134] where is the power of the DMRS symbols for the antenna port l (carrying the spatial layer 1) while Qj is the power of the data symbols for the antenna port l, per resource element. The DMRS symbols and the data symbols carried by the resource elements of the antenna port / are intended for the same user 11.

[0135] Estimating the channel associated with the antenna port l requires knowledge of / Wœ = PUiSqUe

[0136] Control information, acquired semi-statically or dynamically via a control channel, for example the PDCCH channel in 4G or 5G, dedicated to user equipment u, indicates time frequency resource elements RE assigned to user equipment u.

[0137] According to the invention, a control channel for data transmission between the transmitter and the user equipment u indicates power ratios, qj ^PMRS) or ,, / z3 / ADMRS}, per radio resource element RE time frequency, Ppdschj Q between DMRS reference symbols and data transmitted to user equipment 11, ve being the number of antenna ports allocated to a data transmission to user equipment u.

[0138] According to the invention, the control information intended for the U receivers thus comprises v quantified power ratios for all spatial layers l — 1, ..., v:

[0139] o _ / ^DMRS) OR / „ \2 _ n / n{DMRS). PpDSCHJ \ PpDSCHJ / “ *11

[0140] Figure 5 schematically illustrates a method according to the invention implemented by an access point, a base station in the case of a mobile access network. The base station comprises M2 transmit antennas and at least v antenna ports. The base station implements communication method 1 to communicate at a given transmission interval (TTI) with U user devices (receivers). 2 < u < U. Each user device comprises >1 receive antennas. The system comprising the base station and the U user devices therefore includes receive antennas. The base station transmits, to Each user device u, among the U user devices, receives information indicating radio resource elements, transmission time frequency, allocated to at least one antenna port, taken from among the v antenna ports, assigned to the user device u. This transmission is done using a dedicated control channel per user device u.

[0141] According to mobile telecommunications standards published by 3GPP, for example 4G, 5G, this control channel is the PDCCH channel. The information is carried by the DCI (Data Control Information) and indicates the time-frequency radio resource elements (RE) for each transmit antenna port allocated to a user. According to method 1, control information dedicated to the user equipment u indicates a PDCCH ratio of powers per time-frequency resource element (RE) between DMRS reference symbols and data transmitted to this user equipment u for a given antenna port l.

[0142] For downlink transmission, the control information is transported via the control channel which is dedicated to a user device and which is typically the PDCCH channel.

[0143] When it receives the DMRS reference symbols associated with an antenna port / which is intended for him, the receiver u estimates the pre-coded çDMRS channel By exploiting the(s) "z power ratio(s), PpDSCniou ( p 1received in the information of V PpDSCHJ / control according to the invention, the receiver 11 weights the pre-coded channel estimated at starting from the reference symbols, it obtains the pre-coded channel which allows it to ul- later to equalize the received data associated with the same antenna port / .

[0144] In the case of DMRS reference symbols belonging to different configuration groups, "CDM group", according to the invention the method can transmit only one power ratio, PPDSCH, per CDM group, per user, in the case where the ratio between the power of the DMRS and data carried by the same antenna port is the same for all DMRS ports belonging to a CDM group.

[0145] Considering, for illustrative purposes only, a configuration with NCDM CDM groups, the channel estimation for a receiver u from the DMRS symbols is based on a set A„ of CDM groups taken from the NCDM CDM groups. Each user 11 being mapped to one or more antenna ports, the set Au indicates the port numbers assigned during transmission to that user M.

[0146] The method according to the invention determines the power ratios, fiPDscHc, between the data and the reference symbols for each user 11 and for each port assigned to that user, CG Au. According to one embodiment, the power ratios fiPDscHc are quantified, each on a number of bits n > 1 (for example four), and are part of the control information transmitted to the receiver M, for example transmitted in the control channel (PDCCH / DCI) in the downlink. MU-MIMO Downstream Channel Inversion

[0147] The signal emitted from the M antennas is: 101481 w =

[0149] With:

[0150] v — [ tT vT v 1T <= C'x! the data vector to be transmitted to V receivers •V «A- 2 9 ** ff JV—' A scheduled simultaneously, at a given transmission interval (TTI), each component of which has a normalized power of one with:

[0151] Xu GC " The data vector intended for the receiver w G {1, ..., t / } and which is mapped onto vu spatial layers,

[0152] W = [ W? ... W„ ... ] G çM-kv ia pre-coding matrix where v is the total number of transmitted spatial layers _ ,

[0153] e ie pre-coding associated with receptor ll,

[0154] q — Q ... Q ... Q ) ('rx the power diagonal matrix transmitted via antenna port, [01551 O =rfWp ...,O ) the diagonal power matrix associated with spatial layers onto which the reference symbols and data intended for the receiver u are mapped.

[0156]

[0157]

[0158]

[0159]

[0160]

[0161]

[0162]

[0163]

[0164]

[0165]

[0166]

[0167]

[0168]

[0169]

[0170]

[0171] The received signal „ e for a receiver w can be written: 'U With : jj e the transmission channel from the transmitter with M transmitting antennas to the receiver M with receiving antennas with Nr < M, n u e C' v * le vecteur de bruit tel q ue R, = E {} = diag ( h ■ ■ ■, ) y jj ipz / v interference due to signals intended for other receivers i * u, ie{l, The invention is situated within the context where the U receivers are scheduled simultaneously. The signal vector received by the U receivers is given by: y = HWQ^x + n with F y- 1^ N. – the total number of receiving antennas, H the combined (global) channel matrix jj _ [ jj jj^ n = [nf, / <, n^]T e Cv the noise vector such that E {nn'} = dia^R^, ■ ■ ■, R^. Subsequently it is assumed that E {tin} - dia^l ■ ■ ■, cr^). According to a first embodiment of the invention, the transmitter's knowledge of the MU-MIMO channel (of matrix H, i.e., the channel between the M transmitting antennas and the A' receiving antennas, H e cNxM, _ yu is the number (total number of receiving antennas) is assumed to be perfect, the pre-coding matrix used W is that obtained by the pseudo-inversion of the channel (obtained according to the so-called Zero Forcing technique) which allows to cancel the interference between the v spatial layers, i.e., the layers are orthogonal to each other. For a receptor u, the channel matrix Hu can be decomposed into singular values ​​(SVD, Singular Value Decomposition) as follows: jj _ jj £ y " e with jj e the matrix containing the set of eigenvectors called output vectors, ye the matrix containing the set of eigenvectors called input vectors and £ _ diag^ / Ç7 v" ) GQ^^'r a matrix a square containing the positive real singular values ​​arranged on the coefficients diagonals such as £ TO 1L H • The received signal can then be written as:

[0172]

[0173] y= Or again: n} ■ UV- [VT W2 ... Wjô? "td

[0174]

[0175] Or, considering the composition of the EU matrix: y= VL ■ *1 ■ X2 + «2 .XU. .naked.

[0176] According to the first embodiment of a process according to the invention, the Zero Forcing (ZF) technique is applied to the matrix, _ r T / j 1T f AxM. that is to say on the matrix containing the set of eigenvectors called input vectors. Thus, the process performs a pseudo-inverse of this matrix Z to obtain: ZF _ 1 E CMxN' ^cttc pseudo-inverse matrix can be expressed by blocks ■. mzf - [Mzf MZF, - - •, MzF, ... Mzf ] with mzf e çM^r.

[0177]

[0178]

[0179]

[0180]

[0181] The pre-coding applied in MU-MIMO transmission for the data and reference symbols intended for receiver 11 corresponds to the first columns of MZF- Wu=\mSm^ ... «K]e(*^ The combined pre-coding matrix, corresponding to the overall channel, is w = [W1 w2 ... wu] e CMx” where r is the total number of transmitted spatial layers „ vü „ . V^L^Vu It is clear that ZW e CVxv is an identity matrix Iv to which has been added, at each initial position j NUR- vM zero rows for all me {1, .... U] ■ The diagonal power matrix Q — diag^Q^ ■ ■ ■, Qv) e Cwv describes the power allocated per spatial layer / G {1, ..., v) where Q g corresponds to the first eigenvalues ​​of Eu. To simplify the notation, it is considered that Q — diagjQp ■ ■ ■, Q ■ H It follows that QuJ = Qs^+l for V ( Z) G { 1, ..., U] x {1, ..., vJ by defining S(u) =

[0182]

[0183]

[0184]

[0185]

[0186]

[0187]

[0188]

[0189]

[0190]

[0191]

[0192] Let yu be the signal received by the receiver u. The signal "' _ ye with u" A'„XvM FJ 1 UH Uu,ï ' ■ ' K«>V„ can be written where — Funu cst a spatially uncorrelated noise vector, that is to say that Subsequently, the eigenvalue 4^(w)+f = V ( u, / ') e { 1, x {1, ..., vM}, is called the eigenvalue associated with the layer / e {L ..,, y}. Similarly, the variance = V ( w, / ) e {1, ..., ¢7} x {1, ..., vw}, is called the variance of layer / e {1, ..., y}. ZF precoding (pseudo-channel inversion) allows for the cancellation of interference between different receivers (or, in other words, ZF precoding allows for the cancellation of the term y / f WjCFx) in order to obtain parallel channels without interference. The noise is further reduced. Interference, SINR, of the receivers are uncorrelated with each other because the receivers are spatially separated due to the ZF pre-coding. Thus, according to the first embodiment, with Zero-Forcing pre-coding, the signal-to-noise plus interference ratio, SINR, of the spatial layer / is given by: 2,Q, SINR; = with the eigenvalue corresponding to the spatial layer Z, Q] the power allocated to the spatial layer / , the noise power of the spatial layer l. In the event that the channel suffers from poor radio conditions (corresponding to equity (zz') (very broad), ZF pre-coding can lead to values ​​of Low SINR at reception. However, by simultaneously scheduling in MU-MIMO transmission only receivers in good radio condition according to a particular embodiment of the invention, the use of ZF pre-coding remains an efficient pre-coding solution. Given the limitations of ZF pre-coding in high-noise (or low SNR) scenarios, according to a second embodiment of the invention, a re- version

[0193]

[0194]

[0195]

[0196]

[0197]

[0198]

[0199]

[0200]

[0201]

[0202]

[0203]

[0204]

[0205]

[0206]

[0207]

[0208] Regularized channel inversion, called MMSE (Minimum Mean Square-Error) pre-coding, is performed. This regularized version shows a significant gain at low SNR and converges towards the performance of ZF pre-coding at high SNR. MMSE pre-coding is given by: WMMSE = É[ZÉ + aI )4 Since intra-user and inter-user interference can no longer be considered zero at low SNR, MMSE pre-coding attempts to maximize the SINR per spatial layer in reception. Data power allocation per antenna port The power allocation Qz to the data is determined to maximize the sum data rates per antenna port which define channels orthogonal to each other due to pre-coding, under power constraints per antenna P an[ : niaxE M log2(l + SNR Z ) subject to the following constraints: [wQW1] <p y me 11, ...,a4 t            -w,m                      t            > These constraints can also be written in the form: me(l.....MJ-Q^O v 1 e(1, ..., »] According to the first embodiment, ZF pre-coding, the signal-to-noise ratio can express in the form gjqR — where ^7 cst 'a eigenvalue corresponding to the spatial layer l, Qt is the power allocated to the spatial layer / and is the power noise for the layer / . Regardless of the pre-coding H = [ fFj W7 ... WL! ] e CMx', for a receiver of the type LMMSE (with a suitable whitening filter), the signal-to-noise ratio of the layer {1, ..., v„] can be expressed in the form: SINR^Q,}= Qsi= SINRs{„)+, Indeed, the signal received "for a receiver 11 can be written: Xi + n By defining IP Mu = Hu ! ■ ■ ■ m^, ] GC^xü comes: yu = Mu^xu + Y^HuW^Xt + nu 1.1.1 — m..,; 93 ^. ^ + ? +' z S- ■ / > •'*' / s ' ; Xi i X: b ru : ; 1 = m,. / Q2 iljXsÿj HJ- / .>y.

[0209] with: < + + RK / Xi ï covarna.-ce du brus s i yes sps »3 c? is^ya u ïjïisaïeyf ï siT$.i'r&î's.Lj£!? s^ySï'e u&iîisajçïîi's

[0210] To resolve the problem of optimization under the constraints, what soit the précodage, the invention recourt à un lagrangien défini par: [ ° 2H1 L(Q n) = E>g2(l + SINR,) - E" d W Q W t'P") 102121 =E^ î iog2(i+siNR l ) - EI^JELqAU 2 'M

[0213] With . ( 1 the corresponding Lagrangian multiplier vector “[Pf • • • ' P v / J to the M power constraints per antenna.

[0214] According to the first mode, ZF pre-coding, Lagangrian can be written in the form: 102151 L(Q. H) =E>g2(l + ^J - E" dWQW^-P»)

[0216] Vv 1 / i, VM (Vv I |2 n \ *^ / =1^»? | 1 + ' j - Wmj I ■ P ant j

[0217] Regardless of the pre-coding, the dual Lagrange function g is defined by:

[0218] g(y) = maxL(Q, p) Q

[0219] Since the utility function (sum of flow rates) is convex, minimizing the function g is equivalent to optimizing the utility function. Calculating the vector M that minimizes § allows us to optimize the utility function and therefore the flow rate.

[0220] The method according to the invention maximizes the Lagrangian L with respect to the power matrix Q for a given value of l1, updates the value of the vector l1 taking into account its gradient and maximizes the Lagrangian L with respect to P. Thus, the method according to the invention alternates between a maximization of the Lagrangian L with respect to the power matrix Q for a given value of M and, after an update of the value of the vector P taking into account its gradient, a maximization of the Lagrangian L with respect to M.

[0221] The repetition of these alternating maximizations according to a particular algorithm allows the process to converge towards the power allocation matrix which meets the constraints on the power per antenna.

[0222] An implementation of this algorithm according to a first mode, when the pre-coding is of type ZF, is given in a condensed form in Annex A and described below.

[0223] In an initial step, the following variables or parameters are initialized. The vector M is initialized to a value p(0\ p = p(°) = the power matrix is ​​initialized to 0 vw, q^A) _ q , the iteration index is initialized to zero, i = 0, the tolerance factor e is initialized to a given value, for example e = 1 () \ the rates learning values ​​fi corresponding to the Lagrangian multiplier are initialized to a given value, for example fi — 0.2.

[0224] The function corresponding to the maximization of the Lagrangian with respect to the power matrix Q for a given value of 'U is denoted:

[0225] q" - argmax L ( Q, p ) Q

[0226] The power allocated to the spatial layer l G {1, ..., v} corresponds to a maximum of the function Q, i.e., when its derivative is zero: 4L(Q,p) _ ~ and therefore:

[0227] / \ ,2 (4) W „ ...... ...... _ \ / In( 1

[0228] The diagonal matrix Q* of allocated powers is given by:

[0229] Q = diag ( ) (5)

[0230] As by definition the dual Lagrange function is: g(p) = maxL(Q, p)

[0231] al°rs : g(p) , pj-

[0232] The process can calculate the initial values ​​of Q*: Q^o> — Q^p(°)j

[0233] After the initialization step, the sequence of the algorithm in Annex A is as follows. - 1. Increment of the index of the current iteration, i — i+ 1, - 2. Given the value obtained in the previous iteration, the value The current of g(p) is given by gQjQ _ pj and the gradient by relation to vector I1 is calculated as follows for all m G {1, ..., M};

[0234] । >2 (6), "dp" dp "* t IM mi I 1 m * - 3. Based on the gradient, the update of each multiplier La The Grangian function for all m in G {1, ..., M] is given by:

[0235] = + 0 A 2 rmrmrm iri

[0236] with G [0,1] the learning rate for the Lagrangian multiplier - 4. Calculation of qW _ 7) j according to the equation Q^) = diag(ôiW ¢0

[0237] And repeat from 1-4 until L impn t J m,»?

[0238] At the last iteration when the difference is less than the threshold e, the matrix de- finished e'XÛ = diag(G-^ , qj includes the powers q _ Q*^ For the data for each antenna port.

[0239] According to a second mode, when the precoding is arbitrary, for example an MMSE type precoding, the calculation of Q* = argmax L ( Q, g ) ^asé on the derivation Q given by equation (4) requires a numerical differentiation of the Lagrangian. An implementation is given in condensed form in Annex B. According to this method, to obtain the derivative with respect to the Lagrangian, 4L(Q,|i), the process according to the invention follows a gradient descend / ascend type approach which relies on a numerical derivative of Æ(Q,m) assuming that the function of the lagrangian is locally convex (around dQ; ).

[0240] Power allocation constrained by exposure limit to electromagnetic fields

[0241] According to one embodiment of the invention, it takes into account an additional constraint related to exposure to electromagnetic fields (EMFE). The EMFE constraints are considered on a set £2 of points in space, of arbitrary shape which does not necessarily correspond to a sphere: the received power Pr at each point r £2 must not exceed an EMFE value determined by a regulatory text (Pr EMFE^ r £2). Each point r £2 can be identified by a distance from the emitter.

[0242] The power received Pr at a point r 6 42 in space is given by:

[0243] Pf.^\hrWQ1 / 2x\2

[0244] with

[0245] x the transmitted data vector x _ [xr xt j7 e C1 whose each component has a power normalized to 1,

[0246] v the total number of transmitted spatial layers v _ y77 ,

[0247] Q e C'X1 the diagonal power matrix which describes the power allocated per spatial layer l G {1, ..., v],

[0248] W = [ Wj ... VFr] gCw>' the precoding matrix and

[0249] hr e the vector representing the channel between the transmitter and point 7.

[0250] These EMFE constraints are added to the antenna transmission power constraints, and the performance maximization problem is now given by: [025i] maxR ,log.( 1 + SNRj

[0252]

[0253]

[0254]

[0255]

[0256]

[0257]

[0258]

[0259]

[0260]

[0261]

[0262]

[0263]

[0264] subject to the following constraints: [WQWl <P ant , V mek And subject to the following constraints: |^WQwV| <EMFE max V ren Qz>0 V le{l, ..., v| To solve the constrained optimization problem, the invention uses a Lagrangian defined by: L(Q, g) =Eklog2(1+SINRi) - iLMiLQMd2 ]ArWQwV| - EMFEnrax) aveC p = g Kw+|X2[ the vector of multipliers Grangien whose first M elements correspond to the M power constraints per antenna and the last elements correspond to the |X2| EMFE constraints. According to this embodiment and with ZF pre-coding, the Lagrangian becomes: L(Q,g) = ^( 1 + ^) - -Pant) |ArWQwV|- EMFEmax) The algorithm described for the first embodiment can be adapted by considering all the constraints: antenna transmission power and EMF constraints on the different points £2 in space. Thus, at iteration i, the calculation of the gradients for me {1, ..., M} corresponding to the antenna transmission power constraints is that given by expression (6) and the calculation of the gradients for m G [M 4-1, ..., M + |} corresponding to the EMFE constraints on the The different points 42 in space are given by the following expression: A(» = = EMFEnlax-|hrWQs(1'i, WV ...,M + |Q|1 Thus, according to this embodiment of the invention, the powers Q, determined according to the method, comply with the power constraint per antenna and are such that the power received at any point r G 42 does not exceed the regulatory EMFE constraint. Figure 6 is a diagram of an embodiment of an access point according to the invention in the context of a 4G or 5G telecommunications system. The access point PA comprises M transmitting antennas ANT_E, at least v antenna ports, a transmitter EM1, a receiver RE1, a memory MEM1 including a buffer, and a computer (microprocessor) pPl whose instructions enable the implementation of a telecommunications method 1 according to the invention. At initialization, the code instructions of the program Pgl are, for example, loaded into the buffer MEM1 before being executed by the microprocessor qPl. The microprocessor pPl controls the various components of the access point, the transmitter EM1 the receiver RE1.

[0265] The access point PA further includes a transmission chain which generally includes at least one DATA input channel COD encoding, one MIMO encoding and one OFDM multi-carrier modulation to generate TBu data packets mapped to the time-frequency resource grids of the antenna ports assigned respectively to the user equipment u (terminals).

[0266] The EM1 transmitter allows these TBu data packets to be sent to user equipment u via a PDSCH data channel.

[0267] The EM1 transmitter also allows the following to be transmitted: - DMRS reference symbols, - control information indicating L antenna ports, among the v antenna ports, allocated to data transmission to user equipment 11 and identifying resource elements RE time transmission frequency allocated to the v antenna ports, according to semi-static (RRC) and / or dynamic mechanisms via a PDCCH control channel dedicated to user equipment M, - a power ratio Ppdschj Pæ" resource element RE time frequency, between DMRS reference symbols and data transmitted to this user equipment u, for a given antenna port l, this (or these) ratio(s) being carried by the PDCCH control channel dedicated to the user equipment M.

[0268] The RE1 receiver receives DMRS reference symbols transmitted by the terminals to estimate the uplink channel of each terminal for the purpose of demodulating the data received by the REL receiver

[0269] By executing the instructions, the microprocessor pPl drives the transmitter EM1 to transmit control information indicating RE time frequency resource elements allocated to at least one antenna port assigned to user equipment u among the v antenna ports, via semi-static (RRC) and / or dynamic mechanisms (via a dedicated PDCCH control channel per user equipment M), and for the PDCCH control channel dedicated to user equipment u to indicate at least one PpDScm Pæ" power ratio RE time frequency resource element between DMRS reference symbols and data transmitted to that user equipment11 by the access point, for a given antenna port l.

[0270] The simplified structure of an embodiment of a terminal according to the invention capable of implementing a receiving method according to the invention is illustrated by [Fig.7].

[0271] The UEW terminal comprises N''r > 1 RXu receiving antennas, an RE2 receiver, a The device consists of an EM2 transmitter, a MEM2 memory comprising a buffer, and a microprocessor (pP2) whose operation is controlled by the execution of a program Pg2 whose instructions enable the implementation of a telecommunication method 2 according to the invention. At initialization, the code instructions of program Pg2 are, for example, loaded into the MEM2 buffer before being executed by the microprocessor pP2. The microprocessor pP2 controls the various components of the terminal, the receiver RE2, and the EM2 transmitter.

[0272] The RE2 receiver includes an OFDM / MIMO / COD 1 receive chain which performs the inverse operations of the access point's transmit chain to receive one or more TBu data packets transmitted by that access point via a PDSCH data transmission channel.

[0273] The RE2 receiver also allows the receiver to receive: - DMRS reference symbols issued by the access point, - control information indicating which antenna ports, among the antenna ports, are allocated to data transmission to user equipment 11 and identifying resource elements RE time transmission frequency allocated to the antenna ports, according to semi-static (RRC) and / or dynamic mechanisms via a PDCCH control channel dedicated to user equipment M, - at least one PPDSchi Pæ" power ratio RE time frequency resource element, between DMRS reference symbols and data transmitted to this user equipment M via antenna port / , this ratio or these ratios being carried by the PDCCH control channel dedicated to the user equipment u

[0274] The EM2 transmitter allows: - to issue DMRS reference symbols.

[0275] Thus, by executing the instructions, the microprocessor pP2 drives the receiver RE2 to receive DMRS reference symbols, control information indicating RE time frequency resource elements allocated to at least one antenna port assigned to this user equipment among the v antenna ports, as well as at least one power ratio pdschj per RE time frequency resource element between DMRS reference symbols and data transmitted per antenna port / to this user equipment u.

[0276] By executing the instructions, the microprocessor pP2 determines a first estimate of the pre-coded MU-MIMO channel by exploiting the received DMRS reference symbols, for an antenna port l assigned to that user, and weights this first estimate with the power ratio pdschj. To obtain an estimate of the channel Pre-coded TBu data intended for user H and received by receiver RE2.

[0277] Power control of the reference signals for channel estimation at the receiver allows for channel estimation quality independent of the power allocated to the data per spatial layer. This power control strategy requires transmitting v power ratios from the transmitter to the receiver, where v < min(My N} is the number of spatial layers selected for transmissions to the user equipment.

[0278] Accordingly, the invention also applies to one or more computer programs, in particular a computer program on or in an information medium, adapted to implement the invention. This program may use any programming language and may be in the form of source code, object code, or code intermediate between source and object code, such as in a partially compiled form, or in any other form desirable for implementing a method according to the invention.

[0279] The information medium can be any entity or device capable of storing the program. For example, the medium can include a storage means, such as a ROM, for example a CD-ROM or a microelectronic circuit ROM, or a magnetic recording means, for example a USB flash drive or a hard drive.

[0280] On the other hand, the information carrier may be a transmissible medium such as an electrical or optical signal, which may be transmitted via an electrical or optical cable, by radio, or by other means. The program according to the invention may, in particular, be uploaded to a network such as the Internet. Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the process in question.

[0281] According to techniques known to those skilled in the art, the transmitter, in order to calculate the pre-encoder and the power ratios between the DMRS and the data, must know the transmitter-to-receiver channel in advance. This knowledge can be based on the principle of reciprocity in TDD. The transmitter estimates the receiver-to-transmitter channel using reference signals transmitted by the receiver. Channel reciprocity allows the transmitter-to-receiver channel to be deduced from the estimated receiver-to-transmitter channel. In the downlink direction, the reference signals transmitted by the receiver (UE terminal) for channel acquisition by reciprocity are the SRS, while in the uplink direction the reference signals used are the CSLRS (transmitted by the access point).

[0282] Another approach can be based on the uplink, from the receiver, of the "quantized" (codebook-based MIMO) transmitter-to-receiver channel. For example, in the direction ] TO DO : On the downlink, the receiver (UE terminal) can estimate this channel using the CSLRS transmitted by the transmitter (PA access point).

[0283] Annex A

[0284] MU-MIMO power allocation algorithm - 1. Initialization of learning rates fie [0,1] V m — 1, ..., M for the constraints M, for example fi — 0.2, V m — 1, .... M, - 2. Initialization of the tolerance threshold, for example g = 10'3, - 3. Initialization of Lagrangian multipliers A to an initial value ^(°), by example ^(0) = 1^, - 4. Initialization of the matrix = owv' - 5. Initialization of the iteration index, i = 0, 6. Calculation £^(0) =diag[Q^\ ...,g^°' j

[0285] with / -^(0) _1 21, 7. As long as jy-M .r _ r1^ [ m~' t J mm 1 I 8. i = i+\ - 9. Calculation of the derivative (gradient) a (0 for all m = 1, ..., M: [°286] ( / ) ..... dg{ / l) v 2, Ah, from the 1st ant~ lWWl - 10. Based on the gradient, calculate the updates to the mul- values Lagrangian tiplieurs for all m — 1, .... M:

[0287] ^> = ^-^ + / 3 A rmr ni m fn

[0288] with fi the learning rate for the Lagrangian multiplier n-CalculgV) = ârg^^ =diag(2*0), ..., 2*0)) with Q

[0289] Q^i} = i__ ln(2)E'f . / / 'j x. / mj m.ii - 12. End of while, - 13. Supply of 2*

[0290] Annex B

[0291] MU-MIMO power allocation algorithm for any pre-coding W - 1. Initialization of learning rates afe [ 0,1 ] VZ = 1, .... v, fie [0,1] V m = 1, ..., M, for the M constraints, for example ^, = 0.2, V77î — 1, - 2. Initialization of the tolerance threshold e, for example f — | (y \ 3. Initialization of Lagrangian multipliers P to an initial value, for example ^(0) = l)xM, 4. Initialization of the matrix ziGO_n, 5^ 5. Initialization of q(°) _ 1 maxEfei|n'„,yp 6. Initialization of the iteration index, i = 0, 7-While [£*: ( t Jmjn lj 8. i = i+ 1, 9. Calculation of the numerical derivative gp) of the Lagangrian L at the value Qj for

[0292] ' \ dQ, / m 10. Calculation of the derivative * (i) for all m = 1, ..., M: There m

[0293]

[0294]

[0295]

[0296]

[0297] 11. Based on the gradient, calculate the updates and the values ​​of the Lagrangian multipliers: / / 9 = / / '-b + / 3 AV m — 1, ..., Mr mrm! m ' and power: 12. End of while, 13. Supply of Q.

Claims

Claims

1. Communication method (1) intended to be implemented by a base station of a MU-MIMO system with U > 2 user equipments (UE), the system comprising M > 2 transmitting antennas, at least v antenna ports and N > 2 receiving antennas distributed in Nur > 1 receiving antennas per user equipment, , using transmission time-frequency radio resource elements (RE) allocated to at least one antenna port assigned to the user equipment 11, among the r antenna ports, 1 u < U, and dedicated to the communication between the base station and the user equipment u, characterized in that: - control information of the communication between the base station and the U user equipments (UE), transmitted by the base station, indicates, for a user equipment u,at least one power ratio (Ppdschi) Pæ" resource element (RE) radio time frequency between reference symbols (DMRS) and data transmitted to this user equipment M for a given antenna port (fi.,

2. Communication method (1) according to claim 1, according to which the control information is transmitted to the user equipment u via a control channel (PDCCH) dedicated to this user equipment M.

3. Communication method (1) according to one of claims 1 and 2, according to which the control information further indicates a configuration of the reference symbols associated with the antenna ports.

4. Communication method (1) according to one of claims 1 to 3, according to which the control information, for the user equipment 11, indicates as many power ratios as antenna ports allocated to the user equipment 11.

5. A communication method (1) according to claim 3, wherein the configuration is such that the antenna ports are grouped to form Ncdm groups such that the reference symbols are spatially code multiplexed between the antenna ports of the same group and wherein the control channel (PDCCH) dedicated to the user equipment w indicates a unique power ratio for each set of ports antennas belonging to the same group and allocated to the user equipment u.

6. Communication method (1) according to one of claims 1 to 5, further comprising a use of the same pre-coding matrix (VF) to pre-code the data and the reference symbols to be transmitted to the user equipment u and the composition of which depends on an estimation of the channels between the base station and each of the U user equipment, all of the channels forming a global channel called MU-MIMO.

7. Communication method (1) according to claim 6, such that the pre-coding matrix (VF) is based on a pseudo-inversion of the MU-MIMO channel.

8. Communication method (1) according to the preceding claim, further comprising a singular value decomposition of the matrix Hu of the channel between the base station and each user equipment u such that, f] = [j V y'- with ye the matrix containing the set of so-called output eigenvectors, VG 'a matrix containing the set of so-called input eigenvectors and = ) e a square matrix containing the positive real singular values ​​arranged on the diagonal coefficients such that < Àit^ £ ■ ■ ■ -, the pseudo inversion being carried out on the matrix „ r T7+ T7t 17t 1T rNxM, called matrix containing the set of Z = 1 V) V2 ... Ff / I And so-called input eigenvectors of the MU-MIMO channel.

9. Communication method (1) according to claim 6, such that the pre-coding matrix (VF) is based on a regularized version of the inversion of the MU-MIMO channel.

10. Communication method (1) according to one of claims 1 to 9, further comprising a determination of an allocated power (Qz) to the data allocated to antenna ports and intended for the user u to maximize a sum of the user rates, under the constraint of a maximum power per transmit antenna, by alternating between a maximization of a Lagrangian L with respect to the power matrix Q for a given value of P and then, after an update of the value of the vector P taking into account its gradient, a maximization of the Lagrangian L with respect to P, with .. _ [., ■, 1 ” [Rp • • • ' HmJ vector of Lagrangian multipliers corresponding to the M power constraints per antenna.

11. Communication method (1) according to claim 10, such that the determination of a power allocated (Q;) to the data further takes into account an EMF constraint at at least one point in the space defined by a distance to the base station.

12. Communication method (2) intended to be implemented by a user equipment taken from among U user equipments (UE) of a MU-MIMO system with U > 2 user equipments (UE) and a base station, the system comprising M > 2 transmitting antennas, v antenna ports and N > 2 receiving antennas distributed into > 1 receiving antennas per user equipment, y _ using transmission time-frequency radio resource elements (RE) allocated to at least one antenna port assigned to the user equipment w, from among the v antenna ports, 1 u < U, and dedicated to the communication between the base station and the user equipment M characterized in that: - information for controlling the communication between the base station and the user equipment (UE), transmitted by the base station,indicate at least one power ratio per radio time frequency resource element (RE) between reference symbols (DMRS) and data transmitted to this user equipment 11 for a given antenna port (Z).,

13. Communication method (1) intended for a MU-MIMO system with U £ 2 user equipments (UE), the system comprising M > 2 transmit antennas, at least v antenna ports and N > 2 receive antennas distributed in NUR > 1 receive antennas per user equipment, y _ Vy», characterized in that the power allocation distinguishes between power allocated to the data and power allocated to the reference symbols intended for one of the user equipments and takes into account a maximum power per transmit antenna.

14. Access point (PA) comprising M transmitting antennas (ANT_E), at least v antenna ports, a transmitter (EM1), a receiver (RE1), a computer (pPl), such that the transmitter (EM1) is capable of transmitting: data (TBu) to U >2 user equipments, control information indicating radio time frequency transmission resource elements (RE) allocated to at least one antenna port, among the r antenna ports, assigned to a user equipment u,, l <u< U, characterized in that the transmitter (EM1) is further capable of transmitting: via a control channel (PDCCH) dedicated to the user equipment Zi, at least one power ratio (Ppdschi) Pæ" resource element (RE) time frequency between reference symbols (DMRS) and data transmitted to this user equipment 11 for a given antenna port ( / ).

15. Telecommunications terminal (UEu) comprising > 1 receiving antennas (RXu), a receiver (RE2), such that the receiver (RE2) is capable of receiving: reference symbols (DMRS) transmitted by an access point (AP) having at least v transmitting antenna ports, control information indicating radio time frequency transmission resource elements (RE) allocated to at least one antenna port assigned to the terminal among the v antenna ports, transmitted by the access point, characterized in that the receiver (RE2) is further capable of receiving: via a control channel (PDCCH) dedicated to this terminal, at least one power ratio (fi pose ni) Pæ" resource element (RE) time frequency, between reference symbols (DMRS) and data transmitted to this terminal for a given antenna port (1), transmitted by the access point.

16. Digital signal received by a terminal, comprising reference symbols (DMRS) transmitted by an access point (AP) having at least v antenna ports, control information indicating at least one power ratio (PPDSchÎ) Pæ" radio resource element (RE) transmission frequency time allocated to at least one antenna port assigned to the terminal among the v antenna ports, between reference symbols (DMRS) and data transmitted to this terminal for a given antenna port ( / ).

Images