Data transmission method in an OMAMRC system and associated devices

Abstract

Data transmission method in an OMAMRC system and associated devices The invention relates to a data transmission method comprising the following steps, following an inability of the destination device to decode messages emitted successively by source devices and an inability of the destination device to decode at least one message received from at least one source device called "undecoded source device", each message emitted by a source device comprising a first redundancy version resulting from an encoding of an information message associated with the source device: a selection of a set comprising several undecoded source devices and a set of several relay nodes associated with the set, according to a quality criterion of a communication channel; a transmission, to the selected relay nodes, of a retransmission instruction;and a reception of the second redundancy versions resulting from the encoding of the information messages from the source devices. Figure for the abbreviation: Fig. 3.;

Description

Title of the invention: Method for transmitting data in an OMAMRC system and associated devices technical field

[0001] The present invention belongs to the general field of digital communications. It relates more particularly to the transmission of data between source devices and a destination device having several receiving antennas, via a relay node.

[0002] As discussed in more detail below, an orthogonal multiple-access scheme, denoted OMAMRC ("Orthogonal Multiple-Access Multiple-Relay Channel" according to Anglo-Saxon terminology), is for example used for transmissions between relaying nodes and the destination device.

[0003] The invention finds a particular, but not limiting, application in the transmission of data via mobile networks, for example from sensors capable of autonomously collecting and transmitting data to a base station (uplink communication). The network then comprises, for example, several source devices corresponding to sensors, several relay nodes, and a destination device corresponding to a base station.

[0004] The invention also finds a particular application in the transmission of data between several nodes of an ad hoc wireless network. The network then comprises several nodes, some of which act as source devices, relay nodes, or destination devices. Previous technique

[0005] Conventionally, the source devices of an OMAMRC-type communication system transmit messages to a single destination, if necessary via relay nodes. The relay nodes are, for example, of the "Half Duplex" type (i.e., a relay node cannot receive and transmit simultaneously) and receive messages from the source devices before decoding them. When retransmission of these messages is required by a destination device, the relay nodes access the communication channel linking these nodes to a destination device using an orthogonal multiple access scheme to transmit the messages to the destination device, thereby limiting interference.This orthogonality is usually obtained by time multiplexing in the form of disjoint time intervals (such as "Time Division Multiple Access", TDMA), but orthogonality resulting from frequency multiplexing in the form of sub-bands of . Disjoint frequencies (such as "Frequency Division Multiple Access", FDMA) is also conceivable.

[0006] The document "Efficient Cooperative HARQ for Multi-Source Multi-Relay Wireless Networks", Cerovié, S. & Al., In 2018 14th International Conference on Wireless and Mobile Computing, Networking and Communications (WiMob) proposes a version of the OMAMRC protocol which considers an incremental redundancy retransmission of the HARQ type ("Hybrid Automatic Repeat reQuest").

[0007] Fig. 1 is a representation of a frame transmission cycle in an OMAMRC type communication system.

[0008] The data transmission cycle in an OMAMRC type communication system comprises three phases, an initial INIT phase and, for each TF frame to be transmitted, a first PI phase and a second P2 phase.

[0009] During the INIT initialization phase, the destination device determines a number of bits 9 carried by a modulation and a coding efficiency R for the first transmission from each source, according to a quality criterion of the communication channel.

[0010] To do this, the destination device determines a representative value of a quality, for example the signal-to-noise ratio (SNR) of the direct links between this destination device and relay nodes using known techniques based on the use of reference signals (e.g., pilot symbols, SRS sound reference signals as defined by the 3GPP LTE consortium). During a retransmission, a relay node is associated with a source device and may correspond either to the source device to which it is associated, to another source device, or to an intermediate node located between the source device and the destination device.Similarly, source devices and intermediate nodes determine representative quality values ​​for source / source, intermediate node / intermediate node, and source / intermediate node links, which are then transmitted to the destination device.

[0011] Based on these representative quality values, the destination device determines, for each source device, a coding efficiency R and a number of bits carried by a modulation for a number A of channel uses ("Resource Element" according to the terminology of the 3GPP consortium or "channel use" according to terminology inherited from information theory). This data is then transmitted, by the destination device, to each of the source devices.

[0012] Thus, the size L of the message to be transmitted by a source device is defined as follows: L - NY x R x q. The message is encoded with a low-yield mother code Rq. (for example, 1 / 3), and this coded message of length Lc — L / Rq > L is subsequently stored in a circular buffer, an example of which is described with reference to [Fig. 2].

[0013] The data transmission cycle in an OMAMRC-type communication system further includes a first PI phase during which the M sources successively transmit their message during M time slots, respectively using the modulation and coding schemes determined during the INIT initialization phase. During this first PI phase, the number of times the channel is used is fixed and identical for each of the source devices.

[0014] In response to a determination that one or more messages transmitted by these source M devices have not been decoded by the destination, the second phase P2 is implemented, during which the undecoded messages are cooperatively retransmitted by the relay nodes. A relay node associated with source devices then retransmits the undecoded messages from these source devices at the end of the first PI phase, so as to allow the destination device to decode the messages from all the source devices without error. More precisely, this relay node transmits to the destination device a redundant version of a message from a source device that it has correctly decoded and that had not been decoded by the destination device at the end of the first PI phase.Thus, when the relay node is separate from the source device to which it is associated, said relay node must have received the message transmitted by the source device during the first PI phase, then re-encode the received message and store it in its own circular buffer. This second phase P2 lasts a maximum of TMAX time intervals. During this phase, the number N2 of channel uses is fixed and identical for each of the relay nodes.

[0015] Finally, if the messages from all source devices are decoded without error by the destination device within a time interval TUSED < TMAX, the destination device broadcasts an acknowledgment. In this case, a new frame transmission cycle begins with the erasure of the memories of the relay nodes and the destination device, and with the transmission of new messages by the source devices.

[0016] Figure [Fig.2] represents a circular buffer allowing selection of redundancy of the message to be transmitted.

[0017] As mentioned previously, during the INT initialization phase, a message encoded with a yield Rq is stored in a circular buffer. As illustrated in Figure 2, this circular buffer has several read start positions POS0, POS1, P2, and POS3. Such a circular buffer contains the encoded bits of a message of a source device encoded by a low yield mother code / ?o (eg, 1 / 3) and allows selection of a particular redundancy of the message to be transmitted, depending on a starting reading position.

[0018] Indeed, these starting read positions POS0, POS1, POS2, and POS3 are associated with different redundancy versions: RVO, RV1, RV2, and RV3. In the chosen example, there are four possible redundancy versions. For each redundancy version, a node reads the number of coded bits to be sent. This number corresponds to the number of channel uses for a given modulation and message size, starting from the corresponding redundancy position, by moving through the circular buffer in the direction of the initial filling (in this example, clockwise). The selected coded bits are then interleaved and modulated. The first redundancy version, RVO, can be decoded independently of the other versions.

[0019] Thus, during the first PI phase, a source device transmits a TX1 message corresponding to the first RVO redundancy. To do this, this source device determines the bits to be transmitted by reading, in the circular buffer, the N fl bits encoded from position POS0.

[0020] During the second phase P2, the relay node determines the bits to be transmitted during a first retransmission TX2 by reading, from the circular buffer, the N2q bits encoded from position POS2, where N2 is the number of possible uses of the channel for retransmissions. If necessary, this source device then determines the bits to be transmitted during a second retransmission TX3 by reading, from the circular buffer, the N2q bits encoded from position POS3, and the bits to be transmitted during a third retransmission TX4 by reading, from the circular buffer, the N2q bits encoded from position POS1. Thus, the sequence of transmissions for a source s> is, for example, as follows: TX1, TX2, TX3, TX4, TX1, TX2, TX3, TX4, ...

[0021] However, existing methods do not focus on improving spectral efficiency, particularly when the relay nodes include a plurality of transmitting antennas and the destination device includes a plurality of receiving antennas. Description of the invention

[0022] The present invention aims to remedy all or part of the disadvantages of the prior art, in particular those set out above, by proposing a solution which makes it possible to exploit spatial multiplexing, i.e., the possibility of simultaneously transmitting several spatial layers on the same time or frequency radio resource.

[0023] To this end, and according to a first aspect, the invention relates to a data reception method comprising the following steps, implemented by a destination device comprising several receiving antennas, following the reception of M > 2 messages emitted successively by M > 2 source devices and the inability of the destination device to decode at least two messages received from at least two source devices, referred to as "undecoded source devices", each message emitted by a source device comprising a first redundancy version resulting from an encoding of an information message from the source device:

[0024] - a selection, from a set 5* comprising several non-source devices decoded and a set Bj of several relay nodes associated with set 5*, according to a quality criterion of a communication channel between the relay nodes of set B, and the destination device, a relay node of set B, corresponding to a source device or an intermediate device knowing the information messages of the undecoded source devices, at least one relay node of set B, having several transmitting antennas;

[0025] - a transmission, to the selected relay nodes, of an instruction to retransmission, so that the relay nodes simultaneously transmit second redundancy versions resulting from the encoding of information messages from the undecoded source devices of set 5*; and,

[0026] - a reception, by the destination device, of said second versions of redundancy arising from the encoding of information messages from the source devices of the set, the said second redundancies being used by the destination device to decode the messages received from the source devices of the set S*.

[0027] By allowing several relay nodes to simultaneously transmit redundant messages from source devices that have not been decoded without error, and by allowing the destination device to receive this data via its multiple receiving antennas, the invention improves known methods. In particular, the simultaneous transmission of data between one or more relay nodes having multiple transmitting antennas and a destination device having multiple receiving antennas makes it possible to exploit several spatial layers, and thus improves spectral efficiency.

[0028] A relay node in set B* is aware of a message to be retransmitted, either because it is its own message (and the relay node fa* in this case corresponds to a source device), or because the relay node has previously decoded it correctly. In the following description, a relay node is said to be associated with a set of source devices when this node of relayage has previously decoded the information messages from all source devices in this set, and is therefore able to retransmit new redundancy versions of the messages from all source devices in this set.

[0029] In a particular embodiment, an orthogonal multiple-access multiple-relay channel scheme, denoted OMAMRC (Orthogonal Multiple-Access Multiple-Relay Channel), is applied to the transmission channel between the relay nodes and the destination. The system then implements a cooperation strategy called IR-HARQ (Incremental Redundancy Hybrid-ARQ) based on selective relaying known as SDF (Selective Decoded and Forward).

[0030] Generally speaking, the steps of a process should not be interpreted as being linked to a notion of temporal succession.

[0031] In particular embodiments, the transmission method may further comprise one or more of the following characteristics, taken individually or in all technically possible combinations.

[0032] In a particular embodiment, the transmission method is iterated until a stopping criterion is reached (for example, a certain duration TMAX or a certain number of iterations). The use of a stopping criterion offers the advantage of limiting the use of the communication channel, particularly when the latter is noisy.

[0033] In a particular embodiment, the method further includes a transmission, by the M source devices, of first redundancies (also referred to as "redundancy versions") of their information messages, during M consecutive time intervals.

[0034] In a particular embodiment, each of the source devices transmits a first redundancy of its information message using a single antenna in omnidirectional transmission or after applying omnidirectional precoding to the antennas of that source device. This step is implemented within the PI phase previously described with reference to [Fig. 1]. This feature is advantageous since it increases the number of relay nodes in the system capable of receiving and decoding these messages.

[0035] In a particular embodiment, the retransmission instruction transmitted to the relay nodes includes the set 5* and a precoding matrix V associated with the source devices of the set 5*.

[0036] In a particular embodiment, the retransmission instruction further includes an ordered list b of the relay nodes of set B. Transmission of the list b to the relay nodes allows each of the nodes to relaying, to identify the precoding coefficients of the precoding matrix V to be applied.

[0037] In a particular embodiment, the retransmission instruction transmitted to the relay nodes further includes an indication of a distribution of a transmission power transmitted by each relay node Q,*) of the set Bp between the undecoded source devices (.^) of the set £*.

[0038] According to a particular implementation, this indication takes the form of an ordered list P of powers to be applied to each of the source devices in the set 5*. When the powers allocated to the selected source devices s* e 5* are different, this ordered list allows a relay node to determine the power it must allocate to each source device.

[0039] In a particular embodiment, the retransmission instruction transmitted to the relay nodes further includes an ordered list a of antennas to be activated or deactivated by each of the relay nodes of at least a subset of the set Bp. The activation or not of certain transmitting antennas makes it possible to optimize a channel quality criterion, since only the strongest transmitting antenna / destination device links are considered.

[0040] In a particular embodiment, this quality criterion is a mutual information of the communication channel between the relaying nodes Q,*) of the set B^ and the destination device (d).

[0041] In the following description, this "mutual information" is also referred to as "equivalent mutual information" when it relates to the quality of a multilayer channel. Of course, other quality criteria can be considered, such as a receive bit error rate. Furthermore, during selection, one can seek to optimize such a quality criterion or, alternatively, ensure that this quality criterion reaches at least a given threshold value.

[0042] In a particular embodiment, the selection includes:

[0043] - a determination of eligible sets of undecoded source devices and of sets B of relay nodes associated with eligible sets such as 2 < ISI < N?h with the cardinality operator and the total number of transmitting antennas of the relay nodes of set B;

[0044] - a determination, for each of the combinations of eligible sets S and of sets B of relay nodes, of the quality criterion of the communication channel between the relay nodes of set B and the destination device (d);

[0045] and the sets $* of source devices ) and B," of selected relaying nodes optimizing said quality criterion.

[0046] Thus, in a particular embodiment, an exhaustive search of eligible set combinations 5 and B is carried out, and the one providing optimal quality is selected.

[0047] In a particular embodiment, the selection further includes, for each of the eligible subsets $ of source devices, a step of determining the relay nodes in set B that have decoded all the source devices in the eligible set $ under consideration. When the communication channel between the relay nodes and the destination device is known (e.g., when the links between the relay nodes and the destination device are direct links), considering relay nodes that have decoded all the source devices and not just some of them offers the advantage of eliminating interference at the receiver.

[0048] In a particular embodiment, a maximum number QMAX of source devices whose messages can be simultaneously retransmitted by the relay nodes is parameterized, and the determination of eligible sets S of source devices is limited to subsets such that I SI Q^ax- This feature offers the advantage of reducing the number of iterations required to determine the eligible sets S, and therefore a fortiori reducing the time complexity of the process.

[0049] In a particular embodiment, the determination of the quality criterion for the communication channel between the relay nodes of assembly B and the destination device includes:

[0050] - a calculation, for each source device of the eligible set S, of a criterion of single-source quality for each source device in the eligible set S, based on a precoding matrix and an emission power to be allocated to said source device; and,

[0051] - a calculation of the quality criterion of the communication channel, based on the criteria single-source quality calculated for each source device in the eligible set S.

[0052] In a particular embodiment, the method further includes a selection of antennas to be activated for each of the relay nodes of at least one subset of the set Bp. This step can be implemented if the total number Ntb of transmitting antennas of the relay nodes of the set B is greater than the number of source devices of the set 5*.

[0053] In a particular embodiment, the antenna selection includes:

[0054] - a determination of subsets A of antennas that can be activated by each of the relay nodes of at least one subset of the set Bp

[0055] - for each determined subset Ât of antennas, a calculation of the quality criterion of a communication channel between the antennas of said sub-assembly Ât, and the destination device;

[0056] - a selection, among the subsets Ât, of the antenna subset to activate optimizing said quality criterion.

[0057] In a particular embodiment, the quality criterion of the communication channel between the relay nodes of the assembly and the destination device is determined according to a combination of single-source quality criteria of sub-channels of said communication channel, each sub-channel being defined for a source device of the assembly 5*.

[0058] According to a second aspect, the invention relates to a data transmission method comprising the following steps, implemented by a relay node comprising several transmitting antennas, following the reception of M > 2 messages emitted successively by M > 2 source devices and the inability of the destination device to decode at least two messages received from at least two source devices, referred to as "undecoded source devices", each message emitted by a source device comprising a first redundancy version resulting from the encoding of an information message associated with the source device:

[0059] - receiving an instruction to retransmit second versions of redundancy arising from the encoding of information messages from undecoded source devices of a set S*, said relay node knowing said information messages from the undecoded source devices of the set S*', and,

[0060] - a simultaneous transmission of the second redundancy versions of the messages information from the source devices of set g* by the transmitting antennas of said relaying node and to the destination device, said second versions of redundancies being used by the destination device to decode the messages received from the undecoded source devices of set 5*.

[0061] According to a third aspect, the invention relates to a computer program comprising instructions for implementing a receiving method according to the first or a transmission method according to the second aspect, when said program is executed by a processor.

[0062] This program may use any programming language, and be in the form of source code, object code, or code intermediate between source code and object code, such as in a partially compiled form, or in any other desirable form.

[0063] According to a fourth aspect, the invention relates to a computer-readable recording medium on which the computer program according to the invention is recorded.

[0064] The information or recording 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 hard disk drive.

[0065] On the other hand, the information or recording medium can be a transmissible medium such as an electrical or optical signal, which can be transmitted via an electrical or optical cable, by radio, or by other means. The program according to the invention can, in particular, be downloaded onto an Internet-type network.

[0066] Alternatively, the information or recording medium 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.

[0067] According to a fifth aspect, the invention relates to a destination device (or "destination node" or "destination") comprising several receiving antennas and configured to implement a receiving method according to the invention.

[0068] According to a sixth aspect, the invention relates to a relay node comprising several transmitting antennas and configured to implement a transmission method according to the second aspect.

[0069] According to a seventh aspect, the invention relates to a system comprising M > 2 source devices, a destination device according to the fifth aspect and a plurality of relaying nodes of which at least one conforms to the sixth aspect.

[0070] According to an eighth aspect, the invention relates to a communication method including the data reception and data transmission methods previously mentioned, for example implemented by the system conforming to the sixth aspect. Brief description of the drawings

[0071] Other features and advantages of the present invention will become apparent from the description below, with reference to the accompanying drawings, which illustrate an example of an embodiment without being limiting in any way. In the figures:

[0072] [Fig-1] The [Fig. 1], previously described, is a representation of a frame transmission cycle in an OMAMRC type communication system;

[0073] [Fig.2] Fig.2, previously described, represents a circular buffer allowing select a redundancy of a message to be transmitted;

[0074] [Fig.3] [Fig.3] schematically represents a communication system according to a particular embodiment of the invention;

[0075] [Fig.4] [Fig.4] represents modules embedded in a destination, according to a example of implementation of the invention;

[0076] [Fig.5] [Fig.5] schematically represents an example of hardware architecture of a destination device;

[0077] [Fig.6] [Fig.6] represents, in the form of a flowchart, a particular mode of implementation of a communication process in an OMAMRC system, according to a first example.

[0078] [Fig.7] Figure 7 represents, in the form of a flowchart, a particular method of implementing a step of selecting a set g* of source devices whose messages are to be retransmitted simultaneously to the destination device and a set Bz of relaying nodes adapted to this retransmission;

[0079] [Fig.8] [Fig.8] represents, in the form of a flowchart, a particular method of implementing a method for calculating equivalent mutual information of a channel linking the transmitting antennas of the relaying nodes to the receiving antennas of the destination device;

[0080] [Fig.9] [Fig.9] represents, in the form of a flowchart, a particular mode of implementation of a selection of antennas of a relay node to activate (resp. to deactivate);

[0081] [Fig. 10] [Fig. 10] represents, in the form of a flowchart, a particular method of implementing a communication process in an OMAMRC system, according to a second example; and,

[0082] [Fig. 11] [Fig. 11] represents, schematically, an example of implementation of the communication process in an OMAMRC system. Description of the implementation methods

[0083] In general, the invention proposes cooperation between different devices of a communication system to enable a destination device to correctly decode the information messages from the source devices of that system. To this end, the invention relies on selecting a set of several source devices to be "assisted" and a set of several relay nodes adapted to these source devices. However, the invention is described in a more general context, and more particularly is integrated in the following description into an algorithm that can, in certain cases, select a set Bz comprising a single multi-antenna relay node, for example, when this configuration offers optimal quality.

[0084] Figure 3 schematically represents a communication system according to a particular embodiment of the invention.

[0085] As illustrated in Figure 3, the SYS communication system comprises three sources si, s\ s\ a destination device d and two intermediate nodes r2 positioned between the source devices and the destination. An intermediate node se It differs from a source device in that it does not have its own message to transmit. It only retransmits (relays) messages from other nodes.

[0086] In the present embodiment, and for the sake of simplicity in the description, the communication system is considered to have a single destination d. It should be noted, however, that there is no limitation on the number of destinations d. The following developments can indeed be easily generalized by those skilled in the art to cases where more than one destination device is considered. It should be noted that the number of source devices is not a limiting factor of the invention, provided that this number is greater than or equal to 2. Finally, there is no limitation on the number L of intermediate nodes, which only need be such that L > 0. Thus, in a particular implementation, the SYS communication system does not include any intermediate nodes.

[0087] In the following description, the set of nodes b comprising the intermediate nodes ri> and source devices ^i' ■ • - ' sm operating as relays responsible for retransmitting messages from other nodes to the destination is called the "set N of relaying nodes". Thus, each of these sources si, s2, s3 can operate, at different times, either exclusively as a source device or as a relay responsible for retransmitting messages from other nodes to the destination.

[0088] Furthermore, at least one of the relay nodes comprises several transmitting antennas. Hereafter, the number of transmitting antennas of a node be N is denoted

[0089] For the sake of simplification, it is also assumed that the source devices and intermediate nodes are equipped with a single receiving antenna; that the source devices, intermediate nodes and destination device are perfectly synchronized; and that the source devices are statistically independent, i.e. there is no correlation between them.

[0090] Furthermore, the antennas of all the nodes of the system are co-phased before any transmission. A phase calibration procedure is, for example, carried out periodically, and can be initialized by the destination device d.

[0091] In a particular embodiment, an orthogonal multiple-access scheme, denoted "OMAMRC" - an acronym for "Orthogonal Multiple-Access Multiple-Relay Channel" - is applied. The system then implements a cooperation strategy called IR-HARQ ("Incremental Redundancy Hybrid-ARQ" according to Anglo-Saxon terminology) based on selective relaying called SDF ("Selective Decoded and Forward" according to Anglo-Saxon terminology).

[0092] Thus, the M source nodes and the L intermediate nodes access the transmission channel according to an orthogonal multiple access scheme that allows them to listen, without interference, to the transmissions of the other source nodes and the other intermediate nodes. In the following description, it is assumed that this orthogonality is obtained by time division multiplexing in the form of disjoint time intervals (such as Time Division Multiple Access, TDMA). However, orthogonality resulting from frequency division multiplexing in the form of disjoint frequency sub-bands (such as Frequency Division Multiple Access, FDMA) is also conceivable.

[0093] Figure 4 represents modules embedded in a destination device, according to an example of an implementation of the invention.

[0094] As illustrated in [Fig. 4], the destination device d includes, in particular:

[0095] - a MOD_SEL module of a set 5* comprising several source devices undecoded and of a set B, of several relay nodes associated with the set S*, according to a quality criterion of a communication channel between the relay nodes of the set and a destination device, a relay node of the set B, corresponding to a source device or to an intermediate device knowing the information messages of the undecoded source devices, at least one relay node of the set B, having several transmitting antennas;

[0096] - a M0D_TX transmission module, to the selected relay nodes, of a retransmission instruction, so that the relay nodes simultaneously transmit second redundancy versions resulting from the encoding of information messages from the undecoded source devices of set 5*;

[0097] - a M0D_RX module for receiving, by the destination device, said second versions of redundancy resulting from the encoding of information messages from source devices (sJ) of the set £*, said second redundancies being used by the destination device (d) to decode messages received from undecoded source devices.

[0098] Their functionalities are described in more detail below with reference to different modes of implementation.

[0099] Figure 5 schematically represents an example of the hardware architecture of a destination device

[0100] As illustrated in Figure 5, the destination device d has the hardware architecture of a computer. Thus, the destination device d includes, in particular, a processor 1, random access memory 2, read-only memory 3 and non-volatile memory 4. It also has communication means 5.

[0101] The read-only memory 3 of the destination device d constitutes a recording medium according to the invention, readable by the processor 1 and on which is A computer program, PROG, according to the invention, has been recorded, comprising instructions for executing steps of the receiving process according to the invention. The PROG program defines functional modules of the destination device that rely on or control the hardware elements 1 to 5 of the destination device d mentioned above. These functional modules are illustrated in [Fig. 4] by way of no limitation, and are described in more detail below with reference to different implementation methods.

[0102] In the implementations described below, the communication means 5 enable the destination device to obtain the messages transmitted by the source devices during the first phase, as well as the messages retransmitted by the relay nodes during the second phase. The communication means 5 also enable the destination device to transmit the retransmission instruction to the relay nodes. To this end, the communication means 5 include a wired or wireless communication interface capable of implementing any suitable communication protocol.

[0103] Fig. 6 represents, in the form of a flowchart, a particular method of implementing a communication process in an OMAMRC system, according to a first example.

[0104] This communication method includes a data reception method implemented by a destination device (d) and comprising steps S300 to S380, and a data transmission method implemented by a relaying node (b) and comprising steps S210 to S240.

[0105] As illustrated by Figure 6, the communication method includes a first link initialization and adaptation step S300 during which the destination device d determines a number of bits 9 carried by a modulation and a coding efficiency R for the first transmission of each source si' • • • ' sm, according to a communication channel quality criterion.

[0106] To this end, the destination device determines a representative quality value, for example the signal-to-noise ratio of the direct links between this destination device and the relay nodes (e.g., between the destination device and the source devices, but also between the destination device and the intermediate nodes) using known techniques based on the use of reference signals (e.g., pilot symbols, SRS sound reference signals as defined by the 3GPP LTE consortium). Similarly, the source devices and intermediate nodes determine representative link quality values ​​between two sources, two intermediate nodes, and / or between a source device and an intermediate node, which are then transmitted to the destination.

[0107] Based on these representative quality values, the destination device determines, for each source device, a coding efficiency R and a number of bits 9 carried by a modulation for a number of uses of the channel. This data is then transmitted, by the destination device to each of the source devices • • • ' sm.

[0108] During this same step S300, each source device sv • - ■ ' sm encodes a message of size L = N} x R xq to be transmitted to the destination device d with a low yield mother code Rq (for example of 1 / 3), and the coded message of length Lc — L / Rq > L is subsequently stored in a circular buffer, an example of which is described with reference to [Fig.2]

[0109] The communication method further includes a step S100 during which the M source devices si' • • • ' sm successively transmit their message during M consecutive time intervals, using respectively the modulation and coding schemes determined during step S300. During this first phase, the number A of channel uses is fixed and identical for each of the source devices si' •••' sm. The redundancy versions RV0}, RVO^ transmitted by the source devices sr • • • ' sm during this step S100 correspond to the first redundancies RV0 of their messages.

[0110] These initial redundancies are transmitted, for example, using a single omnidirectional transmitting antenna or after applying omnidirectional precoding to the transmitting antennas. Thus, for a source device comprising N transmitting antennas, the precoding results, for example, from a superposition of discrete Fourier transform vectors of dimension N. This characteristic is advantageous since it increases the number of relay nodes in the system capable of receiving and decoding these messages.

[0111] At least some of the M messages ÆV0t, ..., RV0M are received by the destination device during a step S310. During a step S320, the destination device d determines the set of source devices whose messages have been decoded without error during the time interval t- 1. The method further includes a step S330 during which the destination device transmits, to the relay nodes 6 e N, an ACK / NACK message to indicate whether it has successfully decoded all the source devices or not.

[0112] In the following description, a source device whose message could not be decoded by the destination device is called an "undecoded source device". If at least one message from a source device could not be decoded by the destination device (i.e., if a NACK message is received), each relay node transmits its respective decoding set S^ to the destination device d. This set of The decoding S^_i corresponds, for a relay node b GN, to the source devices whose messages have been decoded without error by that relay node. Each of the decoding sets is received by the destination device d during an S340 step. Thus, at a given interval / gf, the destination device knows the decoding set of all the relay nodes bg € N.

[0113] The communication method further includes a step S350 in which the destination device selects, based on a quality criterion, a communication channel between the antennas of said subassembly Aj and the destination device:

[0114] - a set g* among the undecoded source devices comprising several source devices s*, and

[0115] - a set B* of relay nodes associated with the set S* from among a set B of bt relay nodes.

[0116] This S350 step is implemented, for example, by the MOD_SEL module described with reference to [Fig. 4]. A particular implementation method for this S350 selection step is described with reference to [Fig. 7].

[0117] A relay node corresponds to a network node that has correctly decoded messages that were not decoded by the destination device during a previous time interval Fl, and is therefore able to transmit, to the destination device, a new redundant version of these messages during the retransmission time interval z. This previous time interval Fl does not necessarily correspond to the time interval immediately preceding the time interval f. Let be the decoding set of a node b at the time interval t-1. The latter contains all the identifiers of the source devices that node ba was able to decode without error before the retransmission interval f. Thus, a relay node b is therefore a potential candidate for transmitting the second redundancy versions if the intersection between its decoding set and the complement of the decoding set of the destination device is different from the empty set, i.e., si ^d / -i * .

[0118] The communication method further includes an S360 step during which the destination device calculates a precoding matrix V = [y V ] E to be applied by each of the relaying nodes fy* eg*, with A the set of complex numbers, NTS* the total number of transmitting antennas of the relay nodes fy* eg*, and v the number of source devices s* of which The messages must be retransmitted simultaneously. In other words, v-, with the cardinality operator.

[0119] During this same S360 step, the destination device d transmits, to the relay nodes eg*, the set Ej of identifiers of the source devices whose messages are to be retransmitted and the precoding matrix V. This S360 step is for example, implemented by the M0D_TX module described with reference to [Fig. 4].

[0120] According to a particular implementation, the destination device transmits also the vector . , , , .t tj*^' with the cardinality operator and P- [Pj, 0iB*|] G [M] r'a transpose of the matrix M. This vector b includes the selected relay nodes eg*, so that each relay node can determine the precoding coefficients to apply from each vector el = 1 y of the precoding matrix V. The relay nodes listed by the vector b are by examples are ranked according to their index, in ascending order. Knowing the other active nodes and their number of antennas, each relay node determines then determines the position of the precoding coefficients to be applied, then applies the first coefficient associated with it to its first antenna, and the second to its second antenna. etc.

[0121]

[0122] In a particular implementation mode, each of the selected source devices e sJ applies the same power Pi = P^ 1=1, ..., v. Alternatively, the selected source devices es* apply different powers, and the destination device d also transmits, during this S360 step, a vector p_pj containing the power to be applied for each source device y* g S*; According to a particular implementation, the powers P, Py are ranked in ascending order of indices of the source devices s?. The set £* of source devices s?, the precoding matrix V, the vector b, and optionally the vector P are received by the nodes eg* during an S230 step. In response, each relay node eg* simultaneously transmits the same redundancy of messages from the v selected source devices, i.e., E'' „ „ ._ li xj ] fn^Uors during an S240 step. In order to facilitate For clarity and conciseness, the index “n + j^i), l” is described in more detail with reference to Figure 8. These redundancies are received by the receiving antennas of the destination device d during an S370 step.

[0123] Based on this retransmission, the destination device attempts to decode the messages from the source devices e Sf during an S380 step.

[0124] In a particular embodiment, the communication process is iterated until a stopping criterion is reached. This stopping criterion corresponds, for example, to a duration or a number of iterations reached. Alternatively, the communication process is iterated until all messages are decoded without error by the destination.

[0125] Figure 7 represents, in flowchart form, a selection process corresponding to a particular implementation of the S350 selection step, at a certain interval f of a set of source devices whose messages are to be retransmitted simultaneously to the destination device and a set of relaying nodes adapted for this retransmission. This selection process is implemented by the destination device d.

[0126] As illustrated by Figure 7, this selection process includes a first step S510 during which the destination device d calculates the set parts of the system are limited to subsets with a cardinality less than or equal to the number of NR receiving antennas and distinct from the empty subset. This constraint on the cardinality of the subsets aims to address the assertion that it is undesirable to simultaneously retransmit more streams than there are receiving antennas, in order to limit interference.

[0127] It is important at this stage to recall that the power set of a set A, denoted H(A), corresponds to the set of all possible subsets of A.

[0128] Subsequently, the subsets of cardinality ge (1 . / VjJ are denotedllv ofSjyi) telquen (ç \_|^rr ( cv

[0129] The selection process further includes a determination step S520, for each subset of source devices G i) 2V}' the set of relay nodes B that have decoded the messages from all the source devices of £. Considering the relay nodes that have decoded all the source devices and not just some of them has the advantage of eliminating interference at the receiver. During this same step, only the eligible subsets S of and of B are selected such that IEI < with - Y, the total number of transmitting antennas in set B of the 11 J, B relay nodes considered.

[0130] Step S520 is iterated for different values ​​of corresponding qe to different cardinalities of the set of sources. According to an implementation Specifically, if no subset is eligible for a value Qw, the other values ​​Q > Qq are not tested. Indeed, if no eligible subset S is determined for this value Qo, no eligible subset S can be determined for a value Q > Qq.

[0131] The selection process further includes a step S530 in which a representative value of a quality criterion of a communication channel is calculated, for one of the parts of the set B of relay nodes, B being the set of relay nodes knowing S.

[0132] This value represents a corresponding quality criterion, in this particular implementation, to a mutual information known as "equivalent mutual information," which characterizes the quality of a communication channel through which several messages are transmitted simultaneously by a plurality of transmitting antennas and received by a plurality of receiving antennas. An example of calculating equivalent mutual information is described in more detail with reference to [Fig. 8].

[0133] According to a particular implementation, this step S530 is repeated for each of the parts of set B, and each source set S.

[0134] Finally, an S540 step is implemented in which the destination device d selects the set L of sources, the part of the set B of relay nodes, and the set Âf of transmitting antennas that optimizes the quality criterion (e.g., equivalent mutual information). This S540 step is implemented, for example, by the MOD_SEL module described with reference to [Fig. 4].

[0135] In a particular embodiment, the selected combination corresponds to the one that maximizes or minimizes this quality criterion, depending on how this quality criterion is defined. Alternatively, the selected combination corresponds to one of the possible combinations of parts providing a representative value of the quality criterion that is above (respectively below) a certain threshold.

[0136] The set £ of sources selected at this stage is denoted and the set B of relay nodes is denoted B}-

[0137] Illustrative example

[0138] Steps S510 to S530 of this selection process are illustrated below by means of an example. Let the set of source devices whose messages have not been decoded without error by the destination device be such that g} _ || 4 5}, and NR = 3, Nf 1 = 2, N= 3, = 1. In this case, the set ) of the parts of ] is written as = {|1}, {4}, {5}, |1,4], {1,5}, {4,5}, {1,4,5}[

[0139] If S = {4,5} and B = [5] then L is not eligible because ÎEI > Ntb = Nt,5 ~ 1-

[0140] If S = {4,5} and B = {4} then S is eligible even though there is only one active node because IE1 < Nt# = Nta = 3.

[0141] E = {1,4, 5} and B = {1} then E is not eligible because ILI > -2.

[0142] E = {1,4, 5} and B = {1,5} then E is eligible although there are only two active nodes to retransmit three source devices because IE î — = Nf j + = 3.

[0143] E “ {1,4, 5] and B = {4} then E is eligible even though there is only one active node because 1X1= 2V jg — ^TA = 3-

[0144] Complexity of the selection process

[0145] When all cardinalities q { [, are considered, a search The exhaustive search is performed on the set 14^(¾^). The cardinality of this set is as follows:

[0146] yk '~^k=\ k\$d^k}

[0147] Thus, the complexity increases with the number of streams transmitted to the destination. One way to control this increase is to set a maximum number Q < N. & ^max “ of messages to be retransmitted simultaneously. In this case, the search is limited to EG subsets Qe {1, ..., min{Qmax, j}. The number of iterations in this case is given by the following equation:

[0148] [Srfj4{ / \ ^k=l ' k \2 /

[0149] Thus, for [S^il =7, NR — 8 and Qmav — 2, 28 iterations are executed against 127 iterations in the case where the number of flows is not limited.

[0150] Fig. 8 represents, in the form of a flowchart, a particular method of implementing a method for calculating equivalent mutual information of a channel linking the transmitting antennas of the relaying nodes to the receiving antennas of the destination.

[0151] Method for calculating equivalent mutual information

[0152] As illustrated by Figure 8, the calculation process includes a first step S610, of calculating a precoding matrix aimed at maximizing a quality criterion of a channel in the event of simultaneous transmission of messages from the source devices s* of the set 5* by the destination device 4. This quality criterion corresponds, in this example, to the equivalent mutual information previously mentioned.

[0153] More formally, a vector x G Xvc°m Taking the v = IEI source devices in E sorted in order of increasing indices is generated. The source devices E are jointly decoded by the set of relay nodes B€ with > IEI, where Nr# = From the set of nodes of relay B, a vector T gg® is constructed, such that b i< ••• <^61^2,€ V€ ze{l, ..., IBi|-

[0154] The channel equivalent to the transmission from the source devices listed in x by the nodes in B is represented by a matrix pj e The channel between the antenna The index re {1, ..., NR} of the destination device d and the transmitting antenna n G {1, ...,} of the relay node b{ e B, corresponds to the coefficient of the matrix, such that Jd_ y'--1 represents the total number of transmitting antennas P\ll ~

[0155] nodes b^^.^ with k the index representing the order of the nodes in the vector b. The precoding matrix y — [yi v^. j G is made up of the v best orthogonal eigenvectors of pp jj £ (eg, of the v vectors whose eigenvalues ​​are the most important), with JJ the adjoint matrix of H. The kèmc vector e is applied as precoding for the source such that the node b; e B uses N Tp. coefficients of the vector v<

[0156] The precoding coefficient applied by the antenna „ — fi I of the i-th tj 1, . . , , xV J1 element b: in b then corresponds to the element n+f$ of the vector Ni, i.e., vn+^i)l with / At vr 1 In this way, antenna n of the relay node b;e B transmits: P\l) ~ ^k=\^Tbk

[0157] yv v M

[0158] And the signal received by the destination device is then expressed as follows:

[0159] y = HV x+w #(4)#

[0160] with x the emitted signal, w a noise vector, E{ MW} = &2Inr with E {} the expected value, 67 the noise power, and Inr an identity matrix.

[0161] By applying (HV) ' ' we recover the v symbols emitted without interference: [°162] y^XlXl + W}^ { =

[0163] °ù E [ | Wj |2} = with the / _th largest eigenvalue of gj{.

[0164] The calculation method further includes a step S620 in which a power Pi to be allocated to each spatial layer l = 1, ..., v (or equivalently to each source in the set ■S) is determined. It is important to note here that this power allocation is done per spatial layer (or equivalently per source) and not per antenna, so as to comply with the orthogonality constraint of the precoding vectors. This step S620 is described in more detail below.

[0165] Then, during an S630 step, a "single-source" quality criterion is calculated for each spatial layer l = 1, ..., v (or equivalently for each source s of the set 5') as a function of the precoding matrix V and the transmission power to be allocated to said spatial layer 1 (or equivalently to said source *). In a particular embodiment, the single-source quality criterion associated with each spatial layer ^ corresponds to a signal-to-noise ratio.

[0166] The signal-to-noise ratio SNR of the 1st spatial layer is expressed as follows: SNR — ' with 'a the 4th largest eigenvalue of H'H' ° the power of the noise and Pi the power to allocate to the zth spatial layer.

[0167] Furthermore, the equivalent mutual information L is defined as the sum of the mutual information of the different spatial layers, and is expressed as follows: [01681 4)

[0169] In the case of Gaussian inputs, the equivalent mutual information L is given

[0170] by : Ev / ^iPj \ / \ fcl log2(l+") #(7)

[0171] It is important to note that in the case where the covariance of the noise (including interference) is not a multiple of the identity Inr, i.e. E {MW'} = R, then y = jy, V, ] EX^3** is made up of the v best eigenvectors orthogonal to g

[0172] Power calculation step S620

[0173] The transmission power of a node is limited by a power budget P which is shared by the antennas of this node. According to a particular implementation, an amplifier is connected to each antenna, so as to allow each of these antennas to transmit a signal at power P. In this case, the power Pi to be allocated to each spatial layer l = 1, ..., v (or equivalently to each source5 of the set S) is determined by maximizing the equivalent mutual information L (an expression of which is, by way of example, given by equation #(7)) subject to the following local constraints:

[0174] 1 |2r> n *1 idi 44 1 1 V n+p(i)lI 7* / —P, ï— L •••, IBI # (8)

[0175] The power allocation problem can then be written as:

[0176] max Y [og ] + j # ( 9 ) - = i- œi#(io)

[0177] Any method of satisfying constraints known to a person skilled in the art for solving such a problem may be considered; the choice of a particular method corresponding to only one implementation variant of the invention. According to a particular embodiment, this problem is solved by applying Lagrange multipliers with IBI constraints.

[0178] A possible suboptimal solution is as follows:

[0179] P=P =----------, Z - 1, ..., v # (11)

[0180]

[0181] Variant s In a particular implementation mode, the matrix y _ oa V| is the the i-th column of V is quantized, and for example chosen with respect to a predefined set of vectors B€ such that $ _ æ-oniin HV - VH 2 is an arbitrary norm. 1 * veB

[0182] In a particular embodiment, the matrix V is received noisy by the relay nodes during step S230. In this case, the signal received by the destination device becomes:

[0183] j = HVx + W#(12)

[0184] with y the noisy precoding matrix.

[0185] This signal no longer ensures zero interference at the receiver. In this case, the receiving device can use an equalizer that minimizes the squared error.

[0186]

[0187] Minimum mean (Linear Minimum Mean Squared Error, LMMSE, according to Anglo-Saxon terminology) of the type "* / ry'rr' " V1 with H = HV Which allows H \a H +JC) ur to reduce interference between the v spatial layers. The destination device d then calculates the v signal-to-interference-plus-noise ratios (SINR) at the output of the equalizer. Then the mutual information is calculated by considering the matrix y, and the v signal-to-interference-plus-noise ratios (SINR) fy. Figure 9 represents, in flowchart form, a particular method of implementing a selection of antennas of a relay node to be activated (resp. deactivated). In the implementations described so far, all antennas of a given relay node are activated, even if one of the links between a transmitting antenna of the relay node and a receiving antenna of the destination device is of poor quality. However, the equivalent mutual information of the channel could be optimized by selecting, within a relay node, only the antennas offering the highest quality links. This selection is only feasible, however, if > I ∈ Z. Furthermore, the maximum number of antennas that can be deactivated is equal to Nj - q - i ∈ ZI, in order to maintain the eligibility criterion for the entire Z.

[0188] More formally, let rig( { 1, .... Aj-b} ) be the set of all antenna subsets that can be disabled with cardinality El}- Ï^Ant-Off designates the set of all antenna subsets that can be deactivated in the network, and is expressed as follows: [01891 n^ / / = UgSr diyi JI,....) # ( 13)

[0190] By way of illustration, consider E = {1, 2, 3, 5}, B = {1, 4, 7}, Nta - t, Nta - 2, and NTJ - 3. The set E is indeed eligible for antenna selection since NtJi — T. 1 3" N ta "b NT,7 — 6 > I € EI — 4.

[0191] Furthermore, the maximum number of antennas that can be deactivated is equal to EI = 2. In other words, Q e |q, p 2}- If Q - Q, then no antenna can be deactivated. If Q=1, the antenna subsets that can be deactivated are: {1}, {2}, {3}, ,,,, {6}}. If Q=2, the antenna subsets that can be deactivated are: {{1,2], {1,3}, ..., {1,6], {2,2}, ..., {5,6}}.

[0192] Thus, the set ^Ant-Off of antenna subsets that can be deactivated is expressed as follows: {H, {2}, {3}, .... (6), {1,2}, {1,3}, .... {1,6}, {2,2}, .... {5,6}}

[0193] A* denotes the set of antennas to be deactivated in the system. To determine A*, an exhaustive search is performed on, based on the information equivalent mutual channel between the activated antennas (complement of A}e Qff in {1, ..., N tb} ) and the NR antennas in reception at the destination.

[0194] As illustrated in Figure 9, the method for selecting the antennas of a relay node to be activated (resp. deactivated) includes a first step S710 during which subsets Af of activatable antennas are determined the relay nodes bt e B.

[0195] The selection process further includes a calculation step S720, for each antenna subset Af determined in step S710, of a value representing a quality criterion for a communication channel enabling simultaneous transmission of messages from source devices "e S", by the antennas of said subset Af to the destination device. This value representing a quality criterion for a communication channel corresponds to the equivalent mutual information mentioned above, and is, for example, calculated in accordance with the calculation method described with reference to [Fig. 8].

[0196] Finally, an S730 step for selecting a subset of antennas that optimizes (e.g., which maximizes) the representative value of a quality criterion (e.g., equivalent mutual information) is implemented. Alternatively, the subset has * antennas selected corresponds to one of the subsets having a representative value of a quality criterion exceeding a threshold.

[0197] We note hereafter Q' the number of antennas to be deactivated. The cardinality of the set on which the exhaustive search for antennas is performed is expressed as follows:

[0198] | ^Ant-Off ] ~ ^q'^qq] (Nj-sQ"). # (14)

[0199] In a particular embodiment, the antenna selection corresponds to a selection of a subset of the set of antennas of the nodes of the set B offering the best links with the destination device d. In this particular case, a reduced-complexity search is performed on rr ih. with ^AnWff ~ Og' =0 Ag' Ag> the Q' antennas with the worst connections to the destination, for example, in terms of signal-to-noise ratio. Thus, the cardinality of the set on which the antenna search is performed is expressed as follows:

[0200] |{(), ..., 2VrîB-l€ ZI Zl + l # (15)

[0201] Figure 10 represents, in the form of a flowchart, a particular method of implementing a data communication process in an OMAMRC system, according to a second example.

[0202] This particular mode of implementation of a communication process differs from that described with reference to Figure 6 in that the destination device transmits, to the relay nodes b GN, the set of source devices whose messages have been decoded without error and / or the set S^_i of source devices whose messages could not be correctly decoded, rather than an ACK / NACK message.

[0203] Steps S300, S100, S310 and S320 are identical to those described with reference to [Fig.6], and are not re-described, for reasons of conciseness.

[0204] The method further includes a step S330-1 in which the destination device transmits to the relay nodes &g N the set of source devices whose messages have been decoded without error. Alternatively, the destination device transmits to the relay nodes b GN the complement of the decoding set of the destination device. This set is received by the relay nodes bE N in a step S210-1. Then, in a step S215-1, each relay node b GN determines the intersection S'^ between their decoding sets. respective S^, bE N, and the complement of the decoding set of the destination device S^, i.e., S'^ = AS^. This set of source devices whose messages have not been decoded during the time interval Cl by the destination, but which have been decoded by the relay node be N, is transmitted by said relay node during a step S220-1, and received by the destination device during a step S340-1.

[0205] Finally, steps S350, S360, S230, S240, S370 and S380 are identical to those described with reference to [Fig.6], and are not re-described, for reasons of conciseness.

[0206] At this stage, it is important to note that in the case where the method of selecting the antennas of a relay node to be activated or deactivated is applied, the set of antennas to be activated by each relay node is also transmitted, by the destination device and to the relay nodes, during the S360 step (in addition to the precoding matrix V and the vector b).

[0207] The set of antennas to be activated by each relay node takes, for example, the form of a bitmap vector of size denoted a, which indicates the subset of antennas to be activated. If the element of the vector corresponding to antenna n of node bt is equal to 1 (i = 1), then the antenna is activated (i = A), otherwise if 0, it is not it is not (Il EA^), with p(i)

[0208] This vector a is for example constructed from the set a} in the following way: the vector a is first initialized with "1"s, then the elements of the vector a which correspond to the index a GA, are instantiated as "0".

[0209] Illustrative example

[0210] Let - {1,2,3,5}, B* = {1,4,7}, Nr,l - 1, ^TA = 2 and A77 = 3 and assume that Aj = {2,5}. The bitmap vector a of dimension = 6 is as follows:

[0211] «= [1,0, 1,1,0,1]7

[0212] Alternatively, a vector 0 containing the indices in A^ is passed during this S630 step. Thus, considering the previous example, the vector a is expressed as follows:

[0213] a=[2,5]r

[0214] On the relay node side, this vector can be interpreted because the nodes activated during retransmission f are known thanks to the vector b, and the number of antennas of each node is also known. Thus, node 2 deactivates its first antenna and node 3 deactivates its second antenna.

[0215] It is also important to note that the set of antennas to be activated by each relay node must also be considered to determine the position of the precoding coefficients in the y — f vJ matrix. This matrix has the dimension xy, where A, denotes the set of antennas to be deactivated in the network. Thus, if antenna n of node bj is activated, i.e., n & A^„ this antenna applies the coefficient corresponding to the element ,, .. of the vector vtsoit Par P(A) = L / =1 aj Subsequently, each activated antenna simultaneously transmits the same redundancy of information messages from the selected source devices.

[0216] Figure 11 schematically represents an example of OMAMRC systems in which the method according to the invention is implemented. In this example, a system is considered comprising two source devices, siet s2, and a destination device 4, with Nf j ~ 2, = 2 and NR = 4.

[0217] As illustrated in Figure 11, the decoding set Slm of the source device at interval t-1 is such that = the decoding set of the source device s2 at interval t-1 is such that Sy / .j — {$}, ^21 ct the decoding set of the destination device d at interval t-1 is such that $ - {0}'

[0218] Also, three configurations are possible: S = {sJ cl B = {"j, $2} or (¾} and B = {$4,¾} or L = and B = We assume that the The configuration in which L = {s15 s2} and B = {jq, s2] yields the best equivalent mutual information. In this case, the two transmitting antennas of the source device s2 simultaneously transmit the same redundancies of the source devices and s2, and x = [jq, s2]7-

[0219] The eigenvector yj ( y2 ( V3 ] V4 ] ^corresponding to the largest value Proper A is applied as precoding of the source device s i' and the eigenvector y2 = [tq2 V99 2 V4 ?] ^corresponding to the second largest eigenvalue 22 of H'J?est aPPÜqué as precoding of the source device s2.

[0220] In other words, the first antenna of the source device '^.transmits the signal Vj 1 * / ?VXS)+ V] the second antenna of the source device if.transmits the signal v2.i * RVXs}+ V? 2*RVXs2, the first antenna of the source device s2.transmits the signal 1'3 । * RVXS + V3 and 'a second antenna of the device source s2.transmits the signal V4 j * ^RVXs...

[0221] Also, the power transmitted by each relay node is expressed as follows:

[0222] p --------,____.____.___P----------,----.______ 0 max^'j 4^ / 4^2 / 4^2 / ),

[0223] and the equivalent mutual information: j _ j + ) + । + ^0 .

[0224]

[0225]

[0226]

[0227]

[0228]

[0229]

[0230]

[0231]

[0232] Furthermore, since = 4 > ! € E ! = 2, the process of selecting source devices and antennas to be activated by each relay node can be applied. The antenna subsets that can be deactivated are as follows: ^Ant-Off= {0, OK {2}, {3}, {4}, {1,2}, {1,3}, {1,4}, {2,3}, {2,4}, {3,4}} with indices 1 and 2 corresponding to the antennas of source si and indices 3 and 4 to the antennas of source E. We also assume that the subset of antennas to be deactivated is such that A = {3}, and that it is the first antenna (index 3) of the source s2 that is deactivated. A new H channel is defined between the active transmitting antennas of the relay nodes and the NR receiving antennas of the destination. The eigenvector jb, J 7 corresponding to the largest eigenvalue of R' is applied as precoding of the source if, the eigenvector v( _ [ v2 ? v3 ? r corresponding to the second largest eigenvalue of JJ'1 Jjf is applied as precoding of the source s2, in other words, the first antenna of the source device si transmits the signal vJ j * RVXSi+ V] 2^VXS?, the second antenna of the source device transmits the signal v2 i * v2 ^RVXS?, and the second antenna of the source device s2 transmits the signal v3 j * ^RVXSr The power transmitted by each relay node is then expressed as follows: p -------—£-- and the equivalent mutual information I in the case of Gaussian inputs is expressed as follows: T 1 / 1 \ . / A-,Pq \ 1 = log9(1 + -J + log ? ( 1 + —;

Claims

1. Demands A data reception method comprising the following steps, implemented by a destination device (d) comprising several receiving antennas, following the reception of M > 2 messages transmitted successively by M > 2 source devices and the inability of the destination device to decode at least two messages received from at least two source devices called "undecoded source devices", each message transmitted by a source device (including a first redundancy version (RVO) resulting from an encoding of an information message from the source device ($ gm) • a selection (S350), of a set g* comprising several undecoded source devices and a set B* of several relay nodes Q,*) associated with the set S^, according to a quality criterion of a communication channel between the relay nodes ( of the set B* and the destination device (d), a relay node of the set B^ corresponding to a source device (seM) or to an intermediate device (r) knowing the information messages of the undecoded source devices, at least one relay node ( of the set B, having several (N^ y) transmitting antennas; • a transmission (S360), to the selected relay nodes Qp, of a retransmission instruction, so that the relay nodes Qp simultaneously transmit second redundancy versions resulting from the encoding of the information messages from the undecoded source devices (sf) of the set; and, • a reception (S370), by the destination device (d), of said second redundancy versions resulting from the encoding of the information messages of the source devices ($0 of the set $*, said second redundancies being used by the destination device (d) to decode messages received from source devices in set 5*.

2. A receiving method according to claim 1, wherein said quality criterion is mutual information of the communication channel between the relaying nodes (^) of the assembly and the destination device (d).

3. A receiving method according to claim 1 or 2, wherein the selection (S350) comprises: • a determination (S530) of eligible sets 5' of undecoded source devices (S'J) and sets B of relay nodes associated with the eligible sets S, such that 2 < 1S1 < Nt-g, where 11 is the cardinality operator and Nts is the total number of transmitting antennas of the relay nodes in set B; • a determination (S540), for each of the combinations of eligible sets 5' and sets B of relay nodes, of the quality criterion of the communication channel between the relay nodes in set B and the destination device (d); and the sets 5* of source devices (s't) and B, of relay nodes (^*) selected optimizing said quality criterion.

4. A receiving method according to claim 3, wherein the determination (S540) of the quality criterion of the communication channel between the relay nodes of assembly B and the destination device (d) includes: • a calculation (S630), for each source device(s) of the eligible assembly of a single-source quality criterion for each source device(s) of the eligible assembly as a function of a precoding matrix and a transmit power to be allocated to said source device(s); and, • a calculation (S640) of the quality criterion of the communication channel, as a function of the single-source quality criteria calculated for each source device(s) of the eligible assembly 51.

5. A receiving method according to any one of claims 1 to 4, further comprising a selection of transmitting antennas to be activated by each of the relay nodes of at least one subset of the Bj- assembly

6. A receiving method according to claim 5, wherein the antenna selection comprises: • a determination (S710) of antenna subsets Ât that can be activated by each of the relay nodes Q,*) of at least one subset of the set Bp; • for each determined antenna subset At, a calculation (S720) of the quality criterion of a communication channel between the antennas of said subset Ât, and the destination device (d); • a selection (S730), from among the subsets Ât, of the antenna subset a* to be activated optimizing said quality criterion.

7. A receiving method according to any one of claims 1 to 6, wherein the quality criterion of the communication channel between the relay nodes (^*) of assembly B* and the destination device (d) is determined based on a combination of single-source quality criteria of subchannels of said communication channel, each subchannel being defined for a source device hV) of the assembly ■

8. A data transmission method comprising the following steps, implemented by a relay node (jf,*) comprising several (Nt [>) transmitting antennas, following the reception of M > 2 messages transmitted successively by M > 2 source devices and the inability of a destination device to decode at least two messages received from at least two source devices, referred to as "undecoded source devices," each message transmitted by a source device ($ e M) comprising a first redundancy version (RVO) resulting from the encoding of an information message associated with the source device ( sgM): • a reception (S230) of a retransmission instruction of second redundancy versions resulting from the encoding of the

9.

10.

11.

12.

13. information messages from undecoded source devices (s*) of a set $*, said relay node knowing said information messages from the undecoded source devices (s*) of the set; and, • a simultaneous transmission (S370) of the second redundancy versions of the information messages from the source devices (sf) of the set g* by the transmitting antennas of said relaying node and to the destination device, said second redundancy versions being used by the destination device (d) to decode the messages received from the undecoded source devices (s*) of the set £*. Computer program comprising instructions for implementing a receiving method according to any one of claims 1 to 7 or a transmission method according to claim 8, when said program is executed by a processor. Computer-readable recording medium on which a computer program according to claim 9 is recorded. Destination device (d) comprising several (Nr) receiving antennas and configured to implement a receiving method according to any one of claims 1 to 7. Relay node (^) comprising several transmitting antennas and configured to implement a transmission method according to claim 8. System (SYS) comprising M >2 source device(s), a destination device (d) according to claim 11 and a plurality of relaying nodes of which at least one conforms to claim 12.

Images