Cooperative retransmission method in omamrc system

Abstract

The present invention relates to a transmission method for an OMAMRC telecommunications system having M sources (s1,i,..., sM), optionally L relays, and a destination, where M ≥ 2, L ≥ 0. In such a solution, when a source has not been able to be decoded by the destination, the destination determines an active set (Âι <sb / > of nodes from among all the nodes of the system that have decoded the source, and then organizes a simultaneous retransmission, via said active set, of a message transmitted by the destination.

Description

[0001] DESCRIPTION

[0002] TITLE: Cooperative retransmission method in an OMAMRC system

[0003] 1. Field of the invention

[0004] The present invention relates to the field of digital communications. Within this field, the invention relates more particularly to the transmission of coded data between at least two sources and a destination with relaying by nodes which may be relays or sources.

[0005] It is understood that a relay has no message to transmit. A relay is a node dedicated to relaying messages from sources while a source has its own message to transmit and can also in certain cases relay messages from other sources (in this case, the source is said to be cooperative).

[0006] There are many relaying techniques known by their Anglo-Saxon names: "amplify and forward", "decode and forward", "compress-and-forward", "non-orthogonal amplify and forward", "dynamic decode and forward", etc.

[0007] The invention applies in particular, but not exclusively, to the transmission of data via mobile networks, for example for real-time applications, or for example via sensor networks.

[0008] Such a sensor network is a multi-user network, comprising several sources, several relays and a destination, which can use an orthogonal multiple access scheme for the transmission channel between the sources and the destination, noted OMAMRC ("Orthogonal Multiple-Access Multiple-Relay Channel" in English terminology).

[0009] According to this scheme, orthogonality between source and relay transmissions can be achieved by time multiplexing in the form of disjoint time slots.

[0010] 2. Prior art and its drawbacks

[0011] It is known from application WO 2019 / 162592 published on August 29, 2019 an OMAMRC telecommunication system which comprises M sources, possibly L relays and a destination, M ≥ 2, L ≥ 0 with an implementation of an orthogonal multiple access scheme in time of the transmission channel which applies between the nodes taken from among the M sources and the L relays. The maximum number of time slots per transmitted frame is M + T max with M intervals allocated during a first phase to the successive transmission of the M sources and T used ≤ T maxintervals for one or more cooperative transmissions allocated during a second phase to one or more nodes selected by the destination according to a selection strategy.

[0012] The known OMAMRC transmission system comprises at least two sources. Each of these sources can operate at different times either exclusively as a source or as a relay node. The system may optionally further comprise relays. The terminology node covers both a relay and a source acting as a relay node or as a source. The system considered is such that the sources themselves can be relays. A relay is distinguished from a source in that it has no message to transmit of its own, i.e. it only retransmits messages from other nodes. Such an OMAMRC transmission system is described in the article S. Cerovic, R. Visoz, L. Madier “Efficient Cooperative HARQ for Multi-Source Multi-Relay Wireless Networks,” IEEE Eleventh International Workshop on Selected Topics in Mobile and Wireless Computing 2018.The channels between the different nodes of the system are subject to slow fading and white Gaussian noise. Knowledge of all the channels of the system (via the CSI: Channel State Information) by the destination is not available. Indeed, the channels between the sources, the channels between the relays, and the channels between the relays and the sources are not directly observable by the destination, and their knowledge by the destination would require too much information exchange between the sources, the relays and the destination. To limit the cost of feedback overhead, only information on the channel distribution / statistics (CDI: Channel Distribution Information) of all the channels, e.g. average quality (e.g. average SNR, average SINR) of all the channels, is assumed to be known by the destination in order to determine the rates allocated to the sources.

[0013] Channel adaptation is said to be slow, meaning that before any transmission, the destination allocates initial rates to sources knowing the distribution of all channels (CDI: Channel Distribution Information). In general, it is possible to go back to the CDI distribution based on knowledge of the average SNR or SINR of each channel in the system.

[0014] During transmissions of framed source messages, the channel CSIs are assumed to be constant (slow fading assumption). The bit rate allocation is assumed not to change for several hundred frames, it only changes with CDL changes.

[0015] A transmission method implemented in such an OMAMRC system distinguishes three phases: an initial phase and, for each frame to be transmitted, a 1 ere phase and a 2 ndephase. The transmission of a frame takes place in two phases which are possibly preceded by an additional phase called the initial phase.

[0016] During the initialization phase, the destination determines an initial bitrate for each source by taking into account the average quality (e.g. SNR) of each of the system's channels.

[0017] The destination estimates the quality (e.g. SNR) of the direct channels: source to destination and relay to destination using known techniques based on the use of reference signals. The quality of the source-source, relay-relay and source-relay channels is estimated by the sources and relays using, for example, the reference signals. The sources and relays transmit the average qualities of the channels to the destination. This transmission occurs before the initialization phase. Since only the average value of the quality of a channel is taken into account, its refresh occurs on a long time scale, i.e. over a time that allows the rapid variations (fast fading) of the channel to be averaged. This time is of the order of the time required to travel several tens of wavelengths of the frequency of the transmitted signal for a given speed. The initialization phase occurs, for example, every 200 to 1000 frames.The destination returns the initial flow rates it has determined to the sources via a return path. The initial flow rates remain constant between two occurrences of the initialization phase.

[0018] In the first phase, the M sources successively transmit their message during the M time slots using modulation and coding schemes determined from the initial bit rates. During this phase, the number N ± channel use (ie resource element according to 3GPP terminology) is fixed and identical for each of the sources.

[0019] In the second phase, messages from the sources are transmitted cooperatively by the relays and / or by the sources. This phase lasts a maximum of T maxtimeslots. During this phase, the number N2 of channel uses is fixed and identical for each of the selected nodes (sources and relays). The sources, which are independent of each other, broadcast their messages during the first phase in the form of coded information sequences for a single recipient. Each source broadcasts its messages at the initial rate. The destination communicates its initial rate to each source via very limited rate control channels. Thus, during the first phase, the sources each transmit their respective message during time slots, each dedicated to a source.

[0020] Sources other than the one transmitting and possibly relays, of the "Half Duplex" type, receive successive messages from the sources, decode them and, if selected, generate a message only from the messages from the sources decoded without error.

[0021] The selected nodes then access the channel orthogonally in time to each other during the second phase to transmit their generated message to the destination.

[0022] The destination can choose which node should transmit at any given time.

[0023] Although such a solution allows to maximize the average spectral efficiency (utility metric) within the system considered under the constraint of respecting an individual quality of service (QoS) per source, it is desirable to try to further improve the decoding performance of a given source.

[0024] The present invention meets this objective.

[0025] 3. Statement of the invention

[0026] To this end, the present invention relates to a transmission method intended for an OMAMRC telecommunications system with N nodes and a destination (d), the N nodes comprising M sources (s1... , s M) and possibly L relays (r1... , r L ), with M ≥ 2, L ≥ 0, comprising a first phase during which the destination receives first redundancies (RV0) of messages transmitted successively by the M sources, the message from a source having been coded before transmission by an incremental redundancy type coding comprising several redundancies and a second phase comprising the following steps implemented by the destination (d): broadcasting a control message identifying one or more sources for which it has not decoded said transmitted message without error, called non-decoded sources, receiving at least one identifier from at least one source (s i ) not decoded by the destination transmitted by a first set of nodes comprising at least one node, taken from among the N nodes, having decoded without error said message from a source s i, determination, among the nodes of the first set, of a second set of nodes, called the active set (Â i ), associated with the source (s i ), issuing a request for retransmission of said message from the source (s i ), to the nodes of the active set (Â i ), and reception of the same second redundancy of the message from the source (s i ) transmitted simultaneously by at least two nodes of the active set (Â i ) in the same time interval.

[0027] Such a method allows several nodes to simultaneously transmit the same redundancy for the same message from the same source in the same time interval.

[0028] Knowing that each node of the system has its own independent power budget, the redundancy thus obtained improves the raw decoding performance of a source s, by proposing that certain nodes of the system, hereinafter called active nodes having decoded without error a message transmitted by the source s, according to a first redundancy simultaneously retransmit a second redundancy of this message, i.e. by using the same use of the channel ("channel use"). These active nodes form what will be called for the following an active set.

[0029] Thus, the equivalent transmission power for source s is multiplied by the number of active nodes in the system that have decoded a message sent by source s without error and participate in the retransmission. The first and second redundancies can be identical, for example when using a repeating code, or not, and including or not systematic bits.

[0030] In this method, it is specified that the first redundancy is a code word. The fact that the first redundancy is a code word makes it possible to go back to the transmitted message because there is a unique correspondence between code word and message which requires a coding efficiency less than or equal to 1.

[0031] By avoiding systematically soliciting all the nodes of the system, the efficiency of the retransmission is improved. Thus, for example, the nodes whose transmission has a limited power gain because their respective transmission channels are of low power are not solicited to retransmit the message sent by the source s, even if it has been decoded without error by these nodes.

[0032] By avoiding activating certain nodes for the retransmission in question, it is then possible to limit the formation of interference. Finally, the network's energy consumption is reduced, because the nodes that do not provide any real performance gain are not used.

[0033] In one example, determining the active set (Â i ) includes for at least one subset (A i ) of nodes taken from the first set of nodes: the determination of a utility metric as a function of the size of said subset (A i ) and the quality of a channel established between the source s i and the destination (d) via nodes of the subset (A i ), the determination, among the subsets of nodes taken from the first set of nodes, of the subset (A i ) whose determined utility metric is the largest as an active set (Â i ).

[0034] Here, the constitution of the active set is carried out by seeking to maximize a utility metric, so as to find a compromise between the number of simultaneously active nodes (energy efficiency) and the gain in performance (spectral efficiency). The general efficiency of the method is improved, without this resulting in a degradation of the quality of the retransmission.

[0035] In one example, determining the utility metric of a subset (A i ) includes the determination of mutual information representative of the quality of a channel established between the source s i and the destination via the nodes of the subset (A i ), called mutual information relating to the subset (A i ), the utility metric being a function of said mutual information thus determined.

[0036] In this case, the quality of the channel established between the source and the destination via nodes of the subset is represented by the mutual information relating to the channel established between the source and the destination via the nodes of the subset.

[0037] In one example, the utility metric of a subset is proportional to the mutual information about that subset. In one example, the utility metric of a subset (A i ) is inversely proportional to an increasing function of the size of said subset (A i ), said increasing function exhibiting logarithmic growth.

[0038] In this example, we place ourselves in a scenario, called the reference scenario, corresponding to equal fading for all the links between the sources belonging to (A i ) and the destination (d). In this case the power received at the destination is proportional to the cardinality of the subset (A i). This allows us to obtain an approximation of the asymptotic behavior of the behavior of the mutual information relative to the subset (A i ).

[0039] Here, the growth of mutual information is weighted by a denominator whose growth is logarithmic. The choice of a growth quotient at least logarithmically comes from the fact that mutual information is a quantity that grows logarithmically at the asymptote, that is to say when the cardinal of the active set becomes very large. This choice of denominator, of logarithmic growth, makes it possible to counterbalance this logarithmic growth at the asymptote. The presence of such a logarithmic growth denominator makes it possible to determine an active set of smaller size than if the utility metric depended only on mutual information.

[0040] The denominator can grow logarithmically or faster than logarithmically. More precisely, in a power-limited regime (or low signal-to-noise ratio (SNR) regime), the mutual information exhibits a linear growth (with respect to the received power, i.e. with respect to the size of the subset whose utility metric is calculated for the reference scenario). In a band-limited regime (or high SNR regime), the mutual information grows logarithmically. Thus, adding an additional active node to the subset is only allowed if it contributes to at least a logarithmic increase in spectral efficiency given by the value of the mutual information in number of bits per channel use or bits per second per hertz (at least the gain of the band-limited or high SNR regime).

[0041] Furthermore, the discrete nature of the channel inputs, taken into account in the calculation of mutual information, leads to the fact that the mutual information is capped by the number of bits q carried by the modulation. In fact, increasing the power (and therefore the size of the subset whose metric is determined) asymptotically (i.e. when the size of this subset becomes very large) only leads to negligible gains in spectral efficiency. In other words, adding a node to this large cardinal subset only increases the mutual information by a small amount, since this mutual information is bounded by q, and the subset with a large cardinal already has mutual information close to q. The potential increase in mutual information then becomes negligible compared to the logarithmic growth of the denominator.The active set thus determined (as optimal in the sense of the metric among the sets whose metric is calculated) is of reduced size, compared to the first set comprising all the nodes.

[0042] In one example, the method comprises calculating the utility metric M(A i ) is performed for all subsets A L taken from the first set of nodes.

[0043] This scheme for determining an active set is said to be exhaustive. Here, the destination determines the utility metric of all subsets of the set, before determining the best subset in the sense of the utility metric. This advantageously makes it possible to find the optimal active set for the utility metric. In one example, the method further comprises constructing a subset (A i) initially equal to the empty set, said construction comprising at least one iteration of the following steps: determining a node (y) outside the subset (A i ) having the highest signal-to-noise ratio (SNR), if the addition of said node (y) to the subset (A i ) improves the utility metric of the subset (A i ), add node (y) to subset (A i ), at the last iteration, the subset (A i ) thus constructed being the active set (Â i ).

[0044] The determination scheme for this example of determining an active set is heuristic. In other words, it is an approximation, compared to the exhaustive scheme described above. This heuristic scheme is considerably faster to execute, as the number of nodes that can potentially help increases.

[0045] Furthermore, this scheme is optimal when we are in a so-called "equal gain combining" case, in which all the relay nodes know the CSI of their channel with the destination. Each of these nodes can then know the phase of its channel with the destination, and compensate for this phase. This allows the destination to receive all the messages at the same time. The combination of these redundancies is then consistent. In this case, the best mutual information for a given number of active relay nodes is that linked to the set of N relay nodes having the best SNR (i.e. the best channel qualities with the destination). In another example, the method further comprises determining, for at least one source (s i ) not decoded by the destination, of a set (H i ) associated including the nodes having decoded without error the message sent by said source not decoded by the destination, and the determination of the mutual information (SNR Hi) relative to said set (H i ) associated, and determining, among the at least one undecoded source, the source whose associated set has the highest relative mutual information, and determining the active set of said source thus determined, the active set of each of said at least one undecoded source being equal to said active set thus determined.

[0046] The scheme of this exemplary implementation is heuristic, and suboptimal, compared to an exhaustive scheme, but faster to compute. Indeed, rather than determining the active set for each source taken in isolation, the destination first determines the source presenting the best channel established with the destination via the nodes of its set H i— assuming retransmission via the nodes in this set. Then, the destination searches for the best subset of the set Hi, and uses this subset thus determined as the active set for all sources.

[0047] The invention further relates to a system comprising M sources (s1, ... , s M ), L relay and a destination (d), M ≥ 2, L ≥ 0, for an implementation of a method of transmission as described above.

[0048] The invention further relates to a computer program product comprising program code instructions for implementing a method according to the invention as described previously, when executed by a processor.

[0049] The invention further relates to a computer-readable recording medium on which is recorded a computer program comprising program code instructions for executing the steps of a method according to the invention as described above. Such a recording medium may be any entity or device capable of storing the program. For example, the medium may comprise 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 key or a hard disk.

[0050] On the other hand, such a recording medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means, so that the computer program contained therein is remotely executable. The program according to the invention may in particular be downloaded over a network, for example the Internet.

[0051] Alternatively, the 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 method which is the subject of the aforementioned invention.

[0052] 4. List of figures

[0053] Other aims, characteristics and advantages of the invention will appear more clearly on reading the following description, given as a simple illustrative, and non-limiting, example, in relation to the figures, among which:

[0054] [fig. 1] represents an embodiment of the invention described in the context of an OMAMRC system,

[0055] [fig. 2] represents a transmission cycle of a frame,

[0056] [fig. 3] represents the different stages of the transmission method which is the subject of the invention implemented by the system of figure 1,

[0057] [fig. 4] represents a circular buffer allowing the selection of a redundancy of the message to be transmitted,

[0058] [fig. 5] represents step E5 of the process shown in figure 3,

[0059] [fig- 6] represents a first example of carrying out step E50 of step E5 represented in figure 5,

[0060] [fig- 7] represents a second example of embodiment of step E50 of step E5 represented in figure 5,

[0061] [fig- 8] represents a destination belonging to an OMAMRC telecommunications system with M sources, possibly L relays and a destination, M ≥ 2, L ≥ 0 according to an embodiment of the invention.

[0062] 5. Detailed description of embodiments of the invention

[0063] 5.1. General principle of the invention

[0064] 5.1.1. Redundant OMAMRC System

[0065] In connection with [Fig. 1] an embodiment of the invention is presented, described in the context of an OMAMRC system, with the support of the diagram of [Fig. 2] which illustrates a transmission cycle of a frame.

[0066] This system includes M sources which belong to the source set = {s 1; ..., s M ], L relays that belong to the relay set and a destination d. By convention, it is considered that , in other words, we can confuse a source and its index, and a relay and its index (shifted by a value M, the number of sources).

[0067] Each source of the game communicates with the single destination d with the help of other sources (user cooperation) and cooperating relays. To simplify the description, the following assumptions are made subsequently, concerning the OMAMRC system:

[0068] 1. sources, relays and destination are equipped with a single transmitting antenna;

[0069] 2. sources and relays are equipped with a single receiving antenna;

[0070] 3. the destination is equipped with N R receiving antennas;

[0071] 4. the sources, relays, and destination are perfectly synchronized;

[0072] 5. the sources are statistically independent (there is no correlation between them);

[0073] 6. all nodes transmit with the same power;

[0074] 7. use is made of a CRC code assumed to be included in the K s information bits corresponding to the message from each source s to determine whether this message is correctly decoded or not;

[0075] 8. Channels between different nodes suffer from additive noise and fading. Fading gains are fixed during the transmission of a frame carried out for a maximum duration M + T max time intervals, but can change independently from frame to frame. T max ≥ 1 is a system parameter;

[0076] 9. the instantaneous quality of the channel / direct channel in reception (CSIR Channel State Information at

[0077] Receiver) is available at the destination, sources and relays;

[0078] 10. returns are error-free (no errors on control signals).

[0079] Nodes include relays and sources that can behave as a relay when not sending their own message.

[0080] The nodes, M sources and L relays, access the transmission channel using a time-orthogonal multiple access scheme that allows them to listen to the transmissions of other nodes without interference. The nodes operate in a half-duplex mode.

[0081] The following notations are used:

[0082] • Hi is the set of nodes a having decoded the message u without error i emitted by the source Si,

[0083] • HAS i ⊂ H i a subset of the active transmitting nodes selected by the destination for use of channel k

[0084] • is the coded symbol for using channel k emitted by the node

[0085] • ya,b,k is the signal received at node b of channel k corresponding to a signal emitted by the node

[0086] • ys -,b,k is the signal received at the node of channel k corresponding to the signals emitted by the nodes

[0087] • Ya,b is the average signal-to-noise ratio (SNR) which takes into account the effects of channel attenuation (path-loss) and masking (shadowing),

[0088] • h a, b is the channel fading gain which follows a zero-mean, circularly symmetric complex Gaussian distribution of variance y a, b (the received power which is proportional to the emitted power), the gains are independent of each other,

[0089] • n a,b,k or n si,b,kare identically and independently distributed white Gaussian noise (AWGN) samples that follow a complex Gaussian distribution of circular symmetry with zero mean and unit variance.

[0090] R s is a variable representing the initial flow rate of source s. R s can take its values ​​in the finite set itself, a s is a variable representing the ratio N2 / N 1 s and can take its values ​​from a finite set A The signal received at the node of the channel k corresponding to the signal emitted by node a ES during the first phase can be written: (1)

[0091] The signal received at the node of the channel k corresponding to the signals emitted by the nodes uds belonging to the set during the second phase can be written: (2) where x , i.e., the same redundancy version on message s i is transmitted by all s the nodes is a phase correction term with respect to the channel h a, d with j 2 = —1. For the destination node, the received signal is written as

[0092] [Fig. 3] represents the different stages of the transmission method, the subject of the invention, implemented by the system described above.

[0093] During a first Phi phase of M time intervals, each source sends a message, encoded using a code allowing incremental redundancy type retransmissions which transforms the message of length L M into a coded sequence of length L c = L M / R0> L M - The coded sequence includes a first RVO redundancy which is a code word transmitted during N l s channel uses, k E {1, ... , 1V 1 S ], the number N l s of source-dependent channel uses.

[0094] By exploiting reference signals (pilot symbols, 3GPP LTE SRS signals, etc.), the destination can determine the gains (CSI Channel State Information) of the direct channels: h dir = i.e. source to destination and relay to channels destination, and can therefore deduce the average SNRs of these channels.

[0095] On the other hand, the gains of the channels between sources, the channels between relays and the channels between sources and relays are not known to the destination. Only the sources and the relays can estimate a metric of these channels by exploiting reference signals in a similar way to that used for the direct channels. Given that the channel statistics are assumed to be constant between two initialization phases, the transmission, to the destination d, of the metrics by the sources and the relays only occurs at the same rate as the initialization phase. The channel statistics of each channel are assumed to follow a centered circular complex Gaussian distribution and the statistics are independent between the channels. It is therefore sufficient to consider only the average SNR as a measure of the statistics of a channel.

[0096] The sources and relays therefore return to the destination metrics representative of the average SNRs of the channels that they can observe.

[0097] The destination thus knows the average SNR of each of the channels.

[0098] 5.1.2. First phase of transmission

[0099] During an initial channel adaptation phase (not shown in the figures) which precedes the transmission of several frames, the destination transmits for each source s a representative value (index, MCS, rate, etc.) of an initial rate Rt and a value âj.

[0100] Each of the initial bit rates unambiguously determines an initial Modulation and Coding Scheme (MCS) or conversely each initial MCS determines an initial bit rate. The recovery of the initial bit rates and the ratios is carried out via very limited flow control channels.

[0101] Each source transmits its framed messages to the destination with the help of other sources and relays.

[0102] A frame occupies time slots during the transmission of M messages from M sources respectively. The transmission of a frame (which defines a transmission cycle) takes place during M + T used time intervals: M intervals for the first phase of respective capacities N channel utilizations for each source i, T used intervals for a second phase which will be described later in this document.

[0103] Still during the first phase, each source S transmits after coding a message u s of K s bits of information being the two-element Galois field. The message u s includes a CRC-type code that allows the integrity of the message to be verified. s . The message u sis encoded according to the initial MCS. Since the initial MCSs may be different between sources, the lengths of the encoded messages may be different between sources.

[0104] The coding applied uses an incremental redundancy code, which can be based, for example but not exclusively, on existing codes such as convolutional codes, turbo codes, LDPC, etc.

[0105] The principle of this type of code is as follows: a message sent by each source is encoded into a coded sequence of bits (there may be a segmentation of the message into several independently encoded sub-blocks if the message is too long) by a very low rate mother code (for example 1 / 3), the coded bits are then placed in a circular buffer represented in [Fig. 4] comprising several reading start positions Pos. 0, Pos. 1, Pos. 2 and Pos. 3. Such a circular buffer contains the coded bits of a message from a source encoded by a (possibly) systematic low rate mother code and making it possible to select a particular redundancy of the message to be transmitted according to a reading start position in the circular buffer.

[0106] These reading start indices Pos. 0, Pos. 1, Pos. 2 and Pos. 3 correspond to different redundancy blocks / versions. In the chosen example, there are four possible redundancy versions. For each redundancy block / version, a node will read the number of coded bits to be sent, corresponding to the number of channel uses available for a given modulation and message size, from the corresponding redundancy position by moving in the circular buffer in the direction of the initial filling. Whether or not the incremental redundancy code is of the systematic type, it is such that the first version of the redundancy block / version can be decoded independently of the other blocks / versions.

[0107] Thus, during the first phase, the M sources successively transmit the first redundancy RV0 of their messages u srespective coded during the M intervals, with modulation and coding schemes respectively determined from the values ​​of the initial rates.

[0108] Each message u s transmitted corresponding to a source , a correctly decoded message is assimilated to the corresponding source by abuse of notation.

[0109] When a source transmits, other sources and relays listen and attempt to decode the received messages at the end of each time interval.

[0110] 5.1.3. Second transmission phase A second transmission phase comprises steps E1 to E6. In a first step E1, the destination determines the success or failure of the decoding of the received messages using the CRC.

[0111] In this second phase, a given node, source or relay, can indeed act as a relay by cooperating with the sources to help the destination correctly decode messages from all sources. This given node transmits (i.e. cooperates by transmitting) a redundancy version of a message from a source that it has correctly decoded. The second phase includes at most T max time intervals (time slots) called rounds. Each round t ∈ {1, ... , T max ] has a capacity of N2 channel uses.

[0112] If the decoding of all sources is correct, the destination broadcasts an ACK message. In this case, a transmission cycle of a new frame begins with the erasure of the memories of the relays and the destination and with the transmission of new messages by the sources.

[0113] If the decoding of at least one source is erroneous, a retransmission method comprising steps E2 to E6 is implemented. In a step E2, the destination broadcasts an MSG message identifying the source(s) for which it has decoded the transmitted message without error. Such sources are called decoded sources. The MSG message may be addressed to the relays, to the sources that can act as relays, or to both. This message is a control message.

[0114] Such a message broadcast by the destination includes, in a first implementation, identifiers of the sources for which the destination has decoded the transmitted message without error. In this first implementation, the nodes intercepting the broadcast message determine the sources for which the destination has not decoded the transmitted message without error.

[0115] In a second implementation, the message broadcast by the destination includes identifiers of the sources for which the destination did not decode the transmitted message without error. In this second implementation, the nodes intercepting the broadcast message immediately know the identity of the sources for which the destination did not decode the transmitted message without error.

[0116] The destination informs the nodes using a limited-feedback control channel to transmit the MSG message. This MSG message is based on the decoding result of the messages received by the destination. The destination thus controls the transmission of the nodes using this MSG message, which improves spectral efficiency and reliability by increasing the probability of decoding all sources by the destination.

[0117] Upon receiving an MSG message, a node who correctly decoded the message u correctlys from one or more sources not correctly decoded by the destination at the end of the previous time interval (round) noted transmits to the destination, in a step E3, the identifier of these sources.

[0118] By convention, we note the set of messages (or sources) correctly decoded by the node at the end of time interval t (round t), t ∈ {0, ... , T max}. The end of the time interval (round) t = 0 corresponds to the end of the first phase. The number of time intervals (time-slots) used during the second phase T used = {1, ... , T max ] depends on the success of decoding at the destination.

[0119] During a step E4, the destination selects a source s i for which a retransmission is required. Such a source i is selected from the set of sources correctly decoded by one or more nodes at the end of the previous time interval t (round t), t ∈ {0, ..., T max}. Thus, rather than leaving the choice of a message to be retransmitted to the nodes having decoded without error a message sent by a source, the destination imposes the choice of the message and therefore of the source for which a retransmission is required.

[0120] In a first implementation, the source s i selected by the destination is the source for which a signal-to-noise ratio SNRi associated with a composite transmission channel, , established directly between each of the nodes having decoded the message sent u without error i by the source s i and the destination, is the highest.

[0121] By choosing the source for which the composite transmission channel has a high signal-to-noise ratio, the destination increases its chances of decoding the message u without error i during its retransmission.

[0122] In a step E5, once the destination has selected the source s i for which a retransmission is required, the destination d: determines, among the sources and relays of H i , a subset, hereinafter called the active set  i associated with the source s i , of nodes, called active nodes, intended to contribute to retransmitting the message sent by the source s i , and issues an RTM retransmission request to the active nodes belonging to the active set (Âi). This RTM retransmission request includes an identifier of the source Si-

[0123] In step E6, a retransmission of a redundancy of the message u t emitted by source s i , is carried out.

[0124] Upon receipt of the retransmission request by the nodes of the active set, each active node, in step E6, transmits the same redundancy, modulated by a phase factor. Here, the factor ha, d represents the established transmission channel between node a and the destination corresponds to the conjugate h a * d of the canal of transmission h a, d established between node a and destination d divided by its norm | h a, d | in the same time interval. The factor Φ a, d represents the phase of the established transmission channel between this active node and the destination d. The transmission power of each node in this step E6 is denoted P.

[0125] In a first implementation of this step E6, none of the active nodes are aware of the phase The combination at the destination of the retransmissions by these nodes assets is then incoherent. The signal-to-noise ratio SNR i of the composite transmission channel established between the source and the destination via the nodes of the active set is expressed in the form: where N ois the spectral density of noise and interference and  i represents the active set.

[0126] In a second implementation of this step E6, each of the active nodes knows the phase Φ a, d . The combination at the destination of the retransmissions by these active nodes is then coherent, because each node can compensate this phase by a factor so that all the messages arrive at the destination at the same time, ensuring the coherent nature of the combination of these messages. The signal-to-noise ratio SNRi of the composite transmission channel is therefore expressed in the form:

[0127] More precisely, this second implementation of the transmission is carried out so that all the redundancies transmitted by the active nodes are received at the same time by the destination in a coherent manner. Thus, the composite channel in this case is expressed according to the following formula:

[0128] Such a transmission mode, called "equal gain combining", makes it possible to obtain, on the destination side, a coherent combination of all the signals emitted by the active nodes.

[0129] The message redundancy transmitted by each active node is the same. Such redundancy can be the RVO redundancy transmitted during the first PHI phase or any other message redundancy u i . The transmission of redundancies can follow a predefined order of starting positions of reading the circular buffer for a message from a repeating source. For example, with reference to [Fig. 4] for 4 redundancy blocks / versions, a systematic LDPC code and N the order can be Pos. 0, Pos. 2, Pos. 3, Pos. 1 and so on with RVO and RV3 the redundancy versions associated with Pos. 0 and Pos. 3 which can decode independently of the other blocks / versions (every second transmission is self-decoding).

[0130] In a third implementation of step E6, the system comprises a first group D i of active nodes knowing the phase Φ a, d and a second group E i of active nodes not knowing this phase Φ a, d , with  i = D i EU i . In this third implementation, the signal-to-noise ratio SNR i of the composite transmission channel is expressed in the form:

[0131] Upon receipt of the retransmission request, each active node belonging to the first group D i transmits, in a step E6', the same redundancy of the message sent by the source s i modulated by a phase factor and each active node belonging to the second group E i transmits the same redundancy of the said message emitted by the source s iwithout phase modulation in the same time interval, so that all these redundancies transmitted by these active nodes are received at the same time by the destination d.

[0132] This is the case, for example, during a transitional period in which the destination d has not yet been able to determine the information relating to the phase factors for all active nodes. Over time, destination d will be able to provide such information to all active nodes in the system, further improving the quality of transmission.

[0133] In this third implementation as well as for the other implementations, the redundancy of the message transmitted by each active node having decoded the message u without error i emitted by source s iis the same for each of these active nodes. Such redundancy can be the RVO redundancy transmitted during the first PHI phase or any other redundancy of the message u i . The transmission of redundancies can follow a predefined order of starting positions of reading the circular buffer for the same message from a source which would need to be retransmitted several times.

[0134] For example, with reference to [Fig. 4] for 4 blocks / redundancy version, a systematic LDPC code and the order can be Pos. 0, Pos. 2, Pos. 3, Pos. 1 and so on with RVO and RV3 the redundancy versions associated with Pos. 0 and Pos. 3 which can decode independently of the other blocks / versions (every second transmission is self-decoding).

[0135] 5.2. Determination of an active set

[0136] Reference is now made to [fig. 5], which represents in more detail step E5, in which the destination d determines an active set Ât, then requests to retransmit a redundancy of the message a that it could not decode without error to the active nodes constituting this active set  i . For the sake of readability, we will subsequently confuse a message a, the source s i who issued it (and whose destination seeks redundancy) and the index i of this source.

[0137] Step E5 includes a step E50 of determining such an active set  i , and a step E52 of transmission, to the active nodes constituting this active set of a retransmission request.

[0138] Determining the active nodes constituting the active set  i is carried out among the sources and relays of the system The retransmission request is addressed by the destination to the active nodes of the active set  i .

[0139] In an exemplary embodiment, determining the active set  i is performed using a utility metric M: A i -> F(A i ). The utility metric M(A i ) is determined by the destination for one or more subsets A i , these subsets A i , all being included in the set of sources and relays of the system. The destination selects subset A i presenting the utility metric M(A i ) the highest. In other words, the destination selects subset A i the most advantageous in the sense of this utility metric M.

[0140] More precisely, during step E50, the destination determines a utility metric M(A i ) for at least one subset Ai of given nodes. In this exemplary embodiment, the utility metric M is a function of the size |A i | of said subset A i and the quality of the channel established between the source and i and the destination d via the nodes belonging to the subset A i .

[0141] Subset A i whose determined utility metric is the largest is then selected as the active set  i .

[0142] In this case, it is then possible to determine the nodes constituting the active set  i among the nodes belonging to the set H i (previously obtained or determined by the destination). This allows the de facto exclusion of nodes that cannot help the destination receive a redundancy of the message sent by the source (i.e. contribute to retransmitting a redundancy of the message). In other words, the set H i includes the j nodes of such as represents the nodes that were able to decode a message from the source to the previous frame represents the complement (in the set of nodes whose destination was able to decode without error at the previous frame t - 1. 5.3. Mutual information

[0143] In an exemplary embodiment, the utility metric M of a subset A i given is proportional to a discrete input mutual information between the source s i and the destination d knowing the composite transmission channel h eq Ai established between the source and the destination via the nodes of subset A i . Thus, mutual information is a quantity representative of the quality of the channel established between the source s i and the destination d via the nodes of the subset A i .

[0144] Mutual information can be expressed as a difference between the entropy H(x i) of a message sent by source s i and the conditional entropy of said message emitted by the source knowing the message y t received by destination d.

[0145] The mutual information between input x s . and the output y D knowing h eq A . is denoted As is a sufficient statistic for the detection of x si , he comes

[0146] As described above, we have are respectively the entropy of x and the conditional entropy of x given y.

[0147] The entropy H(x si ) knowing that expresses itself (where is the modulation constellation, i.e. the set of symbols potentially transmitted):

[0148] Conditional entropy is expressed as follows:

[0149] From where: where represents the expectation with respect to the probability distribution

[0150] As is a function of y 1( from a L and SNR Ai = , it comes that the mutual information between y t and x s . is a function that depends on and the SNR associated with the equivalent channel. In other words:

[0151] To estimate , it is possible to use an integration method, for example Monte Carlo type, based on L samples drawn according to the distribution

[0152] The signal-to-noise ratio with P transmit power per node does not does not depend directly on the cardinality of the group of active nodes except for the reference scenario where In this case the addition of an active node makes it possible to aggregate its power, i.e., to have an equivalent reception power increased by P, i.e., SNR A . =

[0153] We can define the following formula to calculate the utility metric:

[0154] Where the denominator is an increasing factor with the cardinal of A i . This allows to reflect the fact that the more nodes are integrated into the subset A i increases, the better the quality of the channel established between the source and i and the destination d (so the better the mutual information). Therefore, this improvement in mutual information is weighted by a cost in terms of the number of nodes involved in a retransmission.

[0155] 5.4. Logarithmic Decay

[0156] In an exemplary embodiment, the denominator exhibits growth logarithmic. By logarithmic growth is meant at least logarithmic, that is, the growth of the denominator is either logarithmic or faster (linear, quadratic, exponential, etc.).

[0157] In this case, the utility metric M(A i ) of a subset A i is determined according to the following formula:

[0158] Where is a normalization coefficient, increasing with the cardinal of A i . By croissant, we mean (which allows a constant coefficient).

[0159] Here, the utility metric of a subset A i is therefore inversely proportional to an increasing function of the size of said subset A i said increasing function exhibiting logarithmic growth.

[0160] The choice of a denominator of the type comes from the fact that, in regimes power-limited, mutual information increases linearly with received power while it increases logarithmically at high SNR or band-limited regime. Thus, transmission by an additional active node (i.e. its integration into a subset by increasing the utility metric) is allowed if it contributes to a logarithmic increase in spectral efficiency. Taking into account the discrete nature of the channel inputs via mutual information is also important, because mutual information is capped by the number of bits q carried by the modulation. When the number of nodes included in the active set is very large, increasing the power only leads to negligible gains in spectral efficiency, much lower than a logarithmic increase. In one embodiment, the coefficient is constant regardless of , For example

[0161] In another embodiment, where n is a constant (not necessarily whole) greater than 1. In this case the denominator therefore shows exponential growth.

[0162] 5.5 Individual process of determining the active set

[0163] The general principle of obtaining the active set during step E50 has been described so far.

[0164] We now describe different examples of the active set determination process.

[0165] In a first determination process, shown [Fig. 6], the destination determines, during step E50 and for each source s i for which a retransmission of a redundancy can be requested by the destination, an active set  i This first process is called individual process, that is, source by source.

[0166] The destination initializes the process (E510), by initiating, in a step E520, an iteration loop for each source s i of all sources not decoded by the destination. The destination determines in a step E530 the set H i from a source i (on which we iterate), that is to say the set of all the nodes which can contribute to the retransmission of the message sent by this source s i (ie the nodes as formulated otherwise the nodes having decoded the message sent by the source si). The destination then determines, for this source s i , the active set  i which is associated with it during a step E540. The destination then checks in a step E550 whether there are still sources for which the active set  must be determined i corresponding. If so, the destination loops back to step E520, otherwise it ends the process in step E560.

[0167] 5.5.1. Exhaustive determination of the active set

[0168] In a first embodiment of step E540, called exhaustive determination, step E540 comprises: the determination of all the subsets A i of the set H i , the determination, for each subset A i , of its utility metric M(A i ) corresponding, the determination of the subset A i having the highest utility metric among the set of utility metrics determined as the active set  i from the source s i .

[0169] This embodiment is called an exhaustive process, because the destination determines an optimal subset in the sense of the utility metric exhaustively, that is, a metric is calculated for all possible configurations (i.e. all active subsets).

[0170] 5.5.2. Heuristic determination by decreasing SNR of the active set

[0171] In a second embodiment of step E540, the destination constructs the active set. To do this, step E540 comprises sorting the nodes j constituting the set H i in decreasing order of SNR, the construction of a subset A i initially equal to the empty set, an iteration on the nodes j constituting the set H i in decreasing order of SNR, an iteration loop including: if adding node j to subset A i improves the utility metric of subset A i , actually add this node j to the subset A i and continue the iteration, otherwise stop the iteration and select the subset A i as active set  i for the source s i .

[0172] This embodiment of step E540 is called determination by decreasing SNR of the active set. This determination by decreasing SNR has less complexity than the exhaustive determination, while making it possible to determine an active set which is an approximation of the optimal set in the sense of the utility metric M. In addition, in the “equal gain combining” embodiment described above, the active set determined by the determination by decreasing SNR is the optimal set in the sense of the utility metric M.

[0173] In such a case, the messages are combined in a coherent (i.e. non-destructive) manner as explained above in point 5.1.3. The equivalent channel is then a function of the gains | h a, d |. Thus, choosing the relay node with the best gain (i.e. the best SNR) allows choosing the optimal active set for all relay nodes.

[0174] Formulated differently, we can summarize this determination by decreasing SNR in that it includes the construction of a subset A i initially equal to the empty set, said construction comprising at least one iteration of the following steps: determining the node j not belonging to the subset A i having the highest signal-to-noise ratio (SNR), if the addition of said node j to subset A i improves the utility metric of subset A i , add node j to subset A i , otherwise stop the iteration, the subset A i thus constructed being the active set  i .

[0175] 5.6. Common process for determining the active set

[0176] A first process for determining the active set has been described, source by source (or "individual"). A second process for determining the active set, called the common process and represented [Fig. 7], is now described.

[0177] This common process differs from the source-by-source process in that several steps are shared between the different sources for which a retransmission of the message by redundancy is requested by the destination.

[0178] More precisely, step E50 of this common process determines an active set  i common to all sources i for which a retransmission of a redundancy is requested by the destination. Step E50 thus includes an initialization step E515. During this initialization step, a value I_MAX is initialized to the value 0.

[0179] The destination starts, in a step E525, an iteration loop on the sources si of the set of all sources not decoded by the destination. For the current source s i of the current iteration, the destination determines in a step E535 its set H i . The destination then determines the mutual information relating to a channel established between the source s i and the destination d via the nodes of the set H i . The destination then determines the mutual information in a step E545 (i.e. the mutual information between the source s i and the destination d through the nodes constituting the set . The destination then compares in a step E555 this mutual information with the value I_MAX which represents the largest value of mutual information relating to a channel established between the source s i and the destination d via the nodes of the set H i calculated so far. If (SNR H.) > I. _MAX, then this means that the set H i from the current source s i is the set granting to the source s i the best mutual information calculated so far, and the destination stores this pair (s) in memory i , H i ) in a step E565, then loops back to the start of the iteration in step E525, if there are still sources s i on which the destination has not yet iterated.

[0180] When all the sources to be processed have been processed, the destination obtains the best source-set pair. From this best pair, the destination determines the active set i for this source s i in a step E575. The destination selects this active set as the active set for all sources for which a retransmission of a redundancy is requested by the destination, and terminates the common process in a step E585.

[0181] Step E575 may comprise an exhaustive determination, as described in point 5.5.1. Alternatively, step E575 may comprise a determination by decreasing SNR as described in point 5.5.2.

[0182] In other words, we can summarize this common determination in that it includes the determination, for at least one source s i undecoded, of a set H i associated (i.e. the set) comprising the nodes having decoded without error the message sent by said source, and the determination of the mutual information relating to said set H i partner, and determining, among the at least one undecoded source, the source whose associated set has the highest relative mutual information, and determining the active set of said source thus determined, the active set of each of said at least one undecoded source being equal to said active set thus determined.

[0183] 5.7. Device

[0184] [Fig. 8] represents a destination intended for an OMAMRC telecommunications system with M sources, possibly L relays and a destination, M ≥ 2, L ≥ 0 according to an embodiment of the invention. Such a destination is capable of implementing the transmission method according to Figure 3.

[0185] A destination may comprise at least one hardware processor 51, a storage unit 52, and at least one network interface 53 which are connected to each other through a bus 54. Of course, the constituent elements of the destination may be connected by means of a connection other than a bus.

[0186] The processor 51 controls the operations of the destination. The storage unit 52 stores at least one program for implementing the method according to an embodiment of the invention to be executed by the processor 51, and various data, such as parameters used for calculations performed by the processor 51, intermediate data of calculations performed by the processor 51, etc. The processor 51 may be formed by any known and suitable hardware or software, or by a combination of hardware and software. For example, the processor 51 may be formed by dedicated hardware such as a processing circuit, or by a programmable processing unit such as a Central Processing Unit that executes a program stored in a memory thereof. The storage unit 52 may be formed by any suitable means capable of storing the program(s) and data in a computer-readable manner.Examples of storage unit 52 include non-transitory computer-readable storage media such as semiconductor memory devices, and magnetic, optical, or magneto-optical recording media loaded into a read / write unit.

[0187] Network interface 53 provides a connection between the destination and all nodes

Claims

CLAIMS 1. Transmission method intended for an OMAMRC (Orthogonal Multiple-Access Multiple-Relay Channel) telecommunication system, with N nodes and a destination (d), the N nodes comprising M sources (s1... , s M ) and possibly L relays (r1... , r L ), with M ≥ 2, L ≥ 0, comprising a first phase during which the destination receives first redundancies (RV0) of messages transmitted successively by the M sources, the message from a source having been coded before transmission by an incremental redundancy type coding comprising several redundancies and a second phase comprising the following steps implemented by the destination (d): - broadcasting a control message identifying one or more sources for which it has not decoded the said message sent without error, called non-decoded sources, - reception of at least one identifier from at least one source inot decoded by the destination transmitted by a first set of nodes comprising at least one node, taken from among the N nodes, having decoded without error said message from the source s i , - determination, among the nodes of the first set, of a second set of nodes, called the active set (Â i ), associated with the source s i , - issuing a request for retransmission of said message from source s i , to the nodes of the active set (Â i ), And - reception of the same second redundancy of the message from source s i transmitted simultaneously by at least two nodes of the active set (Â i ) in the same time interval.

2. Method according to claim 1, such as determining the active set (Â i ) includes, for at least one subset (A i) of nodes taken from the first set of nodes: the determination of a utility metric, depending on the size of said subset (A i ) and the quality of a channel established between the source s i and the destination (d) via the nodes of the subset (A i ), the determination, among the subsets of nodes taken from the first set of nodes, of the subset (A i ) whose determined utility metric is the largest as an active set (Â i ).

3. Method according to claim 2, such as determining the utility metric of a subset (A i ) includes: the determination of mutual information representative of the quality of a channel established between the source s i and the destination via the nodes of the subset (A i ), called mutual information relating to the subset (A i), the utility metric being a function of said mutual information thus determined.

4. Method according to claim 3, such that the utility metric of a subset is proportional to the mutual information relating to said subset.

5. Method according to claim 3, such that the utility metric of a subset is inversely proportional to an increasing function of the size of said subset (A i ), said increasing function exhibiting logarithmic growth.

6. Method according to one of claims 2 to 5, such as calculating the utility metric M(X i ) is performed for all subsets (A i ) taken from the first set of nodes.

7. Method according to one of claims 2 to 5, such that it further comprises the construction of a subassembly (A i) initially equal to the empty set, said construction comprising at least one iteration of the following steps: determining a node (y) outside the subset (A i ) having the highest signal-to-noise ratio (SNR), if the addition of said node (y) to the subset (A i ) improves the utility metric of the subset (A i ), add node (y) to subset (A i ), at the last iteration, the subset (A i ) thus constructed being the active set (Â i ).

8. Method according to one of claims 2 to 7, further comprising: determining, for at least one source (s i ) not decoded by the destination, of a set (H i ) associated including the nodes having decoded without error the message sent by said source not decoded by the destination, and the determination of the mutual information relating to said set (H i) associated, and determining, in a set of undecoded sources, the source whose associated set has the highest relative mutual information, and determining the active set of said source thus determined, the active set of each of said at least one undecoded source being equal to said active set thus determined.

9. System comprising M sources (s 1, ... , s M ), L relay (r 1; ... , r L ) and a destination (d), M > 2, L ≥ 0, for an implementation of a transmission method according to one of the preceding claims.

10. A computer program product comprising program code instructions for implementing a transmission method according to claim 1, when executed by a processor.

Images