FR2934736A1 - Data packet transmission method for e.g. mesh wireless communication network, involves selecting transmission mode with or without implementation of coding based on comparison of transmission quality level with predetermined threshold - Google Patents

Abstract

The method involves determining global quality level of transmission between a source node and a destination node in a mesh communication network by considering a set of transmission links used or to be used for implementing a network coding. A transmission mode with or without implementation of the network coding is selected based on the comparison of the global quality level of transmission with a predetermined threshold. Independent claims are also included for the following: (1) a method for reception of a data packet by a destination node from a source node in a mesh communication network (2) a transmitted signal formed of a data packet (3) a computer program product comprising a set of instructions for implementing steps of a transmission method and a reception method (4) a storage medium for storing a set of instructions (5) a device for transmission of a data packet from a source node to a destination node in a mesh communication network (6) a destination node receiving a data packet from a source node.

Description

Method and device for transmission in a network-coded mesh network, receiving method and device, signal, computer program products and corresponding storage means FIELD OF THE INVENTION The field of the invention is that of data transmission systems in mesh communication networks. The invention applies equally to wired or wireless mesh communication networks. More specifically, the invention relates to data transmission techniques in synchronous communication networks that may be subject to fading and masking caused by fixed or moving obstacles. The invention applies more particularly in the context of the broadcasting of data packets, that is to say the cases where a node of the network, called source node or transmitter node, broadcasts data packets intended for another node, called destination node or final receiving node, through the communication network. The invention applies to the case of networks in which the data packets are relayed by one or more nodes of the network, called relay nodes, to the destination node. This transmission mode allows the reception node to receive a set of copies of the same data packet that will be used to improve the quality of service of the application. At least two copies of the copy set take distinct and predetermined transmission paths. Retransmission of packets by the nodes can be done using a switching mechanism (based on a table or retransmission rules) to relay the data to the destination nodes. A retransmission table is thus representative of a distribution and a bandwidth occupation in the network. 2. TECHNOLOGICAL BACKGROUND Communication in a wireless communication network is very frequently subject to interference and masking which can significantly alter the quality of the data received. Traditionally, several methods can be implemented to overcome these sources of imperfections. These methods are generally classified according to the type of diversity they bring: spatial diversity or temporal diversity. Spatial diversity, characterized by a repetition code, makes it possible to solve, in a wireless communication system, the effect of the masks by limiting the losses of data packets due to the physical obstacles which interrupt the communications. Time diversity, meanwhile, improves the robustness of the data by correcting some of the data packet errors related to poor transmission conditions in the network.

These two types of solutions are complementary and can be implemented in the same communication network. In a meshed wireless communication network, the application of a so-called spatial diversity method makes the network robust by increasing, for the destination nodes, the possibility of receiving or reconstructing a correct copy.

Indeed, the source node sends an original data packet which is, depending on the chosen repetition code, relayed or not by another node of the network, commonly called relay node. The selection criterion of the relay node, which is responsible for relaying the data packet, is fundamental and determines the performance of the network.

However, the reception conditions in a communication network may differ from one node to another. The repetition code used, in order to make the communications more robust against interference and masking, must therefore take into account the conditions of the transmission channel specific to each pair of source node / receiver node within the communication network.

In fact, by using an adaptive repetition code, each of the destinations can receive a high copy number, if the transmission conditions to this destination strongly corrupt the data packets, or else low in the opposite case. The copies of data received by the destination node are thus combined to achieve a bit error rate, also BER (for Bit Error Rate in English), the lowest possible. OBJECTIVES OF THE INVENTION The purpose of the invention is notably to overcome these disadvantages of the prior art.

In at least one embodiment of the invention, the invention aims to provide a packet transmission technique for optimizing bandwidth allocation. Another objective of at least one embodiment of the invention is to provide such a technique for optimizing error correction.

A complementary objective of at least one embodiment of the invention is to guarantee a quality of service of the application deployed on the communication network. Another objective of at least one embodiment of the invention is to provide such a technique that can be applied to a communication network 20 comprising a shared medium. Another objective of at least one embodiment of the invention is to provide such a technique which is simple to implement and inexpensive in computing resources. SUMMARY OF THE INVENTION In a particular embodiment of the invention, there is provided a method of transmitting packets of data content from a source node to a destination node in a mesh communication network. comprising a plurality of nodes including at least one relay node implementing or likely to implement a network coding for the transmission of said packets.

Such a transmission method implements the following steps in said relay node: determination of an overall quality level of the transmission between said source node and said destination node, taking into account all the transmission links used or likely to be used for an implementation of said network coding; selecting a transmission mode with or without implementing said network coding, as a function of a comparison of said overall quality level with at least one predetermined threshold.

The general principle of the invention is therefore based on a means which makes it possible to evaluate the performances, that is to say the quality of the communication, of the transmission mode with network coding. This means introduces an objective criterion which allows the nodes to select or not the network coding mode. Network coding is a technique for exploiting the wireless medium by transmitting a combination of packets as a single packet to multiple destinations. This technique can be advantageously used to free resources, that is to say the bandwidth, not necessary for some nodes, then reassign these resources for other nodes. However, this technique deteriorates the quality of the service of the application deployed on the network. Thus, the invention makes it possible to optimize the use of the available bandwidth, while taking care not to deteriorate the quality of the service on the network. More specifically, the transmission links made available can then be allocated to other requesting destination nodes, i.e., nodes desiring to receive additional copies of the transmitted data packets. In this way, the destination nodes can perform a more efficient decoding since these nodes receive more copies.

The invention therefore makes it possible to implement an adaptive redundancy for all the nodes of the communication network. Advantageously, the transmission method implements, in the case of a transmission mode with network coding, the following steps: receiving at least two distinct packets, at least one packet of which is representative of said content to be delivered said recipient node; binary combination of the data of said two separate packets forming a combined packet; - transmission of said combined packet.

An embodiment of the invention is described below in which the two separate packets are transmitted by a single source node. It should be noted that this embodiment is a special and non-limiting case, and that the invention can be applied to a network comprising several source nodes and in which the two distinct packets are transmitted by two distinct source nodes.

In an advantageous embodiment of the invention, if said selected transmission mode does not implement network coding, said relay node transmits at least said packet representative of said content, received from said source node. Thus, it is chosen not to implement network coding to have a better robustness in order to preserve the quality of service required by the application deployed on the network. Advantageously, the transmission method comprises a step of estimating said quality level over a predetermined duration. In this way, the transients are eliminated and a better estimate of the overall quality of the communication is obtained over a predetermined period of time. According to an advantageous characteristic, the transmission method comprises the following steps: - constructing at least a first retransmission matrix of the symbol packets of said data content, defining transmission links for each of said packets, without the use of a coding network; constructing at least a second retransmission matrix of symbol packets of said data content, defining transmission links for each of said packets, using a network coding; if said selected mode implements a network coding: selecting said second matrix; transmission of said combined packet taking into account said second matrix. Thus, switching from one retransmission matrix to another allows a rapid reconfiguration of the packet retransmission scheme in the network, and thus allows an optimization of the transmission mode, taking into account the quality of the communication.

In an advantageous embodiment of the invention, said quality level corresponds to a calculation belonging to the group comprising: a calculation of a bit error rate; a calculation of a symbol erasure rate. The calculation of the overall quality level of the communication may be based on one or more combined elements of this list. In addition, this list is not exhaustive. For example, it is easy to compare the result of a calculation of bit error rates with a correction threshold of the coding code corrector channel implemented between the source node and the destination node.

Similarly, it is easy to compare the result of a symbol erasure rate calculation with a correction threshold of a code when erase decoding is implemented. In a particular embodiment of the invention, said level of quality between a source node and a destination node takes into account a sum of the quality levels of each portion of the link between said source node and said destination node, for at least one of said links. Thus, the global quality level calculated by the relay node takes into account the quality levels calculated for each portion of the path between the source node and the destination node, or even each portion of each possible path between the source node and the destination node. Advantageously, the transmission method comprises the following steps: identification of a time interval freed by the transmission of a combined packet; - Re-transmission of one of said packets of said data content, in the time interval released. Thus, sending redundancy packets in the bandwidth freed by the use of the network coding makes it possible to reduce the failure rate of the decoding of data packets by nodes needing more copies for the correction of errors. . Advantageously, the transmission method comprises a step of inserting into at least one packet of said data content at least one information indicating whether said packet is a combined packet or not.

In this way, the destination node is aware of the use or not of a network coding, and can decode the received packet according to the coding used. For example, if the information indicates that a network encoding has been implemented, the receiving node knows that the packet it is to decode is a combined packet, and the receiving node uses the proper decode table.

In another embodiment, the invention relates to a method of receiving packets of a data content by a destination node transmitted from a source node in a mesh communication network comprising a plurality of nodes including at least one relay node implementing or likely to implement a network coding for the transmission of said packets.

Such a reception method implements the following steps, in at least one of said destination nodes: reception of at least a first packet representative of said content; detecting at least one information indicating whether said packet is a combined packet; if said packet is a combined packet: receiving at least one second packet; combining said first and second packets to obtain an uncombined packet transmitted by said source node.

Thus, the destination node of the packets transmitted according to the previously described transmission method, decodes the packets received taking into account information indicating whether a network coding has been implemented or not. Advantageously, the reception method comprises the following steps: determining at least one piece of information representative of a reception error rate; transmitting said information to at least one relay node. In this way, a relay node obtains information representative of a quality of a portion of a path, via the recipient nodes of the packets transmitted between a source node and these destination nodes, in order to perform the calculation of the overall quality level of communication. Advantageously, the invention relates to a signal transmitted according to the transmission method described above, formed of packets of data content from a source node to a destination node in a mesh communication network comprising a plurality of nodes at least a relay node implementing or likely to implement a network coding for the transmission of said packets. The signal is such that at least one of said packets received by said destination node corresponds to a binary combination of the symbols of two distinct packets transmitted by said source node, forming a combined packet, said packet being inserted into said signal when a with network coding is selected, said selection taking into account a comparison between an overall quality level of transmission between said source node and said destination node, taking into account all the transmission links used or likely to be used for an implementation of a network coding, and at least a predetermined threshold. Advantageously, at least one packet of said data content comprises at least one information indicating whether said packet is a combined packet or not. In this way, the use or not of a network coding for the transmitted packet is signaled so that the decoding by a destination node can be performed with the appropriate decoding table. In another embodiment, the invention relates to a computer program product downloadable from a communication network and / or recorded on a computer-readable and / or executable medium by a processor, characterized in that it comprises program code instructions for implementing the transmission method as described above and / or the receiving method as described above, when said program is run on a computer. In another embodiment, the invention relates to a storage medium, possibly completely or partially removable, readable by a computer, storing a set of instructions executable by said computer to implement the transmission method as described above and or the receiving method as described above. In another embodiment, there is provided a device for transmitting packets of a data content from a source node to a destination node in a mesh communication network comprising a plurality of nodes including at least one relay node setting implement or likely to implement a network coding for the transmission of said packets.

Such a device is implemented by a relay node and comprises: means for determining an overall quality level of transmission between said source node and said destination node, taking into account all the transmission links used or likely to be used for an implementation of said network coding; means for selecting a transmission mode with or without implementing said network coding, as a function of a comparison of said overall quality level with at least one predetermined threshold.

Advantageously, the transmission device comprises, in the case of a transmission mode with network coding: means for receiving at least two distinct packets, at least one packet of which is representative of said content to be delivered to said destination node; binary data combining means of said two separate packets, forming a combined packet; means for transmitting said combined packet. Advantageously, if said selected transmission mode does not implement network coding, said device comprises means for transmitting at least said packet representative of said content, received from said source node. According to a particular characteristic of the invention, the transmission device comprises means for estimating said quality level over a predetermined duration. Preferably, said device comprises: means for constructing at least a first retransmission matrix of the symbol packets of said data content, defining transmission links for each of said packets, without using a network coding; means for constructing at least a second retransmission matrix of the symbol packets of said data content, defining transmission links for each of said packets, using a network coding; if said selected mode implements a network coding: means for selecting said second matrix; transmission means of said combined packet taking into account said second matrix. Preferably, said quality level corresponds to a calculation belonging to the group comprising: a calculation of a bit error rate; a calculation of a symbol erasure rate. According to an advantageous embodiment of the invention, said level of quality between a source node and a destination node takes into account a sum of the quality levels of each portion of the link between said source node and said destination node, for at least one of said links.

Advantageously, the transmission device comprises means for inserting into at least one packet of said data content at least one information indicating whether said packet is a combined packet or not. According to another embodiment, the invention relates to a destination node receiving packets of a data content from a source node through a mesh communication network comprising a plurality of nodes. Such a destination node comprises: means for receiving at least a first packet representative of said content; means for detecting at least one information indicating whether said packet is a combined packet; if said packet is a combined packet: means for receiving at least one second packet; means for combining said first and second packets to obtain an uncombined packet transmitted by said source node. Advantageously, the destination node comprises: means for determining at least one piece of information representative of a reception error rate; means for transmitting said information to at least one relay node. 5. LIST OF FIGURES Other features and advantages of embodiments of the invention will become apparent on reading the following description, given by way of indicative and nonlimiting example (not all the embodiments of the invention are not limited to the features and advantages of the embodiments described hereinafter), and the accompanying drawings, in which: FIG. 1 shows an example of a communication network in which a destination node can receive several copies of the same data sent by a source node; FIG. 2 represents a block diagram of a transmission device according to a particular embodiment of the method according to the invention; FIG. 3 shows an example of a communication network in which the transmission method according to a particular embodiment in accordance with the invention can be implemented, without network coding; FIG. 4 shows an example of a communication network in which the transmission method according to a particular embodiment according to the invention, with network coding, can be implemented; FIG. 5 illustrates an example of curves representing the total bit error rate with network coding (BER NC) and that obtained without network coding (BER); FIG. 6 shows a diagram of a communication network in which the transmission method according to the preferred embodiment of the invention can be implemented; FIG. 7 represents the structure of a data superframe transmitted within the communication network according to one particular embodiment of the invention; FIG. 8 presents a flowchart of a particular embodiment of a table activation / deactivation algorithm according to one particular embodiment of the invention; FIG. 9 presents a flowchart of an algorithm for estimating the bit error rate, according to a particular embodiment of the method according to the invention; FIG. 10 presents a flowchart of a particular embodiment of the reception method of the invention. DETAILED DESCRIPTION The general principle of the invention is based on an evaluation of the quality of the communication to determine whether a network coding can be applied or not. On a mesh network, access to the medium is based on a TDMA protocol (Time Division Multiple Access), where each node can access the medium in turn, cyclically. More specifically, packet transmissions are organized in a temporally ordered sequence that repeats itself periodically. This ordered sequence consists of time slot during which a node of the network has access to the network and can transmit information. In this type of network, the data packets can be relayed by all or some of the nodes of the network to the destination nodes.

This transmission mode allows reception by the final destination node of a group of copies that will be exploited to improve the quality of service of the application. These retransmission and reception rules can be put in the form of a retransmission matrix shared by all the nodes of the network. The transmission of several copies is advantageous since it increases the chances of receiving a copy with a low error rate. In addition, the transmission of several copies may also make it possible to combine only some of the copies, that is to say to make a selection of the copies likely to give the lowest bit error rate, to further improve the quality. on duty. In all the figures of this document, the elements and identical steps are designated by the same numerical reference. FIG. 1 shows an example of a communication network 100 in which a destination node can receive several copies of the same data item 10 sent by a source node. This is a wireless mesh communication system (Mesh in English) composed of several communication nodes referenced 110, 120, 130, 140, 150, 160. The links 111, 112, 113, 114, 115, 121, 131, 141 and 151 represent the path of different representative copies of the same set of symbols and received by the destination node 160. The source node 110 transmits a data packet. This same data packet is received by the relay nodes 120, 130, 140, 150 and by the destination node 160 via links 111, 112, 113, 114 and 115 respectively. The data packet is then relayed by the relay nodes 120, 130, 140, 150 via the links 121, 131, 141 and 151 respectively, to the destination node 160. This transmission mode therefore allows reception by the node 25. recipient 160, a group of copies that will be exploited to improve the quality of service of the application. Each copy is received with a certain error rate due to the interference related to the transmission links borrowed and the transmission channel. The copies of data, borrowing a plurality of different transmission paths, are therefore a priori not identical when they reach the destination node. The transmission method according to the invention is based on an evaluation, at the different relay nodes of the network, of the error rates related to the different transmission links of the different copies of data, to determine whether a network coding can be implemented. implement or not, for the continuation of the transmission of the data, from the relay nodes in question. We now present, in connection with Figure 2, a particular embodiment of a transmission device implementing the transmission method according to the invention.

The transmission device 200 comprises the following elements: a block of execution memory RAM 230 (for Random Access Memory in English); a non-volatile memory block ROM 240 (for Read Only Memory in English); a CPU 260 processing unit (for Central Processing Unit in English); a radio wave transceiver module 250 (or Radio Frequency Transceiver) for transmitting and receiving data via the transmission medium; a processing module 210 which communicates with the CPU 260 and the radio wave transceiver module 250. Moreover, according to the particular embodiment of the invention, the processing module 210 of the decoding device 200 comprises in particular : - an interface module CPU IF 211 which corresponds to the interface between the CPU and the baseband part (baseband in English). The CPU IF 211 interface module manages in particular the data exchanges between the different blocks of the processing module 210 (hereinafter detailed) and the CPU 260; A memory 212 for storing the data, in particular for storing the various tables determined during the execution of the allocation method; an encoder 217 and a decoder 218, for example of the Reed-Solomon type; an estimator 214 of the signal-to-noise ratio SNR (for Signal Noise Rate in English) required for the mechanism for calculating BER bit error rates (for Bit Error Rate) of the different links that form each transmission path. SNR signal-to-noise ratio measurement is performed for each received data packet. The signal-to-noise ratio SNR can be measured at the radio level, for example with a measurement of RSSI (for Received Signal Strength Indication in English), or else at the level of the modulation. A technique presented in the US patent application 2002/0041640 A1 proposes to make an estimate of the noise of the transmission channel, denoted B. This estimate is obtained by accumulating the measurements of the deviation of the power of each received symbol with respect to a theoretical value. The latter corresponds to the power of the symbol which has the smallest distance from the received symbol. The result is divided by the number of symbols considered. This corresponds to the application of the maximum likelihood criterion which equates a received symbol with its nearest neighbor in terms of distance. The formula (1) thus obtained is the following:

Emii [(s; -si)] B = i = 1 with i: an integer; k: the number of symbols per data packet; k (1) 10 15 20 25 min: the smallest Euclidean distance between a received symbol and a theoretical point of the modulation constellation; If *: the position of the received symbol; and Si: that of the theoretical symbol. When calculating the noise of the channel after receiving a data packet has been performed, the signal-to-noise ratio SNR can be estimated by dividing the average power per symbol over the noise power calculated according to the preceding equation (1). The formula (2) giving the signal-to-noise ratio SNR of a data packet is written as follows: k \ (sz / 12

SNR = 1 = 1 (2) Bk It should be noted that the SNR signal-to-noise ratio measurement, according to a particular embodiment according to the invention, is carried out permanently, that is to say as and as the data packets are received by the nodes. Once the SNR is measured, the bit error rate (BER) can be deduced, knowing the modulation used. A modulation curve that links the BER and the SNR must be known by the set of nodes to derive the values of the bit error rates for each SNR measurement. a table activation / deactivation controller 213 which executes the steps of the table activation / deactivation algorithm of the transmission method described below with reference to FIG. 8. The controller 213 integrates an error estimator 219, executing the steps described hereinafter with reference to FIG. 9. An example of a 300/400 communication network in which the transmission methods can be implemented is now presented with reference to FIGS. 3 and 4. and receiving according to one embodiment of the invention.

FIGS. 3 and 4 illustrate an example of calculating the overall quality level of the communication, in the case of a data transmission without network coding application (FIG. 3), and in the case of a transmission of data. data with network coding application (Figure 4).

The network illustrated in FIGS. 3 and 4 presents a source node 310/410 (#l) transmitting two dotted line and dashed packet streams, relay nodes 320/420 (# 2), 330/430 (# 3). ), 340/440 (# 4) and nodes 350/450 (# 5) and 360/460 (# 6), recipients of stream flows in continuous line and dashed line, respectively.

The links between these nodes (312/412, 313/413, 324/424, 334/434, 325/425, 336/436, 345/445 and 346/446) represent communication slots, or time slots, where a packet of data can be transmitted. When two links of the same nature (dashed line or solid line) start from the same node, these links correspond to one and the same communication slot.

It is assumed that the retransmission rules, in the case of a transmission without network coding (FIG. 3) are the following: - the node # 1 transmits a first stream of packets (in dotted line) in slots of time towards the node # 2 and a second stream of packets (in solid lines) to node # 3 in different time slots; - node # 2 transfers the dotted line packets to node # 4 (node # 5 is assumed to be within the radio range of node # 2); - node # 3 transfers the packets in solid lines to node # 4 (node # 6 is assumed to be in the radio range of node # 3); - Node # 4 has two types of time slots: the first to transfer packets in solid lines to node # 5 and the second to transfer dashed packets to node # 6. An evaluation of the communications performance between the source (node # 1) and the nodes # 5 and # 6 is possible using the following equations: (3): BER14_continuous = BER13 * (l-BER34) + BER34 * (l-BER13 ), with: - BER14 continuous: the bit error rate between source and node # 4, via node # 3, - BERij: the bit error rate between node #i and node #j. This equation (3) makes it possible to calculate the bit error rate for a packet sent by the source (node # 1) and received by the node # 4, via the node # 3 (solid line). The overall bit error rate for a packet transmitted by the source (node # 1) and received by the destination node # 5, via the nodes # 3 and # 4 (solid line) is written as follows: 4): BER_continuous = BER45 * (1-BER14continuous) + BER14continuous * (1-BER45), with: - BER continuous: bit error rate between the source and the node # 5, via the nodes # 3 and # 4, - BERij: bit error rate between node #i and node #j. Similarly, the overall bit error rate between source and node # 6 is calculated with the following equations: (5): BER14_deleted = BER12 * (1-BER24) + BER24 * (1-BER12), with: - BER14_pointed: the bit error rate between the source and the node # 4, via the node # 2, 20 - BERij: the bit error rate between the node #i and the node #j. This equation (5) calculates the bit error rate for a packet transmitted by the source (node # 1) and received by node # 4, via node # 2 (dashed line). The overall bit error rate for a packet transmitted by the source (node 25 # 1) and received by the destination node # 5 via the nodes # 2 and # 4 (dashed line) is written as follows: 6): BER_pointed = BER46 * (1-BER14_pointed) + BER14_pointed * (1-BER46), with: 30 - BER_pointed: the total bit error rate of the source (node # 1) at node # 6, via the nodes # 2 and # 4, - BERij: the bit error rate between nodes #i and #j. Assuming that the probability of having the same bit of the same erroneous packet more than once is zero, the preceding equations become: (4): BER continuous = BER13 + BER34 + BER45 (6): BER_pointed = BER12 + BER24 + BER46 We now consider the example illustrated in FIG. 4, with the application of a network coding by the node # 4. According to the retransmission rules of this example, the node # 4 applies a network coding and sends a combination of the packets received from the links 424 (in dashed line) and 434 (in solid line), in the form of a single packet, simultaneously at nodes # 5 and # 6 (445 and 446). A typical combination in the network coding and known to those skilled in the art can be obtained with one or exclusive (XOR) as follows: P (combined) = P (dotted XOR continuous) = P (dotted) XOR 15 P ( continuous), where P denotes a packet. With the packet P (dotted XOR continuous) and the packet P (dotted) transmitted by the node # 2, the node # 5 deduces the packet P (continuous) by applying one or exclusive in the following way: P (continuous) = P (combined) XOR P (dotted) # 6 applies the same method to deduce the packet P (dotted): P (dotted) = P (combined) XOR P (continuous) At node # 4, and after combination of the two packets P (dashed) and P (continuous), the bit error rate of the combined packet before it is transmitted by the node # 4 can be estimated by the following equation: (7): BER4 (combined) = BER14continuous * (l -BER 14_pointed) + BER14_pointed * (l -BER14_continuous) Node # 5 receives the packet encoded by node # 4 with a bit error rate that can be estimated by: (8): BER5 (handset) = 30 BER4 (handset) * (1iBER45) + BER45 * (1-BER4 (handset)) Node # 5 decodes packet P (handset) with packet P (dotted) received from node # 2 by applying a or exclusive. The total error rate of the packet P (continuous) thus obtained after decoding by the node # 5 is given by the equation: (9): BER continuous CR = BERS (combined) * (1ùBER15) + BER15 * (1 - BERS (combined)), with: - BER15 = BER12 * (lùBER25) + BER25 * (lùBER12), - BER continuous CR: bit error rate between source and node # 5, with a network coding applied by the node # 4.

In the same way, we deduce the overall BER_pointilléCR bit error rate of the packet P (dotted) decoded by the node # 6. FIG. 5 shows an example of curves illustrating the comparison between these two BER (BER curve) and BER continuous BER (BER NC) error rates, respectively, in the case where network coding is used. is not used and in case the network coding is used. These two curves show that the network coding increases the number of errors in the packet decoded by the destination. This is due to the application of the exclusive or that adds the errors of the packets during the coding and decoding operations. Network coding, which makes it possible to optimize the bandwidth used since only one transmission is necessary instead of two, results in a loss of quality of service, since the total bit error rate is higher. Thus, a measure of the overall quality of the communication may consist of a calculation of the overall bit error rate, and constitute a selection criterion between a transmission with network coding and a transmission without network coding. Once the global quality criterion has been chosen, it is compared to a threshold, in order to choose whether or not to use the network coding. For example, the bit error rate given by equation (9) can be compared with a correction threshold corresponding to the correction capability of the channel coding applied between the source and the destination. In this case, the comparison must be made for the two final destinations concerned by the network coding, in this case, in the example of FIG. 4, the nodes # 5 and # 6. The equation (9) for calculating the bit error rate after network decoding can be generalized to the case where several packets (1, 2, 3, ..., n) are combined to implement the network coding. In this case, the expression of the bit error rate after network decoding of n combined packets can be written in: BER.BERk ... BERI fl (1ù BER), j, k, ..., 1 1 2i +1 with m = [n / 2] if n is odd and m = [n / 2] -1 if n is even The sign [.] Corresponds to the integer part Now we have in relation to figure 6 an example of a communication network in which the transmission and reception methods according to the preferred embodiment of the invention can be implemented.

Thus, the invention is realized in the case of a wireless mesh communication network 600 where each node transfers to the other nodes of the network, in a synchronous manner, data specific to each destination node. To illustrate this case, we consider in the following description a 60 GHz synchronous communication system consisting of 9 transmission nodes, each of which implements the transmission device described above (Figure 2). Thus, each of the nodes 1a, 2a, 3a, 4a, 5a, 6a, 7a, 8a, and 9a, integrates a synchronous communication module, SCM (for Synchronous Communication Module in English), which integrates the various means for executing the transmission method according to the invention. More particularly, the communication system 400 comprises: m BER =

i = = 0 - eight knots 1a, 2a, 3a, 4a, 6a, 7a, 8a, and 9a of the WAR type (for Wireless Audio Renderer in English or English wireless audio receiver), each of which is equipped with means for restitution digital audio channel (or Digital Audio Channel Amplifier), respectively lb, 2b, 3b, 4b, 6b, 7b, 8b, and 9b each integrating a speaker (speaker in English); a node 5a of the WAD type (for Wireless Audio Decoder in English or French-language wireless audio decoder), comprising a multi-channel audio decoder SSD (for Surround Sound Decoder in English), respectively 5b, for example integrated in a flat screen and capable of transmitting via the communication system 600 and, in a perfectly synchronized manner, the different audio channels associated with the video displayed on the screen. This communication system 600 enables the node WAD to communicate to the various nodes WAR, possibly by relay performed by the nodes WAR, the information related to the different audio channels, according to an organized data transmission by frame, as represented in FIG. -after. FIG. 7 shows the structure of a data superframe transmitted within the communication network 600 according to a particular embodiment of the invention. During a synchronous data transmission cycle SDTC, each communication module SCM # i (i identifying the index of the node or the module) or node of the network, transmits a radio packet Pk_radio 710, in a predefined sequence. It should be noted that a superframe is emitted at each SDTC cycle. A radio packet consists of three data fields 720: an RPH (Radio Packet Header) field, also called a Radio packet header, which makes it possible to identify, for example, the number of superframes transmitted since the start-up. of the system, as well as an identifier of the transmitted radio packet. The header contains a bit denoted M (block 729) which makes it possible to activate or deactivate the retransmission table with network coding; a field of useful transmitted data or RPP field (Radio Packet Payload in English), making it possible to transmit the corresponding data to the digital channels of each WAR node, 721, 722, 723, 724, 725, 726, 727, 728 respectively for the nodes 1a, 2a, 3a, 4a, 5a, 6a, 7a, 8a and 9a; an information data field or RPI 730 field (Radio Packet Information) making it possible, in the context of the invention, to broadcast information concerning the quality measurements of the radio links between each node of the network. Each received measurement here corresponds to a SNR signal-to-noise ratio measurement performed by an SCM receiver node # i during the transmission of a radio packet by an SCM node j. More precisely, the RPI field (730) consists of: a subfield of so-called original information ORPI (740) (in English Original Radio Packet Information), corresponding to the SNR information measured by the node SCM # i during the previous radio packet transmissions by other SCM # x nodes in the network. The (k-1) measures are denoted by Mai (740), where the index i identifies the node having carried out the measurement (here the SCM # 1) and the index j identifies the node with respect to which this measurement is performed (node SCM # 2 for M1,2, SCM # k for Mu, (); - an ORPI data protection CRC subfield allowing the receiving node information, to ensure that the information has not corrupted by disturbances related to the radio channel - a relay sub-field of the data measured by the other nodes (other than SCM # 1 in the example of the figure) noted RRPI (in English Relayed Radio Packet Information ), to ensure the reception by all network nodes of SNR measurements even in the presence of fixed or temporary obstacles in the network.This subfield RRPI is constituted by the concatenation of the (k-1) ORPIs. and their associated CRCs received from the other nodes.

It is important to note that the ORPI and RRPI information is transmitted for each SCM, in a fixed order and known to the other nodes of the system to enable them to identify without additional information the content of the information. The nodes will thus constitute a table containing the information of the SNR measured by all the nodes of the network.

It should also be noted that when relaying information, an SCM node, having received an erroneous ORPI subfield from another SCM node (or an erroneous RRPI subfield of another SCM node), can detect this erroneous field using the associated CRC subfield. By default, if the SCM node must relay information for an SCM node for which it does not have correct information, it will position in the corresponding sub-field information of non-provision of the information. FIG. 8 presents a flowchart of a particular embodiment of the table activation / deactivation algorithm of the transmission method according to the invention.

The mechanism of switching from a retransmission table without network coding to a retransmission table with network coding, and vice versa, is implemented according to the algorithm 800 of the transmission method of the invention. Step 810 corresponds to the initialization of the method in which all the parameters to be estimated are set or reset.

The estimation of the bit error rate between each pair of nodes of the network is made at step 820, described by the algorithm 900 detailed below in connection with FIG. 9. Step 830 estimates the rates of total binary error, according to equation (9), used by the test of step 840.

This step 840 compares the previously estimated value with the maximum bit error rate tolerated by the channel coding implemented between the source and the destinations. If the result of the test performed in step 840 is positive, the network coding may be applied since estimated bit error rates, with network coding, guarantee the quality of service required by the two considered destinations. The algorithm 800 is periodic and depends on the predefined number of packets used during the estimation carried out at step 820. After the activation or deactivation of the retransmission table operated at steps 860 and 850 respectively, the algorithm 800 returns to the initial state and resets the variables to be estimated. This particular embodiment takes into consideration a bit error rate calculation as a criterion of the overall quality level of the communication. According to a second particular embodiment described below, for example in the case where the nodes of the network receive several copies of the same original packet sent by the source node via different relay nodes, the erasure number and not the number of erasures are considered. Indeed, the determination of the number of erasure can be advantageous when using a Reed-Solomon type decoding, since information on the position of the erasures can double the correction capacity of this decoder.

Of course, the present invention also applies in the context of any other decoder allowing decoding by erasure. The direct method for determining erasures is to compare, symbol to symbol, all received copies. When the same symbol is different for each copy, it is considered wrong. Conversely, the symbol is considered correct if at least two values of the same symbol are identical among the received copies. The decoder can correct a maximum deletion number equal to twice the number of errors in the symbol that can be corrected. The following equation makes it possible to estimate the number of deletions in the case of the symbol-to-symbol comparison of M copies each having a BERS bit error rate: kv "(1) (~ NEffacements = ù Li fJ û BER; N i = 1 j = 1j ≠ i with: - NEffacements: the number of symbol erasures; - BER): the bit error rate of the copy j; - k: the size of a bit copy - N: the number of bits per symbol - M: the number of copies Consider in a first example a comparison between two copies 10 of the same packet, transmitted by the source node, and received by the final destination. The probability of having two symbols of the same position in the frame, taken in a random manner in the different copies, is equal to: P = P + P 1. In this case, we neglected a term assuming that the probability 15 for these two erroneous symbols to be equal is zero by multiplying the probability obtained by the total number of symbols that can be parer (that is, the number of symbols per packet), the number of different symbols in the two copies is obtained. This number is equal to:

20 NEffacements = YN • (P + Pj) and corresponds to the number of erasures obtained by comparing the two symbol-symbol copies. In a second example, a comparison is made between three copies.

In this case, the formula obtained is similar but of the third order. With the same hypothesis (the probability that two erroneous symbols are equal is zero), we obtain the following formula for the copies i, j, l: P = PPj + POP + PP. This formula corresponds to the probability that the three copies give a different value for the same symbol. (10) The number of erasures obtained is equal to: NEffacements = N (PP + P'PI + PP1) The generalization of these formulas with m packages gives the equation (10) previously described.

FIG. 9 shows an algorithm for estimating the bit error rate of each of the links of the network, according to a particular embodiment of the transmission method according to the invention. This embodiment of the algorithm 900 corresponds to step 820 of the algorithm of FIG. 8.

Steps 910, 920, 930, and 940 initialize the process for the link between node #i and node #j. Each node of the network calculates the bit error rate for each packet received from each node, as well as the bit error rates of the links between the other nodes of the network by retrieving the SNR information obtained using of the SNR diffusion mechanism described previously. Each estimate of the SNR gives a bit error rate. The latter is added, in step 950, with the values of the previously estimated error rate. Step 960 tests the number of frames estimated with N_MAX (N_MAX is the number of frames used for the estimation of the bit error rate).

If the estimated number of frames is less than N MAX, the estimate continues, otherwise the average value of the bit error rate for the link between the nodes #i and #j is deduced by dividing, at the same time. Step 970. Steps 980 and 990 implement tests to browse the different links of the network to be estimated.

The retransmission table is activated or deactivated (according to the algorithm described previously with reference to FIG. 8) after each cycle of estimation of the bit error rate with network coding. This cycle is defined by the number of frames used for this estimate (that is, NMAX).

Once the cycle is complete, the node that executes the algorithm 800 decides to apply or not the network coding according to the result of the estimate. If the retransmission table with network coding is chosen, this node informs the destinations concerned by the network coding using the M bit (block 729).

If the bit M is "0", the table with network coding is not activated, otherwise, if M is equal to "1", the destinations decode the packet transmitted by the node applying the network coding with the retransmission table with network coding. FIG. 10 presents a flowchart of the reception method according to a particular embodiment of the invention.

Step 1010 corresponds to the initialization of the method where all the parameters are set or reset. A first packet, transmitted according to the transmission method described above, is received during step 1020. In order to know the transmission mode, with or without network coding, used to transmit this packet, it is detected, in step 1030, information as to whether the packet being decoded is a combined packet (a network coding has been used) or not (no network coding). If this information indicates that the packet is a combined packet, then the receiver knows that a network coding has been used, and step 1040 is to receive a second packet, necessary for decoding the combined packet. The first and second packets are then combined at step 1050 using the appropriate decoding rules (previously described) to obtain an uncombined packet transmitted by the source node. If the information indicates that the packet is not combined, then it is decoded in a step 1060, according to the decoding rules to be used in the case where no network coding has been implemented.

Claims (4)

REVENDICATIONS1. A method of transmitting packets of data content from a source node (410) to a destination node (450, 460) in a mesh communication network comprising a plurality of nodes including at least one relay node (440) setting implemented or capable of implementing a network coding for the transmission of said packets, said method being characterized in that it implements the following steps in said relay node: - determination (830) of an overall quality level transmission between said source node and said destination node, taking into account all the transmission links used or likely to be used for an implementation of said network coding; selecting a transmission mode with (860) or without (850) implementing said network coding, as a function of a comparison (840) of said overall quality level with at least one predetermined threshold. 2. Transmission method according to claim 1, characterized in that it implements, in the case of a transmission mode with network coding, the following steps: - reception of at least two separate packets, at least two a packet 20 is representative of said content to be delivered to said destination node; binary combination of the data of said two separate packets forming a combined packet; - transmission of said combined packet. 3. Transmission method according to any one of claims 1 and 2, characterized in that, if said selected transmission mode does not implement network coding, said relay node transmits at least said packet representative of said content, received said source node. 4. Transmission method according to any one of claims 1 to 3, characterized in that it comprises a step of estimating said quality level 30 over a predetermined period. 15. Transmission method according to any one of claims 2 to 4, characterized in that said method comprises the following steps: - construction of at least a first retransmission matrix of the symbol packets of said data content, defining links 5 of transmission for each of said packets, without the use of network coding; constructing at least a second retransmission matrix of symbol packets of said data content, defining transmission links for each of said packets, using a network coding; if said selected mode implements a network coding: selecting said second matrix; transmission of said combined packet taking into account said second matrix. 6. Transmission method according to any one of claims 1 to 5, characterized in that said quality level corresponds to a calculation belonging to the group comprising: - a calculation of a bit error rate; a calculation of a symbol erasure rate. 7. A transmission method according to any one of claims 1 to 6, characterized in that said quality level between a source node and a destination node takes into account a sum of the quality levels of each portion of the path between said source node and said destination node, for at least one of said paths. 8. Transmission method according to any one of claims 2 to 7, characterized in that it comprises the following steps: - identification of a time interval released by the transmission of a combined packet; - re-transmission of one of said packets of said data content, in the time interval released. 9. Transmission method according to any one of claims 2 to 8, characterized in that it comprises a step of insertion into at least one packet of said data content of at least one information indicating whether said packet is a packet combined or not. A method of receiving packets of a data content by a destination node transmitted from a source node in a mesh communication network comprising a plurality of nodes, characterized in that it implements the following steps, in said destination node: reception (1020) of at least a first packet representative of said content; detecting (1030) at least one information indicating whether said packet is a combined packet; if said packet is a combined packet: receiving (1040) at least one second packet; combining (1050) said first and second packets to obtain an uncombined packet transmitted by said source node. 11. Reception method according to claim 10, characterized in that it comprises the following steps: determination of at least one piece of information representative of a reception error rate; transmitting said information to at least one relay node. A transmitted signal according to the transmission method according to any one of claims 1 to 9, formed of packets of a data content from a source node to a destination node in a mesh communication network comprising a plurality of nodes including at least one relay node implementing or likely to implement a network coding for the transmission of said packets, characterized in that at least one of said packets received by said destination node corresponds to a binary combination of the symbols of two packets separate signals transmitted by said source node, forming a combined packet, said packet being inserted into said signal when a transmission mode with network coding is selected, said selection taking into account a comparison between an overall quality level of the transmission between said source node and said destination node, taking into account all the transmission links used or likely to used to implement a network coding, and at least one predetermined threshold. 13. Signal according to claim 12, characterized in that at least one packet of said data content comprises at least one information indicating whether said packet is a combined packet or not. 14. Computer program product downloadable from a communication network and / or recorded on a computer readable medium and / or executable by a processor, characterized in that it comprises program code instructions for the implementation of the transmission method according to at least one of claims 1 to 9 and / or the method of reception according to at least one of claims 10 and 11, when said program is executed on a computer. 15. Storage medium, possibly totally or partially removable, readable by a computer, storing a set of instructions executable by said computer to implement the transmission method according to at least one of claims 1 to 9 and / or the method of reception according to at least one of claims 10 and 11. 16. Device for transmitting packets of a data content from a source node to a destination node in a mesh communication network comprising a plurality of nodes including at least one node relay implementing or likely to implement a network coding for the transmission of said packets, said device being characterized in that it is implemented by a relay node and in that it comprises: an overall quality level of transmission between said source node and said destination node, taking into account all the transmission links used or likely to be used for an implementation of said network coding; means for selecting a transmission mode with or without implementing said network coding, as a function of a comparison of said overall quality level with at least one predetermined threshold. 17. Transmission device according to claim 16, characterized in that it comprises, in the case of a transmission mode with network coding: means for receiving at least two distinct packets, of which at least one packet is representative of said content to be delivered to said destination node; binary data combining means of said two separate packets, forming a combined packet; means for transmitting said combined packet. 18. Transmission device according to any one of claims 16 and 17, characterized in that, if said transmission mode selected does not implement network coding, said device comprises means for transmitting at least said representative packet of said content, received from said source node. 19. Transmission device according to any one of claims 16 to 18, characterized in that it comprises means for estimating said level of quality over a predetermined period. 20. Transmission device according to any one of claims 16 to 19, characterized in that said device comprises: means for constructing at least a first retransmission matrix of the symbol packets of said data content, defining transmission links for each of said packets, without the use of network coding; means for constructing at least a second retransmission matrix of the symbol packets of said data content, defining transmission links for each of said packets, using a network coding; if said selected mode implements a network coding: means for selecting said second matrix; transmission means of said combined packet taking into account said second matrix. 21. Transmission device according to any one of claims 16 to 20, characterized in that said quality level corresponds to a calculation belonging to the group comprising: - a calculation of a bit error rate; A calculation of an erasure rate of symbols. 22. Transmission device according to any one of claims 16 to 21, characterized in that said quality level between a source node and a destination node takes into account a sum of the quality levels of each portion of the path between said node source and said destination node for at least one of said paths. 23. Transmission device according to any one of claims 16 to 22, characterized in that it comprises means for insertion into at least one packet of said data content of at least one information indicating whether said packet is a packet combined or not. 24. Receiving node receiving packets of a data content from a source node through a mesh communication network comprising a plurality of nodes, characterized in that it comprises: - reception means from least a first representative packet of said content; Means for detecting at least one information indicating whether said packet is a combined packet; if said packet is a combined packet: means for receiving at least one second packet; means for combining said first and second packets to obtain an uncombined packet transmitted by said source node. 25. Destination node according to claim 24, characterized in that it comprises: means for determining at least one piece of information representative of a reception error rate; means for transmitting said information to at least one relay node.