WO2012114044A2 - Method for grouping stations together in order to optimize the broadcast of a broadcast/multicast stream - Google Patents

Abstract

The present invention relates to a method for transmitting at least one data stream from a station to N receiving stations, wherein N≥2. The stations, which are connected together via links, belong to a single logic network and, during transmission, implement multi-carrier modulation, each link between the station and one of the N receiving stations being associated with a table containing orders of constellations associated with the sub-carriers, respectively, said orders being determined according to one or more quality criteria of the link, said method consisting of: grouping the N stations into K groups, wherein K≥1, according to a similarity criterion between tables, and determining a common table per group; transmitting, via the station and to the K groups, respectively, a stream in the form of K streams by distributing the stream data over the sub-carriers according to the common table of the group.

Description

 Method of grouping stations to optimize the diffusion of

 broadcast / multicast

Field of the invention

 The present invention relates to the field of telecommunications. Within this field, the invention relates more particularly to the techniques implemented by a station for the transmission of a stream to several stations of its logical network. The logical network is the set of associated stations, i.e. that can exchange (useful) data with each other, regardless of the physical medium. Networks can be wired or wireless. The notion of flow corresponds to a sequence of data having a source address, a destination address knowing that the same sequence of data must be transmitted to one or more recipients.

 When a station has to transmit a stream intended for a single station of the network, it is broadcast in a unicast mode illustrated by FIG. This mode implies that a frame belonging to this stream has a destination address belonging to a single station. When a station must broadcast the same stream to several stations of the network, the broadcast can be performed in a multicast mode shown in Figure lb or in a broadcast mode shown in Figure 1c.

 In a wired network with support dedicated to transmission, the channel can be considered as ideal on each link, therefore it is not necessary to adapt the transmitted signal to the physical link that supports it. Broadcasting a stream for multiple recipients can be done in a multicast mode by setting a common destination address for all destination stations. In the case of an Ethernet network where the signal is transmitted in baseband, the broadcast mode "Broadcast" (which corresponds to a multicast mode where all stations are recipients), is achieved by forcing all the component bits to 1. the destination Media Access Control (MAC) address (FF address: FF: FF: FF: FF: FF).

 In a wired network shared between different types of transmitted signals, for example an electrical network which carries electrical supply signals and telecommunication signals, the transmission channel of the telecommunications signals moves away from an ideal channel. Thus, the CPL channel, Acronyms of Power Line Telecommunications (PLT), induces effects similar to those encountered in an air channel. In particular, the CPL channel, like the air channel, can lead to multiple paths during transmission. The multi-path effect encountered with the CPL channel originates in impedance mismatches at the ends of the electrical network that cause signal reflections.

The multipath channels, in a wireless or wired context, cause a time spread of the received energy. This characteristic results in fading (fading according to the Anglo-Saxon terminology) of the channel located in frequency, all the more numerous as the time spread of the impulse response of the channel is large.

In addition, the dynamics of the received signal can be very important, particularly in the context of wireless networks where the power of the radio waves decreases in 1 / d ² , d being the distance separating the transmitter from the receiver. This loss of energy can also be found in certain types of wired networks, such as CPL networks where the support of the transmission is none other than the electrical network, whose primary function is not to ensure the transfer of electricity. given by broad band modulations. As in wireless networks, the power of the signal received at a station of a CPL network depends largely on the distance between it and the transmitter and the frequency of the transmitted signal. Specifically to CPL networks, the fact that two stations belong to the same electrical circuit or to two different electrical circuits is also a factor strongly influencing the quality of the link separating them.

 To combat frequency fading of the channel specific to multipath channels, it is known to use multi-carrier modulations such as orthogonal frequency division multiplex (OFDM) which are particularly suitable for various reasons.

 On the one hand, they make it possible to limit the influence of intersymbol interference by using a symbol period that is much longer than the time spread of the impulse response. In cases where the OFDM modulation is coupled to a cyclic prefix, sometimes referred to as CP-OFDM, it may cancel this intersymbol interference.

 On the other hand, on the (narrow) band occupied by a given sub-carrier, it is reasonable to assume a flat frequency channel, which makes it possible to use a simple Zero-Forcing equalizer on reception.

 Several telecommunication standards include OFDM modulation. For examples, in the case of wireless networks, mention may be made of the DAB, DVB-T standards, certain standards of the IEEE 802.11 family, in the case of wired networks, mention may be made of ADSL or CPLs.

Furthermore, in the case where the channel is known to the transmitter, the multi-carrier modulations allow to adapt the transmitted signal to the channel in the presence. Since the sub-carriers can be modulated independently, it is possible to adapt the amount of information transmitted by subcarrier, by selecting a constellation by sub-carrier with a greater or lesser order depending on the attenuation specific to the subchannel corresponding to this sub-carrier. This technique, known under the English-bit-loading terminology, makes it possible to get as close as possible to the transmission capacity that the link can theoretically support. Thus, for each of its links shown in FIG. 2, a transmitting station has a constellation table known in the English terminology Tone Map, T _n , which indicates for each subcarrier the constellation (sub-order associated) to use. The selection of the constellation according to the attenuation of the subchannel therefore requires a return channel for the receiver to report to the transmitter the quality of the channel. To be effective, a bit-loading algorithm requires a quasi-static channel. This type of existing channel in the ADSL or CPL networks, the use of a bit-loading technique is therefore particularly suitable. When transmitting a stream, the data of this stream are mapped according to the indications contained in the constellation table of the link considered in order to best exploit its transmission capacity. The determination of the constellation to be applied to the subcarrier m is based on the well-known formula of the capacity:

 SINR (m)

où Q ¹ (x) est la fonction inverse de Marcum, avec Q (x) = -= \_e ^{[ 2 J} . Une sous-porteuse ne pouvant transmettre qu'un nombre entier de bits, il est nécessaire de définir un nombre b_m de bits par porteuse qui soit un nombre entier : C _m = log ₂ (l + (1) where the term SINR (m) corresponds to the signal-to-noise plus interference ratio and Γ is the margin on the SINR ("SINR gap"), defined for a symbol error rate ( SER) by [1]:

where Q ¹ (x) is the inverse function of Marcum, with Q (x) = - = \ _e ^{[2 J.} Since a subcarrier can only transmit an integer number of bits, it is necessary to define a number b _m of bits per carrier which is an integer:

where aj is equal to the largest integer immediately below a, IE _cons; is the finite set containing the amount of bits associated with the different constellations used by the system, d is the difference between | _C _m J and the immediately lower element contained in IE _cons; and b max is the largest element of E const

Thus, for a multi-carrier system comprising M _u useful subcarriers whose indices are classified in a set M _H and assuming that all the channels can use the same modulations, the table T _n constellations or "tone map" adapted to channel n is defined by:

T _n = {bme M _u }. (4) For each subcarrier m the constellation order therefore depends on the quality of the channel determined according to the relations (1) (2) for a symbol error rate and a power of the emitted signal which makes it possible to reach the SINR. This order is a maximum for the given quality of the channel, it is understood that a constellation with a lower order can be used but to the detriment of the flow.

 Prior art

When a gateway station between two different technology networks must transmit the same stream of data from the first network to several stations of the second network, the broadcast mode can be defined in a standard. Thus, the HPAV specification [2] and the recent IEEE P1901 standard [3] recommend the duplication of the multicast stream in N unicast stream, N being the number of receiving stations. In particular, when a CPL gateway station (conforming for example to the HPAV specifications) receives a broadcast broadcast data stream at its Ethernet interface, it must transmit it on its interface connected to the electricity network to all the stations. of the logical network to which it belongs. In order to transmit the stream, the CPL gateway station duplicates this stream as many times as there are stations to join and time multiplexes these N unicast streams in accordance with the constellation table T _n defined exclusively for each link n. The resource being shared in time between the different users, it appears that this method is not optimal and that it significantly reduces the overall capacity of the network. This limitation and degradation of the network are particularly damaging in the classic case of broadcasting a broadcast video stream to several stations of a CPL network.

 Presentation of the invention

 The invention proposes to improve the occupation of the time resource by a stream to be transmitted by a station to several destination stations.

 Thus, the subject of the invention is a method for transmitting a data stream from a first station to N receiving stations, N> 2, the stations, interconnected by links, belonging to the same logical network and setting during the transmission, a multi-carrier modulation, each link between the first station and one of the N destination stations being associated with a table containing constellation orders respectively associated with the subcarriers, these commands being determined according to at least one quality criterion of the link.

 The method comprises the steps of:

 to group the N destination stations in K groups, K> 1, according to a criterion of similarity between tables and to determine, by group, a common table,

 transmitting the stream by the first station in the form of K flows to the K groups respectively, using for each of the K groups the common table of the group for distributing the data of the stream on the subcarriers.

 When a stream must be transmitted to several stations, which typically results in the case of a stream broadcast in broadcast mode by an Ethernet network by a destination address FF: FF: FF: FF: FF: FF in the header of the incoming Ethernet frame at the first station, the first station proceeds according to the invention to a distribution of the different stations to join in subgroups of multicast, creating constellation tables specific to these subgroups respectively.

 Thus, instead of transmitting N unicast stream according to known techniques, the invention makes it possible to transmit only K multicast stream, K <N. Therefore, the invention makes it possible to save the temporal resource and, consequently, to improve the efficiency of transmission and optimize the use of link capacity.

According to one embodiment of the invention, the grouping is performed iteratively, using at each iteration a measure of distance between the tables. For a set {1, ..., N}, there exists max (C ^ XC _ ^~ _P ^F , CXC ^ Z _Q ) partitions having the form: {Ι,.,., Ρ}, {P + 1, ..., Q}, {Q + 1, ..., N}. An iterative method according to the invention makes it possible to quickly converge towards a partitioning solution approaching an optimal solution, by basing the grouping decision in subgroups on a measurement of distance between the tables. The table common to a subgroup is defined by taking for each subcarrier m, the most robust modulation, ie the lowest number of states, contained in the tables adapted to the links belonging to this subgroup.

 According to one embodiment of the invention, the distance measurement is based on the determination of the sum of the absolute difference in the number of bits transported by subcarrier between two tables.

 This measure has the advantage of being not very complex and therefore easy to implement.

 According to one embodiment of the invention, the distance measurement is based on the determination of the sum of the absolute difference of the number of bits transported by subcarrier between two tables weighted by a determination of an intersection between the two tables.

 This measure has the advantage of being more relevant than the previous measure to achieve a goal of reducing a parameter a evaluation of the occupation in time of an information bit.

 The invention furthermore relates to a station intended for a telecommunication logic network comprising the station and N receiving stations connected to the station by links, able to broadcast at least one data stream to the N destination stations, N> 2 .

 The station includes:

means for a transmission according to a multi-carrier modulation,

 means for determining by link a table containing orders of constellations associated respectively with the subcarriers, these orders being determined according to at least one quality criterion of the link,

 means for grouping the N stations in K groups, K> 1, according to a criterion of similarity between tables and for determining, by group, a common table,

 means for transmitting a stream by the station, in the form of K flows to the K groups respectively by distributing the data of the stream on the subcarriers according to the common table of the group.

 The means of the station are adapted to implement the transmission method according to the invention.

 The invention further relates to a telecommunication system comprising a station according to a previous object and N receiving stations connected to the station by logical links.

According to a preferred implementation, the steps of the method of transmitting at least one data stream from a first station to N destination stations, N> 2, are determined by the instructions of a transmission program incorporated into an electronic circuit such as a chip itself which can be arranged in an electronic device such as a transmitter. The transmission method according to the invention can just as easily be implemented when this program is loaded into an organ of calculation such a processor or equivalent whose operation is then controlled by the execution of the program.

 Accordingly, the invention also applies to a computer program, including a computer program on or in an information carrier, adapted to implement the invention. This program can use any programming language, and be in the form of source code, object code, or intermediate code between source code and object code such as in a partially compiled form, or in any other form desirable to implement a method according to the invention.

 The information carrier may be any entity or device capable of storing the program. For example, the medium may comprise storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording means, for example a floppy disk, a disk hard, a USB key, etc.

 Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the method in question.

 Furthermore, the program can be translated into a transmissible form such as an electrical or optical signal, which can be routed via an electrical or optical cable, by radio or by other means. The program according to the invention can be downloaded in particular on an Internet type network.

 Therefore, another object of the invention is a computer program on an information medium, said program comprising program instructions adapted to the implementation of a method for transmitting at least one data stream a first station with N destination stations, N> 2, implemented by the first station, when said program is loaded and executed in the first station intended to implement the transmission method.

 And the invention further relates to an information carrier comprising program instructions adapted to the implementation of a method for transmitting at least one data stream from a first station to N destination stations, N > 2, implemented by the first station, when said program is loaded and executed in the first station for implementing the transmission method.

 The invention furthermore relates to a method of reception by a destination station from among N destination stations, N> 2, of a data stream transmitted by a first station, the stations, linked together by links, belonging to a same logical network and implementing during the transmission a multi-carrier modulation, each link between the first station and one of the N destination stations being associated with a table containing orders of constellations associated respectively with the subcarriers, these orders being determined according to least one link quality criterion.

The method comprises the steps of: receiving a common table transmitted by the first station, following the grouping of N destination stations in K groups, K> 1, according to a criterion of similarity between tables and the determination, by group, of a common table,

 demodulating the data of the stream transmitted by the first station using the common table of the group to which the destination station belongs.

 The subject of the invention is furthermore a station intended for a telecommunication logic network comprising a first station and N receiving stations whose station, connected to the first station by links, N> 2, each link between the first station and a station. N destination stations being associated with a table containing constellation orders respectively associated with the subcarriers, these orders being determined according to at least one quality criterion of the link, able to receive a data stream transmitted by the first station to the N receiving stations.

 The station includes:

 means for demodulating the data of the stream transmitted according to a multi-carrier modulation, means for receiving a table common to a subgroup of destination stations comprising the station, containing orders of constellations associated respectively with the subcarriers, these commands being determined according to a similarity criterion between tables,

 means for extracting the distributed stream data on the subcarriers during transmission, using the common table.

 The means of the station are adapted to implement the reception method according to the invention.

 List of Figures

 Other features and advantages of the invention will become apparent from the following description given with reference to the accompanying figures given by way of non-limiting examples.

 Figure la illustrates a unicast broadcast mode of a stream.

 FIG. 1b illustrates a mode of multicast broadcasting of a stream.

 Figure 1 illustrates a broadeast broadcast mode of a stream.

 FIG. 2 is a diagram illustrating a logical network to which four stations STA1, STA2, STA3 and STA4 belong and illustrating a link 124 between the station STA2 and the station STA4.

 FIG. 3 is a diagram illustrating a logical network to which belongs a relay station STA0 and three stations STA1, STA2 and STA3, the station STA0 transmitting a stream to the three stations STA1, STA2 and STA3 and illustrating a link 120 between the station STA2 and the STA0 station.

 Figure 4 is a flowchart of an embodiment of a method according to the invention.

FIG. 5 illustrates the temporal occupation of a sequence of 20000 bits for a method of the prior art.

Figures 6 and 7 illustrate the temporal occupation of a sequence of 20000 bits for a method according to the invention. FIGS. 8-10 contain curves corresponding to simulation results of the transmission of a stream to five stations, FIG. 8 corresponds to the known method recommended by the HPAV specifications, FIG. broadcast with a common table for the five links, Figure 10 corresponds to a partition in two subgroups with a transmission of two multicast streams to respectively two subgroups, according to the second embodiment of a method according to the invention.

 FIG. 11a shows simulation results making it possible to compare the performances of a method according to the invention as a function of the distance criterion used and to compare a method according to the invention with an optimal solution.

 Figure 1b is a magnifying glass on an area of Figure 11a.

 Figure 12 is a diagram of a station according to the invention.

 Figure 13 is a diagram of a destination station according to the invention.

 Description of an embodiment of the invention

A method according to the invention is placed in the context of a wireless or wired network in which the stations implement "bit-loading" algorithms that make it possible to determine a table T _n of constellations for each of the links between the stations. Thus, a station that can communicate with N stations comprises at least N tables. This may be more because there can be several tables per link, particularly in the case of a dynamic adaptation of the tables to the channels and / or in the case of a QoS quality of service differentiation (for example depending on the service

(type of flow, etc.)). A dynamic adaptation of the tables to the channels can be implemented in the case of a CPL network given that the channel varies during the period of the alternating current.

The invention takes place in the context where an STA station ₀ , which can play a gateway or relay role between two networks of different technology, must transmit a stream to N STA stations _{1 (} STA ₂ , STA _N of a logical network to which it belongs, illustrated by FIG. 3 in the case where N = 3. A logical network is a set of stations that can communicate with each other.The N logical links connecting this station STA ₀ to the other N stations STA _n , ne {1 , N}, allow to determine

Ν tables T _n which can be different between them considering that the links can be of different qualities. For example, the symbol error rate (SER) can be set according to a QoS criterion assigned to the recipients in a differentiated way, for example responding to contractual constraints, to a different tariff between recipients, etc.

Each coefficient b _m ⁿ of a table T _n corresponds to a number of bits transported by sub-carrier. Let B _{n be} the total number of bits transported per OFDM symbol to the STA _n station:

where M _u is the set of indices of the useful subcarriers.

The stream to be transmitted by the station STA ₀ relay is either an incoming stream of data (useful) broadcast upstream broadcast to Ν stations logical network to which the relay station or a traffic flow to diffuse Ν stations of the same logical network.

In the following, the tables T _n are considered in descending order of the total number of bits transported per OFDM symbol to the STA station _n : Bi> B ₂ >...> B _N A method according to the invention partitions the N STAi, STA ₂ , STA _{N stations} into subgroups by basing the partition decision on a similarity criterion of the tables T _n , n e. {1, ..., N}. The table common to a subgroup is defined by taking for each subcarrier m, the most robust constellation, ie the lowest number of states, contained in the different tables T _n corresponding to the links associated with this subgroup .

 The flowchart shown in FIG. 4 illustrates the progress of the method according to one embodiment of the invention.

 Method 1 is preferably carried out iteratively.

At a given iteration, the common table T _k adapted to the k ^th subgroup is such that the total number of bits transported per OFDM symbol (of fixed duration) to the stations of the subgroup k is equal to:

with] L _k = {1, ..., N}, (7) k

the k ^th subgroup is composed of the stations whose index is contained in L _k .

 The method according to the invention determines a partition of the N stations in K subgroups which makes it possible to minimize the time occupation of an information bit evaluated by the parameter a defined by the following relation: k = l k

At initialization 2, the set of N stations is partitioned into N subgroups, each subgroup L _k being formed by one of the N stations:

Initial partition: P = {L _k l L _k = {k}, \ / k G {1, ..., N}}. (9)

All the N stations are associated with their table T _n and the stream to be broadcast by the relay station is supposed to broadcast in unicast on the N links.

Each bit of information being duplicated N times and occupying the resource for a time inversely proportional to the number of bits determined by the coefficient b _m ⁿ associated with the subcarrier carrying them, it comes the initialization value of the parameter a:

 NOT

 (10)

It should be noted that the evaluation of the time occupancy of an information bit may furthermore take into account the effects which depend on the target technology and which are associated with the overhead ((overhead according to the English terminology). Saxon) header frames, inter-frame spaces, signaling traffic) induced by the establishment of multiple streams and the increase of additional signaling traffic induced by the choice of partition between the N stations. To this end, the method can take into account a parameter β function of the current partition of all N stations and varying over the iterations, which is added to the parameter a.

To measure the similarity between two tables, different criteria can be used. According to a first embodiment which corresponds to that illustrated in FIG. 4, the similarity between two tables T and T is measured by a distance e _p defined as follows:

The distances e _p between the different tables in the presence can be represented in the following matrix form, noting that e = e and that e = 0:

0 ^e 2, l ^e 3, l ^β Ν, 1

e2, l 0 ^e 3.2 ^β Ν, 2

E = ^e 3, l ^e 3.2 0 · ^β Ν, 3 (12) eN, l ^e N, 2 ^e N, 3 ^■ 0

 During initialization 2, the method calculates the initial values of the coefficients e of the matrix E.

 After the initialization step, the method proceeds iteratively. For an iteration, the sequence is as follows.

In a step 3, the method determines the pair of indices (i, j) with i ≠ j which corresponds to the smallest coefficient e _ij of the matrix E. Knowing this pair, the method determines a new partition of the N stations by establishing a common table from the tables T _i and T., determined as follows:

L _k , k £ {i, j}

new partition: P _c = {c _k I Lc _k = <L _i \ L. if k = min (, j)>} (13)

 0 if k = max (, j)

The new partition is therefore identical to the previous partition for all L _k with k £ {i, j) and includes a new subgroup resulting from the merger (L _{ (JL _i ) of subgroups L _{ and _j . method determines in this same step, the new number Bc _k total of bits transported per OFDM symbol to the stations of the subgroup k:

Bc _k = Σ m (b _m ⁿ ) (14) me "

as well as the new value of the parameter a, called current value noted a _c :

 N 1

 = Σ (15) k = 1 Bc ,,

In a step 4, the method tests whether the current value of the parameter a is smaller than the previous value: a _c "x?

If the test 4 is negative this means that the last merge of the subgroups L _{ and _j performed in step 3 did not make it possible to gain on the temporal occupation of a bit and thus had a negative impact on the partition. This new partition Lc _k is not preserved for the rest of the algorithm. In a following step, the method deviates from the subsequent treatments the coefficient e _ij giving it an arbitrary value that is very large, that is to say strictly greater than the values that can be taken by the coefficients of the table E, symbolized by the symbol "Infinite" in Figure 4. In a step 6, the method tests whether there remain coefficients e _p with p ≠ q of the matrix E different from infinity, ie different from the very large arbitrary value.

If the test 6 is negative, the process goes out of the loop and goes to step 7 because all the coefficients e were reviewed in step 3. The partition is the partition P that is to say the one preceding the last partition P _c determined in step 3.

 If the test 6 is positive, the process loops back to step 3.

If the test 4 is positive, ie the merging of the subgroups L _t and L _j carried out in step 3 makes it possible to gain on the temporal occupation of a bit, the method proceeds to a step 8 of updating.

 In step 8, the method updates:

- the parameter a by assigning it the current value: a = a _c

the partition P by assigning it the composition of the new partition: P = P _c the coefficients of the common table corresponding to the merger performed in step 3, this common table taking as index the smallest index of the pair (i, j ) (min (i, j)):. ,. = {min (V, b ^j ) J / me M} (16) - the table whose index is the largest index of the pair (i, j):

T,. _Λ = {0 / me M) (17) the coefficients of the rows and columns i and j of the matrix E: the coefficients of the line min (i, j) and of the column min (i, j) are set to day with the relation (11) because the table T _Mn ^ ^ corresponds to the table resulting from the merger carried out in step (3), the coefficients of the line max (i, j) and of the column max (i, j) are forced to a very large arbitrary value, that is to say strictly greater than the values that can be taken by the coefficients of the table E, symbolized by the symbol "infinite", so as to no longer take into account the coefficients of this line / column in the continuation of the treatment.

 In the following step 9, the method tests whether there are still at least two different L and L subgroups (p ≠ q) that are not empty.

 If the test 9 is positive this means that there is still a possibility of fusion to be tested, the method then loop back to step 3 to evaluate the impact of this merger.

 If the test 9 is negative, this means that there is only one subgroup left that contains all N stations, the process goes out of the loop and goes to step 7 because the algorithm has converged to the solution broadcast. The selected partition corresponds to that updated in step 8: P.

 Depending on the embodiment, the arbitrary value used in steps 5 and 8 may be a negative value. Indeed, because of the way in which the coefficients of the matrix E are determined, it is certain that these coefficients can not have a negative value. Forcing the already tested coefficients to a negative value allows them to be removed from the rest of the algorithm by a test on the sign.

In a step 7, which follows step 6 or step 9, the method transmits the flow in the form of K flows to respectively the K subgroups of stations. The method therefore duplicates the data flow in K identical data streams. For each of the K flows intended for one of the K subgroups, the method uses the common table of the subgroup, determined in step 8 to distribute the data of the subgroup. flow on the subcarriers (bit-loading). Thus, the common table of a subgroup from a merge made in step 3 is determined by equation (16).

According to a second embodiment, the similarity between two tables T p and T g is measured by a distance E which consists in weighting the computation of distance between two tables taking into account the intersection between the latter, that is to say of the common table resulting from the merger of these two tables.

 The increase in computation complexity of the coefficients is compensated by obtaining a measurement more relevant than the first measurement according to the first mode relative to the objective of reducing the parameter a.

 The method according to this second mode proceeds in the same way as for the first mode but replacing the equation (11) by the equation (18).

 A flow transmission method according to the invention is implemented by a station. Such a station can simultaneously transmit several incoming streams, for example of different nature (video, data, etc.). The method according to the invention applies both to an incoming stream and to several incoming streams.

 The stations of a subgroup are aware of the common table determined by the station which transmits the stream because the latter communicates this table to the stations of the subgroup. Each destination station can therefore demodulate the received data by using the common table used to modulate the data on transmission. Each received data frame may include in its preamble a table identifier to be used by the destination station.

 FIGS. 5-7 make it possible to compare the temporal occupation of a bit between a method of the prior art, FIG. 5, and a method according to the invention, FIGS. 6 and 7.

 Current specifications in CPL networks do not provide for the common Tone Map definition for streaming broacast / multicast streams over the network. The method described by HPAV for transmitting a stream to the attention of N stations and illustrated in FIG. 5 consists in duplicating this stream in unicast N times, each link having a Tone Map of its own. In the typical case of broadcasting a broadcast video stream and if the network comprises more than 3 or 4 stations, the application of this method significantly reduces the overall capacity of the network, the resource being shared in time between the different users . The temporal occupation of the diffusion of a sequence of 20000 bits according to the HPAV method is represented in FIG. 5. It results from the multiplexing of the five unicast streams.

This temporal occupation of the broadcasting of a sequence of 20000 bits is to be compared with that obtained by a method according to the invention with the creation of a common broadcast Tone Map. for the 5 stations, Figure 6, or with the creation of two common Tone Map corresponding to two multicast subgroups, Figure 7.

The curves shown in Figures 8-10 are simulation results of the broadcast of a stream to five stations. The context is that of the HPAV specifications: "OFDM windowed" multi-carrier modulation with M _u = 917 under useful carriers. The tables (Tone Map) were obtained by simulation by performing a "bit-loading" on class 2 PLC channels [4].

FIG. 8 corresponds to the known method recommended by the HPAV specifications which leads to a duplication of the broadcast stream in as many unicast streams as of destination stations. The transmitting station (relay) multiplexes in time the five identical flows mapped according to the five tables associated with the links respectively. The curve is a synthesis of the temporal occupation of the five unicast flows transmitted to the five destination stations respectively. This occupation is taken as initialization value OC _ini of the parameter of a method according to the invention.

 FIG. 9 corresponds to a transmission in a broadcast mode with a common table for the five links, which may in certain cases correspond to an embodiment of a method according to the invention (convergence of the algorithm to the broadcast mode, test 9 of the negative algorithm), the reduction in the temporal occupation is approximately 31% compared to FIG. 8. The reduction of the occupation in time is calculated as follows:

reduction = 100 x ^ ""^{'~ a} (19)

FIG. 10 corresponds to a partition in two subgroups with a transmission of two multicast streams to respectively the two subgroups, according to the second embodiment of a method according to the invention (distance E based on equation (18) ). A subgroup comprises stations 2 and 3 and corresponds to the Tmi curve, the other subgroup comprises stations 1, 4 and 5 and corresponds to the Tm ₂ curve. The reduction in time occupancy calculated according to equation (19) is about 37% compared with FIG. 8.

FIG. 1a show simulation results making it possible to compare the performances of a method according to the invention as a function of the distance criterion: criterion 1 = distance E based on equation (11), curve C1, criterion 2 = distance E based on equation (18), curve C2, and to compare a method according to the invention to an optimal solution curve Co. The comparison is placed in the context of a 5-link logical network with a random choice of 5 tables in a database containing 100 tables determined from 100 simulated PLC channels. 1000 prints are made, which corresponds to the horizontal scale allowing each time to compare the optimal solution, curve Co, to a method according to the invention for both criteria, the broadcast solution, curve C4, and, at the unicast solution according to the HPAV method, curve C3. The unicast solution serves as reference, curve C3, and sets the reference value 0 of the vertical scale which indicates the reduction of occupation in time with respect to this unicast solution. With each draw, the optimal solution is determined by evaluating the OC parameter for all possible partitions in all tables and retaining only the minimum value of parameter a

 Figure 1b is a magnifying glass on an area of Figure 11a. The comparison of the curves Co, Cl, C2, C3, C4 gives the following indications:

 criterion 2> criterion 1 in 33.40% of cases,

 criterion 2 and criterion 1 converge towards the same solution in 54.90% of cases, criterion 2 <criterion 1 in 11.70% of cases,

 criterion 2 converges to the optimal solution in 62.90% of cases,

 criterion 2 <optimum solution in 37.00% of cases,

 the standard deviation between the reduction in time obtained with criterion 2 and the optimal solution is 1.48%.

 Given the small standard deviation between the results, criterion 2 makes it possible to approach the optimal solution significantly. Finally, whatever the criterion used, a method according to the invention never degrades the time efficiency with respect to the reference solution unlike the broadcast solution which in some cases degrades performance, which is reflected in FIGS. ia and 11b by negative time reduction values (below the reference value 0 corresponding to the unicast solution).

 FIG. 12 is a diagram of a station (gateway, relay) able to broadcast at least one stream according to the invention. The STA0 station is intended for a telecommunication logic network comprising the station and N stations connected to the station by links, such a logical network is illustrated in FIG. 3. The station is able to broadcast at least one data stream to the stations. N receiving stations, N> 2. The station includes:

 means for transmission according to a multi-carrier modulation. These means typically comprise an OFDM MUX OFDM modulator which performs an inverse Fourier transform on NFFT under orthogonal carriers.

means for determining, by link n, a table containing the orders of the constellations associated respectively with the subcarriers, these commands being determined according to one or more quality criteria of the link. These means are typically calculation means, for example a microprocessor or a DSP (Digital Signal Processor), microprogrammed to calculate the equations (I) - (3) and deduce the tables T _n from the relation (4) for each link n.

 means for grouping the N stations in K groups, K> 1, according to a criterion of similarity between tables and for determining, by group, a common table. These means are typically computing means, for example a microprocessor, a DSP signal processor, microprogrammed to implement steps (2) - (9) of the method.

means for transmitting a stream by the station, in the form of K flows to respectively the K subgroups by distributing the data of the stream on the subcarriers according to the common table of the group. These means typically include a dithering module TR and a modulator OFDM. The module TR performs a duplication of the flux in K flux and needle on each subcarrier of the OFDM modulator, the number of bits corresponding to the order of the constellation indicated in the table (bit-loading) associated with the flow to which the pointed bits belong. .

 The means of the station are adapted to implement the transmission method according to the invention.

 FIG. 13 is a diagram of a station STA1, intended for a telecommunication logic network, able to receive a data stream transmitted by a first STAO station to N destination stations, according to the invention. The telecommunication logic network, illustrated in FIG. 3, comprises the first STAO station and N destination stations whose station STA1 is connected to the first STAO station by links, N> 2. Station STA1 includes:

 means for demodulating the data of the transmitted stream according to a multi-carrier modulation. These means typically comprise an OFDM OFDM OFDM demodulator which performs a direct Fourier transform on NFFT under orthogonal carriers.

 means for receiving a table common to a subgroup of destination stations comprising the station, containing orders of constellations respectively associated with subcarriers, these orders being determined according to a criterion of similarity between tables. These means are typically extraction means for extracting from a received data frame the coefficients of the common table, for example a microprocessor or a DSP (Digital Signal Processor) microprogrammed.

means for extracting the data of the stream distributed on the sub-carriers during transmission, by using the common table. These means typically comprise a stripping module TR ^"1. The module TR ^{" 1} extracts from each subcarrier of the OFDM demodulator, the number of bits corresponding to the order of the constellation entered in the common table (bit-loading ¹ ).

 The means of the station are adapted to implement the reception method according to the invention.

An example of implementation of the invention is as follows. The context is that of a relay station (/ gateway / bridge) between an Ethernet network and an electrical network used as a PLC network. In any Ethernet network, the mechanism for discovering the addresses of the stations is done through Adress Resolution Protocol (ARP) requests, particular frames broadcast in broadcast mode on the network (destination address FF: FF: FF: FF: FF: FF ). The CPL relay station receiving an ARP request on its Ethernet interface will relay this frame to all the stations of the CPL network according to the current recommendations defined by HPAV (Conversion broadcast - unicast). The transmission of this type of frame can advantageously be carried out by a method according to the invention since the latter makes it possible to limit the impact (in particular on the bit rate) of the broadcast of a broadcast transmission on the CPL network. [1] JM Cioffi, "Multicarrier primer", Nov. 1991, ANSI Contribution TlEl.4 / 91-157, Clearfield, Fia, USA.

 [2] HomePlug AV white paper, available at:

 http://www.homeplug.org/tech/whitepapers/HPAV-White-Paper_050818.pdf.

 [3] IEEE Standard for Broadband Power Line Networks: Medium Access Control and

Physical Layer Specifications, pp. 652-653, available at:

 http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=5678772&tag=l.

 [4] M. Tlich, A. Zeddam, F. Moulin, and F. Gauthier, "Indoor Powerline Communications Channel Characterization Up to 100 MHz-Part II: Time-Frequency Analysis," IEEE Trans. Power Del., Vol. 23, no. 3, pp. 1402-1409, Jul. 2008.

Claims

A method (1) for transmitting a data stream from a first station to N receiving stations, N> 2, the stations, interconnected by links, belonging to the same logical network and implementing when the transmission a multi-carrier modulation, each link between the first station and one of the N destination stations being associated with a table containing constellation orders respectively associated with the subcarriers, these orders being determined according to at least one quality criterion of the link, said process consisting of  to group the N destination stations in K groups, K> 1, according to a criterion of similarity between tables and to determine, by group, a common table,  transmitting the stream by the first station in the form of K flows to the K groups respectively, using for each of the K groups the common table of the group for distributing the data of the stream on the subcarriers. The method (1) for transmitting data streams according to claim 1, wherein the grouping is performed iteratively, using at each iteration a measure of distance between the tables. The method (1) for transmitting data streams according to claim 2, wherein the distance measurement is based on determining the sum of the absolute difference in the number of bits transported by subcarrier between two tables. The method (1) for transmitting data streams according to claim 2, wherein the distance measurement is based on the determination of the sum of the absolute difference of the number of bits transported by subcarrier between two tables weighted by a determination of an intersection between the two tables. 5. Station (STAO) for a telecommunication logic network comprising the station and N receiving stations connected to the station by links, able to broadcast at least one data stream to the N destination stations, N> 2, comprising:  means for transmission according to a multi-carrier modulation,  means for determining by link a table containing orders of constellations associated respectively with the subcarriers, these orders being determined according to at least one quality criterion of the link, means for grouping the N stations in K groups, K> 1, according to a criterion of similarity between tables and for determining, by group, a common table, means for transmitting a stream by the station, in the form of K flows to the K groups respectively by distributing the data of the stream on the subcarriers according to the common table of the group. 6. Station (STAO) according to the preceding claim wherein the means for grouping perform the grouping iteratively, using at each iteration a distance measurement between the tables.  7. Station (STAO) according to the preceding claim wherein the distance measurement is based on the determination of the sum of the absolute difference in the number of bits transported by subcarrier between two tables.  8. The station (STAO) according to claim 6 wherein the distance measurement is based on the determination of the sum of the absolute difference of the number of bits transported by subcarrier between two tables weighted by a determination of an intersection between the two tables. . 9. Telecommunication system comprising a station according to claim 5 and N destination stations connected to the station by logical links. 10. Computer program on an information medium, said program comprising program instructions suitable for implementing a method of transmitting at least one data stream from a first station to N destination stations, N> 2, implemented by the station according to any one of claims 1 to 4, when said program is loaded and executed in the station for implementing the transmission method. 11. An information carrier comprising program instructions suitable for implementing a method of transmitting at least one data stream from a first station to N destination stations, N> 2, implemented by the station according to any one of claims 1 to 4, when said program is loaded and executed in the station for implementing the transmission method. 12. Method of reception by a destination station among N destination stations, N> 2, of a data stream transmitted by a first station, the stations, interconnected by links, belonging to the same logical network and implementing during the transmission, a multi-carrier modulation, each link between the first station and one of the N destination stations being associated with a table containing constellation orders respectively associated with the subcarriers, these commands being determined according to at least one criterion of link quality said method comprising the steps of: receiving a common table transmitted by the first station, following the grouping of N destination stations in K groups, K> 1, according to a criterion of similarity between tables and the determination, by group, of a common table,  demodulating the data of the stream transmitted by the first station using the common table of the group to which the destination station belongs. 13. Station (STAl) for a telecommunication logic network comprising a first station and N receiving stations whose station, connected to the first station by links, N> 2, each link between the first station and one of the N receiving stations being associated with a table containing constellation orders respectively associated with the subcarriers, these commands being determined according to at least one quality criterion of the link, able to receive a data stream transmitted by the first station to the destination N stations, comprising :  means for demodulating the data of the stream transmitted according to a multi-carrier modulation, means for receiving a table common to a subgroup of destination stations comprising the station, containing orders of constellations associated respectively with the subcarriers, these commands being determined according to a similarity criterion between tables,  means for extracting the distributed stream data on the subcarriers during transmission, using the common table.