Method of transmitting a digital signal for a marc system with a full-duplex relay, a corresponding program product and relay device

1. A method of transmitting a digital signal for a network having at least four nodes comprising two transmitters, a full duplex relay, and a receiver, wherein said four nodes are connected together by non-orthogonal links implementing a spatially-distributed network code, the method comprising:

a coding step per transmitter, each of said two transmitter delivering a code word per block of K information bits, wherein said K information bits are associated with B code words of data streams;

a transmission step in which each of the two transmitters transmits its respective code words during N time slots;

an iterative joint detection/decoding step performed in the full duplex relay to separate data streams of code words coming from the two transmitters and to determine, per data stream of code words, a vector representative each of the respective code words;

a joint coding step in the full duplex relay for coding two vectors derived respectively from the data streams of code words coming from the two transmitters in order to determine redundancy information representative of two code words transmitted during N time slots;

a transmission step of the full duplex relay transmitting redundancy information during N time slots; and

sequencing in the full duplex relay such that the redundancy information as transmitted from the two transmitters during the N time slots corresponds to the two code words transmitted during N preceding time slots; and

sequencing in each of said two transmitters not to transmit a code word during N time slots after one of the two transmitters has transmitted B code words of data streams, wherein B is a parameter.

2. The method according to claim 1 , wherein the iterative joint detection/decoding step implements hard decoding.

3. The method according to claim 1 , wherein the iterative joint detection/decoding step implements soft decoding and the joint coding step implements soft coding and marginalization performed at bit level.

4. The method according to claim 1 , wherein the iterative joint detection/decoding step implements soft decoding and the joint coding step implements soft coding and compression.

5. The A method according to claim 4 , wherein the compression is performed using a minimum mean squared uncorrelated error criterion.

6. A non-transitory data medium including program instructions stored thereon and adapted to implement a digital signal transmission method in a network when said program instructions are loaded and executed in a Multiple Access Relay Channel (MARC) system, wherein the network has at least four nodes comprising two transmitters, a relay, and a receiver, said four nodes are connected together by non-orthogonal links implementing a spatially-distributed network code, wherein the program instructions comprise:

coding instructions that configure each of the two transmitter to deliver a code word per block of K information bits, wherein said K information bits are associated with B code words of data streams;

transmission instructions that configure each of the two transmitters transmits its respective code words during N time slots;

iterative joint detection/decoding instructions that configure the full duplex relay to separate data streams of code words coming from the two transmitters and to determine, per data stream of code word, a vector representative of received code word;

joint coding instructions that configure the full duplex relay to code two vectors derived respectively from the two data streams of code words in order to determine redundancy information representative of two code words transmitted during N time slots;

transmission instructions that configure the full duplex relay transmitting redundancy information during N time slots; and

sequencing instructions that configure the full duplex relay to transmit the redundancy information during the N time slots following the N times slots when the corresponding code words are transmitted; and

sequencing in each of said two transmitters not to transmit a code word during N time slots after one of the two transmitters has transmitted B code words of the data streams, wherein B is a parameter.

7. A reception method for a receiver of a Multiple Access Relay Channel (MARC) system, the method comprising:

a step of storing, in B+1 blocks, received data corresponding to pairs of B+1 code words transmitted respectively by each of two transmitters, each of the two transmitters transmitted during N time slots of a data stream, and representative of pairs of blocks of K information bits, which K bits are associated with B code words of redundancy information transmitted in a redundancy stream by a full duplex relay, each of the B code words being transmitted during N time slots such that the redundancy information transmitted during N time slots corresponds to the code words transmitted during N preceding time slots, where B is a parameter; and

a step of iterative joint detection/decoding in the receiver, performed in parallel on the B+1 received data blocks in order to separate each of the data streams received from each of the two transmitters starting from a first block, and also to separate the data streams coming from each of the two transmitters and from the full duplex relay starting from the following B blocks, and in order to act in parallel for the B blocks, per data stream coming from the two transmitters and per block, to determine a vector representative of K information bits associated with a code word by simultaneously decoding the data stream corresponding to each of the two transmitters and the associated redundancy stream coming from the relay and provided by the following data block.

8. A reception method for a receiver of a Multiple Access Relay Channel (MARC) system, the reception method comprising:

a step of storing in B+1 blocks received data corresponding to pairs of B code words transmitted respectively by each of two transmitters, each of the two transmitters transmitted during N time slots of a data stream, and representative of pairs of blocks of K information bits, which bits are associated with B code words of redundancy information transmitted by a relay in a redundancy stream, each of the B code words being transmitted during N time slots such that the redundancy information transmitted during N time slots corresponds to the code words transmitted during N preceding time slots, wherein B is a parameter; and

a step of iterative joint detection/decoding in the receiver in parallel on the B+1 received data blocks to separate as output from a first block of data streams received from the two transmitters, and as output from the following B−1 blocks the data streams coming from the two transmitters and from the full duplex relay, the B+1 blocks storing only redundancy information corresponding to the code words of the preceding data block, and for determining in parallel for the B blocks, per data stream coming from the two transmitters, and per block, a vector representative of K information bits associated with a code word by simultaneously decoding the data stream corresponding to the two transmitter and the associated redundancy stream coming from the full duplex relay and supplied by following data block.

9. A method of transmitting a digital signal for a network having at least four nodes comprising two transmitters, a full duplex relay, and a receiver, wherein the four nodes are connected together by non-orthogonal links implementing a spatially-distributed network code, the method comprising:

a coding step per transmitter, each of said two transmitters delivering a code word per block of K information bits, wherein said K information bits are associated with B code words of data streams;

a transmission step in which each of the two transmitters transmits its respective code words during N time slots;

an iterative joint detection/decoding step performed in the full duplex relay to separate data streams of code words coming from the two transmitters and to determine, per data stream of code word, a vector representative each of the respective code words;

a joint coding step in the full duplex relay for coding two vectors derived respectively from the data streams of code words coming from the two transmitters in order to determine redundancy information representative of two code words transmitted during N time slots;

a transmission step of the full duplex relay transmitting redundancy information during N time slots;

sequencing in the full duplex relay to transmit the redundancy information as transmitted during the N time slots following the N time slots when the corresponding code words are transmitted, wherein each of the two transmitter transmits a frame of B+1 consecutive code words, where B is a parameter and sequencing in the full duplex relay not to transmit redundancy information corresponding to the B+1 code word.

10. The method according to claim 9 , wherein the iterative joint detection/decoding step implements hard decoding.

11. The method according to claim 9 , wherein the iterative joint detection/decoding step implements soft decoding and the joint coding step implements soft coding and marginalization performed at bit level.

12. The method according to claim 9 , wherein the iterative joint detection/decoding step implements soft decoding and the joint coding step implements soft coding and compression.

13. The method according to claim 12 , wherein the compression is performed using a minimum mean squared uncorrelated error criterion.