Low Density Parity Check Code Design For Half-Duplex Cooperative Relay Channel

For the half-duplex relay channel, cooperative diversity can be achieved via the cooperation among different nodes in a relay network. The employing of relays can greatly enhance the system ability of anti-fading, increase the channel capacity and reduce the power cost . The integration of cooperation with coding has been proven to be a very useful technique to enhance the system performance. Coded cooperation can obtain both spatial diversity and code gain. In the wireless relay network, by processing and forwarding the received data from multi-links, network code can effectively improve the transmission efficiency of the relay node without loss of diversity gain .In order to improve the BER performance of the relay network, we propose the joint network coding and channel coding technique on the relay node.For the orthogonal multi-access relay channel of two sources and one relay system, two relaying strategies based on the joint network-channel coding are proposed in this paper, i.e. code coopertation based on bilayer network LDPC code technique , the joint rate-compatible LDPC code and network code technique. To design the optimal relay code which approach the capacity of the orthogonal multi-access relay channel, we modify the LDPC Gaussian approximation algorithm corresponding to each relay protocol. Based on the modified Gaussian approximation algorithm and the differential algorithm, we obtain the optimization process of optimal relay code ensemble for each code cooperation protocol.The performance of the MARC system with two sources and one relay node employing the proposed two coding schemes is simulated as follows. The results show that, using bilayer network LDPC code with near-optimal degree distribution, the gap between the threshold and the system performance is about 0.4dB, and the gap between the capacity bound and the decoding threshold is less than 0.7dB. When employing network coding technique and rate-compatible LDPC code with near-optimal degree distribution, the gap between the threshold and the system performance is about 0.52dB, and the gap between the capacity bound and the decoding threshold is less than 0.46dB.