Blind recognition algorithm for IRA codes in noisy environment

For noisy environment, the parity-check matrix of Irregular Repeat-Accumulate(IRA) codes is hard to reconstruct, moreover, the relationships of the large-scale complex interleaver are hard to recover. To solve the problems, a novel blind recognition algorithm is proposed. First, the code's length and synchronization are identified by applying rank criteria. Second, by implementing matrix transformation, the dual vectors of codewords are found. Then, by setting a threshold, the effective parity-check vectors of dual space are selected. According to the sparse characteristics of the IRA codes' parity-check matrix, the parity matrix can be reconstructed with effective parity-check vectors Finally, relationships of the interleaver can be recovered according to the characteristics of IRA codes. Simulation results show that the proposal can be used to estimate IRA codes encoding parameters and complete the blind recognition in the non-cooperative context with noise.