An Improved Belief Propagation Decoding of BATS Codes

In this paper, we propose an improved belief propagation (BP) decoding algorithm for decoding BATS codes. In the traditional BP decoding, the BP decoder decodes the input packets only when the rank equals to the degree of the efficient matrix for a batch. For improving the decoding performance, our proposed scheme exploits the fact that some packets can be decoded even when the rank is smaller than the degree. We decode these packets in the process of calculating the rank of batch, which will not induce extra computational cost. The simulation results show that the decoding performance of our proposed method is better than the traditional BATS codes.