Research On Improved Decoding Algorithm Of Polar Codes Based On SC Algorithm

Polar codes,proposed by Arikan,are channel codes based on channel polarization,they have been proven to achieve channel capacity of binary-input discrete memoryless channel(B-DMC)with low encoding and decoding complexity.In this thesis,we study the successive cancellation(SC)and simplified successive cancellation(SSC)decoding algorithms of polar codes,and propose the improved algorithms,which has an important theoretical significance.The main results are as follows:Firstly,a soft-output decoding algorithm based on SSC is proposed aiming at the problem of no improvement in bit error rate(BER)performance compared with SC decoding algorithm.In the proposed algorithm,the soft-output information is introduced and the new message passing rules are proposed,the soft-output information can be transimitted in the code tree,which improves the BER performance of SSC decoding algorithm.Besides,the soft-output information can be directly returned from rate-zero and rate-one nodes because of the simplification for these nodes in SSC decoding,which enormously reduces the decoding latency of soft-output decoder.The simulation results show that the soft-output decoding algorithm based on SSC has better BER performance compared to SSC decoding algorithm and the proposed algorithm has lower decoding latency compared to SCAN under the same BER performance.Secondly,to solve the problem of large decoding latency of SC decoding algorithm for Polar codes,the SC decoding algorithm based on parallel processing is proposed.The proposed algorithm sets the flags for the nodes to be updated in the factor graph and proposes the new message updating rules for log-likelihood ratio and estimation bit.The nodes that satisfy the message updating rules in the same column can be updated in parallel,which reduces the delay waiting process of SC decoding algorithm and decreases the decoding latency.The simulation results show that the proposed algorithm has lower decoding latency without loss in BER performance compared to SC decoding algorithm,the proposed algorithm need less decoding time.