A LDPC Decoding Algorithm based on Convolutional Neural Network

Abstract

At present, low density parity check code (LDPC) has been widely used in channel coding and decoding because of its excellent performance, but with the increase of code length, the complexity of decoding algorithm has became higher and higher. In view of the limitations of decoding algorithm and the rapid development of artificial intelligence technology, it has great research prospects to solve the above problems through deep neural network. Therefore, this paper mainly focuses on the design and improvement of LDPC decoding process, and proposes an LDPC decoding model based on DenseNet neural network structure, which improves the LPDC decoding performance by optimizing DenseNet neural network structure. This method can recover information at the decoding end, avoiding the limitations of traditional short decoding loop and high complexity of decoding algorithm. The simulation results show that the LDPC decoding algorithm based on DenseNet neural network structure improves the decoding performance.