Construction Methods and Performance Evaluation of Regular LDPC Convolutional Codes

What is it about?

In this paper, we present three construction methods of systematic regular LDPC-CCs based on randomly or algebraically structured sparse matrices. We provide an overview of LDPC-CCs and compares between Gallager’s, Mackay-Neal and Quasi-Cyclic (QC) construction methods, which are a good choice for many wireless applications due to their errors correction efficiency and low decoding complexity. At a rate-1/2, for the same size of the parity matrix to obtain the same encoding and decoding complexity, and using the Bit Flipping (BF) decoding algorithm over Additive White Gaussian Noise (AWGN) channel, we simulated the BER performance of a (3,6) regular LDPC convolutional codes, where the results showed that Gallager’s construction method (random) outperforms Mackay-Neal and QC ones.

Featured Image

Photo by Kirill Sh on Unsplash

Photo by Kirill Sh on Unsplash

Why is it important?

We conclude that increasing the size of the syndrome-former memory for LDPC convolutional code using Gallager’s construction, gives an important performance enhancement. Simulation results have also shown that variable and check node degrees are also considered as important parameters with decreased values, that an LDPC-CC can achieve higher performance, due to the matrix sparsity.

Perspectives