提出一种新的时不变LDPC卷积码构造算法。对Tanner等的构造方案进行改进,产生了给定码率下的LDPC卷积码多项式矩阵;根据卷积码特性,对该多项式矩阵进行修正,获得具有最大给定编码记忆和快速编码特性的改进的时不变LDPC卷积码。该算法具有的快速编码特性可以降低硬件实现时的编码复杂度。该算法获得的在给定码率下具有最大给定编码记忆的特性可提高译码性能。对码的特性参数和仿真结果的分析表明,文中构造的时不变LDPC卷积码是优异的。
A novel approach of constructing time-invariant LDPC convolutional codes is proposed. The matrix of an LDPC ( Low-Density Parity-Check) convolutional code with a given rate is generated according to the original matrix generated by Tanner etc. Then, the matrix with maximum given encoding memory and fast encoding property is obtained by modifying the above-mentioned matrix. The fast encoding property of the proposed matrix can reduce the encoding complexity. And the maximum given encoding memory of the proposed LDPC convolutional code can improve the decoding performance. The characteristic parameters and simulation results show that the proposed LDPC convolutional codes are excellent.