论文部分内容阅读
时延是现代通信系统的一个重要指标。滑动窗译码能够在保证性能的基础上降低时延。基于滑动窗译码提出了一种不规则原模图低密度奇偶校验(low density parity check,LDPC)卷积码的构造方法。通过对基于AR4JA(accumulate-repeat-by-4-jagged-accumulate)的LDPC卷积码在AWGN(additive white Gaussian noise)信道下的P-EXIT性能分析发现,利用这种构造方式能够设计出多码率并且在BP译码和滑动窗译码方式下都能逼近容量限的LDPC卷积码。计算机仿真证明了基于AR4JA的LDPC卷积码性能优于规则的LDPC卷积码,而且在滑动窗译码方式下在降低至少56.7%时延的同时表现出了很好的性能。结合提出的不规则原模图LDPC卷积码的构造方法和滑动窗译码得到了一种能够实现译码时延与码字性能良好折中的码字构造有效方式。
Time delay is an important indicator of modern communication systems. Sliding window coding can reduce the delay based on guaranteed performance. Based on sliding window decoding, a new method for constructing low density parity check (LDPC) convolutional codes is proposed. By analyzing the P-EXIT performance of an AWGN (additive white Gaussian noise) channel based on the accumulative-repeat-by-4-jagged-accumulate (AR4JA) LDPC code, it can be found that the multi-code Rate and can approach capacity-limited LDPC convolutional codes both in BP decoding and sliding window decoding modes. Computer simulation proves that LDPC convolutional code based on AR4JA outperforms regular LDPC convolutional codes and shows good performance while reducing the delay of at least 56.7% in sliding window coding mode. Combined with the proposed construction method of LDPC convolutional codes and sliding window decoding, an effective method of codeword construction with good compromise between decoding delay and codeword performance is obtained.