Customized Joint Blind Frame Synchronization and Decoding Methods for Analog LDPC Decoder

Authors:
Publisher website
Google Scholar
Abstract
In this study, a joint blind frame synchronization and decoding method is proposed based on the normalized syndrome satisfaction probability (NSSP) provided by the soft Low-Density Parity-Check (LDPC) codes decoder. To reduce complexity, a stopping criterion is introduced into the iterative process through the convergence of NSSP evolution patterns. In addition to achieving an excellent synchronization performance, this method also eliminates the redundancy caused by the pilot sequence. Furthermore, it is compared with the optimal pilot-based and other code-aided frame synchronization algorithms. According to analytical and simulation results, the proposed technique outperforms other advanced code-aided frame synchronization solutions in synchronization, which makes it applicable to achieve comparable performance to the pilot-based methods in specific coding gains. Due to the introduction of a novel stopping criterion, the average number of joint detection & decoding is reduced by up to 90%. Analog probability processing technology integrates sub-threshold region circuits with probability domain-based iterative message-passing algorithms inspired by high energy efficiency and low complexity. Hence, hardware implementation is presented based on the analog LDPC decoder chip as fabricated in a 0.35-μm CMOS technology for our proposed algorithms. The experimental results demonstrate the effectiveness of the proposed method and the realization loss is within 0.2 dB compared with the theoretical circuit simulation results.
Funding Information
  • National Key Research and Development Program of China (2019YFB1803200)
  • National Natural Scientific Foundation of China (Grants 62301049, 62201054, 62001030, 62371047)
  • Australian Research Council (ARC) Discovery Project (DP220103596)
  • ARC Linkage Project (LP200301482)
  • Beijing Nova Program (Z211100002121161)
  • Beijing Natural Science Foundation (JQ23015)

This publication has 44 references indexed in Scilit: