Logo image
Study on the Decoding Algorithms for LDPC Codes Constructed from Euclidean Geometry
Thesis

Study on the Decoding Algorithms for LDPC Codes Constructed from Euclidean Geometry

Fu-Chen Chien
Masters, 國立清華大學, 產業研發碩士積體電路設計專班
2006

Abstract

歐氏幾何 低密度奇偶檢查碼 快速信息傳衍 Euclidean Geometry Low Density Parity Check Codes Turbo-Decoding Message-Passing TDMP LDPC extrinsic information Scaled-Factor throughput
Turbo-Decoding Message-Passing (TDMP) algorithm has been proposed to improve the convergence rate of the conventional Message-Passing (MP) decoding algorithm for Low Density Parity Check (LDPC) codes. TDMP can achieve a better error performance and memory saving in implementation. However the throughput of the TDMP algorithm is limited, especially when the number of check equations is large. In this thesis, we investigate Concurrent Turbo-Decoding Message-Passing (CTDMP) algorithm which can be used in VLSI implementation to increase the decoding throughput. The CTDMP algorithm is implemented by concurrently decoding the super-codes and exchanging extrinsic information among these super-codes. In addition, TDMP is used in the decoding of these super-codes. Since CTDMP has no loss of convergence speed and is suitable for parallel implementation, CTDMP can be used to increase the decoding throughput. It is known that Scaled-Factor Message-Passing (SF-MP) algorithm can be used to increase the convergence speed of decoding. Hence, we also combine SF-MP and CTDMP algorithms to further increase the decoding throughput.

Metrics

1 Record Views

Details

Logo image