Title :
A Comparative Performance and Complexity Study of Short-Length LDPC and Turbo Product Codes
Author :
Landolsi, Mohamed Adnan
Author_Institution :
Dept. of EE, King Fahd Univ. of Pet. & Minerals, Dhahran
Abstract :
This paper presents a comparative performance and complexity study between low-density parity-check (LDPC) codes and turbo product codes (TPC) of short block length (within 2048 bits). The LDPC codes are of the semi-random (SR) type, characterized by low encoder complexity, and are further optimized by eliminating short cycles of length 4 (minimum girth 6). The TPC codes are obtained from 2D and 3D constructions chosen to match the LDPC codes´ parameters. The numerical results show that the SR-LDPC codes have slightly better error performance (to within 0.5dB at a BER of 10-5 while demonstrating lower computational complexity per decoder iteration but the required number of decoding iterations is larger. However, this disadvantage is significantly reduced for moderately high signal-to-noise ratios (starting from 2.5dB). It is therefore concluded that SR-LDPC codes have a more competitive performance-complexity advantage overall
Keywords :
computational complexity; parity check codes; product codes; turbo codes; SR-LDPC codes; computational complexity; decoder complexity; decoder iteration; low encoder complexity; low-density parity-check codes; short block length; short-length LDPC codes; signal-to-noise ratios; turbo product codes; Bipartite graph; Bit error rate; Block codes; Iterative algorithms; Iterative decoding; Maximum likelihood decoding; Parity check codes; Product codes; Sparse matrices; Turbo codes; Bit Error Performance; Decoder Complexity; LDPC Codes; Turbo Product Codes;
Conference_Titel :
Information and Communication Technologies, 2006. ICTTA '06. 2nd
Conference_Location :
Damascus
Print_ISBN :
0-7803-9521-2
DOI :
10.1109/ICTTA.2006.1684775