A comparison and simulation of LDPC decoders for Chinese digital television multimedia broadcasting(CDTMB) standard

Abstract

For the last 20 years intense research has been carried out on Low Density Parity Check(LDPC) decoders for supporting various transmission technologies. There are various proven LDPC decoder algorithms such as Log domain sum product algorithm(LDSPA), Min sum decoder(MSD), Single scan min sum decoder(SSMSD) and normalized single scan min sum decoder(NSSMSD) which are currently in use. Each of these algorithm have its own distinct pros and cons from implementation point of view that comes to support various transmission standards. In this paper four LDPC decoding algorithms are implemented to test and support CDTMB standard for its three code rates viz. 0.4, 0.6 and 0.8. The performance evaluation of all four decoder algorithms mentioned above are carried out with Additive White Gaussian Noise(AWGN) channel with BPSK modulation scheme. The software implementation of these algorithms show that min sum decoder algorithm reduces complexity when compared to LDSPA. Also SSMSD and NSSMSD contributes to conservation of memory to 23% and take at least 35% less computation time compared to LDSPA and MSD. LDSPA is preferred for low data rate applications for its good BER performance, but it has the drawback of increase in hardware complexity due to the presence of tanh and log functions. SSMSD and NSSMSD is preferable for high data rate applications as this showed low memory requirement and faster performance.