LDPC Coded Random-Polar CP-EBPSK System

Random polar continuous phase extended binary phase shift keying (CP-EBPSK) is a modulation technique of high spectrum efficiency. The special impulse response filter makes it possible that using threshold detection to demodulate. To improve the BER performance of the CP-EBPSK demodulator under low SNR, the LDPC code is introduced. Accurate posterior probability estimation is critical for decoding performance, but the special impacting filter used in CP-EBPSK demodulator brings more difficulty in LDPC decoding after obtaining posterior probability information. Bayesian, support vector machine (SVM) and two decision based posterior probability estimates method are introduced and with detailed analysis in this paper. Simulation results show that, the posterior probability estimation is more accurate obtained by SVM and Bayesian methods than by the two others and the BER performance can be improved 8dB-9dB. Because of the training stage, the complexity of these two methods are relatively high. Decision based posterior probability estimation method which without training is relatively simple and the performance can be improved more than 7dB. Thus, in the random-polar CP-EBPSK system using a posterior probability estimation as the initialization of LDPC decoder is an effective way.