A Soft Demodulation Algorithm with Low Complexity for One-dimensional DPC System

Abstract

Based on the one-dimensional dirty paper coding (DPC) system, a low complexity algorithm is proposed to calculate log likelihood ratios (LLRs) for multi-level quadrature amplitude modulation (M ? QAM). The M ? QAM constellation is split into the in-phase and quadrature $${\sqrt{M}-PAM}$$ signal of mutual independence, and the LLR value of each bit is calculated on the expanded $${\sqrt{M}-PAM}$$ constellation. For both the complete interference pre-subtraction and partial interference pre-subtraction schemes, accurate LLR values can be obtained using the proposed simplified algorithm. The proposed algorithm is applied to the one-dimensional DPC system using convolutional code and low density parity check code respectively. Simulation results indicate that the bit error rate performance is almost not lost with the simplified algorithm while the computational complexity is reduced significantly.