Multiuser Interference Suppression: A Generalized Nearest Neighbor Decoding Approach
Shuqin Pang, Wenyi Zhang · 2024
The architecture of the proposed generalized nearest neighbor decoding (GNND) is investigated under quadrature phase shift keying (QPSK) and utilized for multiuser interference suppression. The GNND is capable of handling the non-Gaussian inter-stream interference effectively, for multiuser channels under QPSK input, while the conventional channel linearization (CL) approach generally fails. Adopting the information-theoretic tool of generalized mutual information (GMI) for measuring the achievable information rate, the optimal GNND in maximizing GMI is derived. A coded modulation scheme based upon GNND is further designed, via connected with the existing low-density parity-check decoders, without requiring iterations between demodulation and decoding. Numerical results show that the GNND significantly outperforms the CL approach in terms of the achievable code rate and the bit-error-rate.