Boosting Low-Complexity Vector Perturbation Variants Based on Rectangular Shaping

Shanxiang Lyu, Zheng Wang, Ling Liu, Hongliang He, Guanggang Geng

Abstract: We identify vector perturbation (VP) precoding as a special case of nested lattice coding where the fine/coding lattice is adopted from the channel matrices, and the coarse/shaping lattice is in the self-similar form of the fine lattice. Since we still have the freedom to optimize the shaping lattice, in this work, we design an efficient algorithm to construct a rectangular shaping lattice for VP. This design is especially well-suited to the low-complexity approximations of VP, i.e., zero forcing (ZF) and successive interference cancellation (SIC) precoders. Simulations show that rectangular shaping based ZF and SIC precoders exhibit significant performance gains compared with their self-similar shaping based counterparts.