Performance of Modified Lagrangian Dual Relaxation Detector in Massive MIMO System

In communication systems the maximum-likelihood (ML) decoding algorithm has a large-scale computation complexity beside with Lagrangian dual relaxation (LDR) detector and lattice reduction-aided (LRA) optimization under the PAM constellation set for massive multi-input multi-output (Massive MIMO) system based on extended up to 256 antenna number. The symbol bound constraints under the optimal ML detector may provide worse effects on the system performance, but focused on applying the lattice decoding (LD) with zero-forcing (ZF) and minimum mean square error (MMSE) regularization, we mitigate the out-of-bound symbol effects which are important part in telecommunications. As a result, the simulations obtain the benefit from 2 to 5 dB in proposed algorithms approach performance compared with conventional ML, LD and other schemes.