Symbol-Level Noise-Guessing Decoding with Antenna Sorting for URLLC Massive MIMO

Providing ultra-reliable and low-latency transmission is a current issue in wireless communications (URLLC). While it is commonly known that channel coding with large codewords improves reliability, this usually necessitates using interleavers, which incur undesired latency. Using short codewords is a necessary adjustment that will eliminate the requirement for interleaving and reduce decoding latency. This paper suggests a coding and decoding system that, combined with the high spectral efficiency of spatial multiplexing, can provide URLLC over a fading wireless channel. Random linear codes (RLCs) are used over a block-fading massive multiple input-multiple-output (mMIMO) channel followed by zero-forcing (ZF) detection and guessing random additive noise decoding (GRAND). A variation of GRAND, called symbol-level GRAND, originally proposed for single-antenna systems, is generalized to spatial multiplexing. Symbol-level GRAND is much more computationally effective than bit-level GRAND as it takes advantage of the structure of the constellation of the modulation. The paper analyses the performance of symbol-level GRAND depending on the orthogonality defect (OD) of the underlying lattice. Symbol-level GRAND takes advantage of the a priori probability of each error pattern given a received symbol, and specifies the order in which error patterns are tested. The paper further proposes to make use of further side-information that comes from the mMIMO channel-state information (CSI) and its impacts on the reliability of each antenna. This induces an antenna sorting order that further reduces the decoding complexity by over 80 percent when comparing with bit-level GRAND.