Efficient Low Complexity MIMO Detection Algorithm Design for Wireless Systems

Abstract

The demand for reliable higher data rates in digital communication is increasing newline day by day with the growing number of users. Improving spectral efficiency is one newline of the key solutions to meet the demand for higher data rates. The multiple-input newline multiple-output (MIMO) system allows the transmission of multiple data streams con newlinecurrently within the same frequency band because of its spatial multiplexing capabil newlineity and hence improves spectral efficiency without requiring additional transmission newline power. Recently introduced large-scale MIMO and massive MIMOsupports hundreds newline of antennas at the base station and is able to serve dozens of users in the same time newlinefrequency resource simultaneously and is considered as a key enabling technology for newline present fifth-generation and upcoming sixth-generation wireless communication stan newlinedards. However, designing MIMO detectors with low complexity and near-optimal newline detection performance is challenging. newline In this thesis, the problem of efficient detection is investigated for small-scale, newline large-scale, and massive MIMO systems in spatial multiplexing mode. First, a low newline switching fixed complexity sphere decoding scheme is proposed for small-scale and newline massive MIMO systems. In addition, a sub-optimal sorted QR decomposition and newline soft-bit information generation scheme is presented. newline Next, a neighborhood searching technique, namely low-complexity pair-wise newline layered tabu search (PLTS), is proposed for large-scale MIMO systems. The complex newlineity of PLTS is reduced by an efficient Gram matrix and matched filtered output update newline strategy. Precomputation techniques are introduced to avoid redundant operations. newline Thenumberofiterations for each reactive tabu search (RTS) operation at a certain layer newline of PLTS impacts detector complexity and bit error performance. Therefore two strate newlinegies are introduced to optimize the effective RTS iterations required for layered tabu newline