A Low Complexity Data Detection Algorithm for Uplink Multiuser Massive MIMO Systems

• Published in: IEEE journal on selected areas in communications Vol. 35; no. 8; pp. 1701 - 1714
• Main Author: Chen, Jung-Chieh
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.08.2017; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Algorithms; Bit error rate; Complexity; Computation; Computational complexity; Computer simulation; Coordinate descent method; data detection; Detection algorithms; Detectors; Error detection; Iterative methods; LMMSE; low complexity; massive MIMO; Mean square values; MIMO; MIMO (control systems); Sensors; Upgrading; Uplink
• ISSN: 0733-8716; 1558-0008
• DOI: 10.1109/JSAC.2017.2710878

A major challenge for uplink multiuser massive multiple-input and multiple-output (MIMO) systems is the data detection problem at the receiver due to the substantial increase in the dimensions of MIMO systems. The optimal maximum likelihood detector is impractical for such large wireless systems, because it suffers from exponential complexity in terms of the number of users. Therefore, suboptimal alternatives with reduced complexity, such as the linear minimum mean square error (LMMSE) detector, are necessary. However, the LMMSE detector still introduces high computational complexity, mainly caused by the computation of the Gram matrix and matrix inversion. To reduce the computational complexity of data detection while achieving satisfactory bit error rate (BER) performance, we initially proposed an iterative data detection algorithm that exploits the coordinate descent method (CDM)-based algorithmic framework for uplink multiuser massive MIMO systems. We then developed a reduced-complexity hardware implementation algorithm by leveraging the "one-at-a-time" update property of the CDM-based algorithmic framework. Simulation results revealed that the proposed CDM-based detector provides the same or improved BER performance than the classical LMMSE algorithm at a lower complexity for different test scenarios.