Reconfigurable implementation with reduced precision of massive MIMO systems
Date
2021-04-30
Authors
Tian, Mi
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In wired communications, where the data is transmitted over a wired medium, the
received signals are of high fidelity at any time. For this reason, wired communications
have innate resistance to performance degradation created by interference and noise.
In wireless communications with mobile users in urban areas, the propagation medium
varies over time, and the received signal is subject to fading and multipath propagation.
Multiple-Input Multiple-Output (MIMO) communications can mitigate these issues and
achieve a high-capacity communication channel with high probability. In Massive MIMO
(MMIMO) systems the number of receive antennas is much greater than the number of
transmit antennas. The net result is that the signals corresponding to each transmitter at
the receiver are more likely to be uncorrelated with each other, simplifying the removal of
the interference between the transmitters’ signals. Since many modern wireless systems
such as cellular communications systems and wireless local area networks are interference limited, this makes massive MIMO a very attractive option for communications systems.
Due to the large number of linear RF receivers and high-speed, high-resolution
analog-to-digital converters in MMIMO base stations, the cost of massive MIMO communications systems is large. In this dissertation ways to reduce the cost of MMIMO base stations are investigated and implementation techniques are proposed. In particular, one way to reduce the cost of the implementation is to build the base stations with inexpensive hardware, which requires the measured signals to be coarsely quantized. Digital MMIMO receivers with reduced word-length for massive multiple-input multiple-output(MIMO) systems are proposed. The assessment of the bit error rate and extra power needed to compensate for the information loss due to the coarse quantization is assessed. The implementation with reduced fixed-arithmetic precision of linear algebra operations including the eigenvalue decomposition, which is the most computationally demanding portion of the data detection and decoding process, is presented. The digital hardware is based on inexpensive FPGAs, showing that the proposed techniques are promising to reduce the cost of massive MIMO systems, which is a key impediment to their widespread deployment.
Description
Keywords
Reconfigurable Implementation, Massive MIMO, Coarse quantization, FPGA