Hesami, Sara
(2020)
Modeling and Linearization of
MIMO RF Transmitters.
PhD thesis, National University of Ireland Maynooth.
Abstract
Multiple-input multiple-output (MIMO) technology will continue to play a vital
role in next-generation wireless systems, e.g., the fifth-generation wireless networks
(5G). Large-scale antenna arrays (also called massive MIMO) seem to be the most
promising physical layer solution for meeting the ever-growing demand for high
spectral efficiency. Large-scale MIMO arrays are typically deployed with high
integration and using low-cost components. Hence, they are prone to different
hardware impairments such as crosstalk between the transmit antennas and power
amplifier (PA) nonlinearities, which distort the transmitted signal. To avert the
performance degradation due to these impairments, it is essential to have mechanisms
for predicting the output of the MIMO arrays. Such prediction mechanisms are
mandatory for performance evaluation and, more importantly, for the adoption of
proper compensation techniques such as digital predistortion (DPD) schemes. This
has stirred a considerable amount of interest among researchers to develop new
hardware and signal processing solutions to address the requirements of large-scale
MIMO systems.
In the context of MIMO systems, one particular problem is that the hardware
cost and complexity scale up with the increase of the size of the MIMO system.
As a result, the MIMO systems tend to be implemented on a chip and are very
compact. Reduction of the cost by reducing the bill of material is possible when
several components are eliminated. The reuse of already existing hardware is an
alternative solution. As a result, such systems are prone to excessive sources of
distortion, such as crosstalk. Accordingly, crosstalk in MIMO systems in its simplest
form can affect the DPD coefficient estimation scheme. In this thesis, the effect of
crosstalk on two main DPD estimation techniques, know as direct learning algorithm
(DLA) and indirect learning algorithm (ILA), is studied.
The PA behavioral modeling and DPD scheme face several challenges that seek
cost-efficient and flexible solutions too. These techniques require constant capture
of the PA output feedback signal, which ultimately requires the implementation
of a complete transmitter observation receiver (TOR) chain for the individual
transmit path. In this thesis, a technique to reuse the receiver path of the MIMO
TDD transceiver as a TOR is developed, which is based on over-the-air (OTA)
measurements. With these techniques, individual PA behavioral modeling and DPD
can be done by utilizing a few receivers of the MIMO TDD system. To use OTA
measurements, an on-site antenna calibration scheme is developed to individually
estimate the coupling between the transmitter and the receiver antennas.
Furthermore, a digital predistortion technique for compensating the nonlinearity
of several PAs in phased arrays is presented. The phased array can be a subset of
massive MIMO systems, and it uses several antennas to steer the transmitted signal
in a particular direction by appropriately assigning the magnitude and the phase
of the transmitted signal from each antenna. The particular structure of phased
arrays requires the linearization of several PAs with a single DPD. By increasing the
number of RF branches and consequently increasing the number of PAs in the phased
array, the linearization task becomes challenging. The DPD must be optimized to
results in the best overall linear performance of the phased array in the field. The
problem of optimized DPD for phased array has not been addressed appropriately in
the literature.
In this thesis, a DPD technique is developed based on an optimization problem
to address the linearization of PAs with high variations. The technique continuously
optimizes the DPD coefficients through several iterations considering the effect of
each PA simultaneously. Therefore, it results in the best optimized DPD performance
for several PAs.
Extensive analysis, simulations, and measurement evaluation is carried out as
a proof of concept. The different proposed techniques are compared with conventional approaches, and the results are presented. The techniques proposed in this
thesis enable cost-efficient and flexible signal processing approaches to facilitate the
development of future wireless communication systems.
| Item Type: |
Thesis
(PhD)
|
| Keywords: |
Modeling; Linearization; MIMO RF Transmitters; |
| Academic Unit: |
Faculty of Science and Engineering > Electronic Engineering |
| Item ID: |
16879 |
| Depositing User: |
IR eTheses
|
| Date Deposited: |
24 Jan 2023 10:40 |
| URI: |
|
| Use Licence: |
This item is available under a Creative Commons Attribution Non Commercial Share Alike Licence (CC BY-NC-SA). Details of this licence are available
here |
Repository Staff Only(login required)
 |
Item control page |
Downloads per month over past year
Origin of downloads
Altmetric
Altmetric