Document Type
Dissertation
Degree
Doctor of Philosophy (PhD)
Major/Program
Electrical and Computer Engineering
First Advisor's Name
Arjuna Madanayake
First Advisor's Committee Title
Committee Chair
Second Advisor's Name
Shubhendu Bhardwaj
Second Advisor's Committee Title
Committee Member
Third Advisor's Name
Gang Quan
Third Advisor's Committee Title
Committee Member
Fourth Advisor's Name
Jayantha Obeysekera
Fourth Advisor's Committee Title
Committee Member
Keywords
Full-Duplex, IBFD, Machine Learning, Machine Learning Accelerators, RFSoC
Date of Defense
3-23-2023
Abstract
The proliferation of wireless devices and the growing demand for higher data rates have imposed new challenges on modern wireless communication systems. An unavoidable consequence of these trends is a heavily congested spectrum. Therefore, improving spectrum efficiency has become critical to the sustainable growth of wireless communications systems. Multiple-input-multiple-output (MIMO) and orthogonal frequency division multiplexing (OFDM) are two widely adapted techniques to improve spectrum efficiency. Simultaneous transmit and receive (STAR) in the same frequency band, also known as in-band full-duplex (IBFD) or simply full-duplex (FD), is another promising technology to improve spectrum efficiency. FD technology has the potential to double the spectrum efficiency by reusing the same spectrum for both uplink and downlink. This can be especially beneficial in the 5G-FR1 (Sub 6 GHz) frequency band due to high spectrum congestion and associated high spectrum licensing cost. Therefore, research in developing FD-capable radios ready for modern wideband communication has significant technological and economic value.
This dissertation addresses the primary challenge in FD communication, known as self interference (SI). The main objective of this work is to investigate novel techniques to effectively remove self-interference in wideband and multiple-input multiple-output (MIMO) radios. Traditional two-way wireless communication devices use frequency division duplexing or time division duplexing to avoid the high-power transmit (Tx) signal damaging the sensitive receiver or shadowing the weak signal of interest (SOI) arriving from afar. Since FD radios perform STAR in the same frequency, the transmit signal leaks into the receiver, causing SI. The large power difference between the transmit signal and the signal of interest makes it challenging to cancel SI using a simple solution. The approach to solve this challenge is to deploy SI cancellation stages across multiple domains like spatial, RF, and digital. This thesis is focused on RF and digital cancellation stages. In particular, active ferrite-less circulators and passive circulator-free RF cancellation techniques are developed and tested. Also, traditional finite impulse response (FIR) and memory polynomial methods are studied for digital cancellation. A novel machine learning-based SI cancellation scheme that is robust to variations in transmit power is developed and experimentally verified. SI cancellation techniques designed for single antenna systems are ineffective for MIMO systems due to quadratic growth in complexity and linear increase in residual SI with respect to the number of transmit antennas. This challenge is addressed by developing RF and digital cancellers for MIMO radios that scale linearly in complexity with the antenna count.
The results of this dissertation demonstrate the feasibility of full-duplex technology for wideband MIMO communication by employing a multi-stage cancellation scheme that utilizes RF and digital cancellation. This work will inspire further investigation into FD radios that are suitable for modern wireless local and wide area networks. The findings of this thesis will serve as a stepping stone toward the commercial deployment of FD radio transceivers.
Identifier
FIDC011035
ORCID
0000-0001-5731-7293
Creative Commons License
This work is licensed under a Creative Commons Attribution-Share Alike 4.0 License.
Recommended Citation
De Silva, Yaddehi Udara, "Full-Duplex Radio Frequency Front Ends and Machine Learning Accelerators for Wideband Wireless Communication" (2023). FIU Electronic Theses and Dissertations. 5287.
https://digitalcommons.fiu.edu/etd/5287
Rights Statement
In Copyright. URI: http://rightsstatements.org/vocab/InC/1.0/
This Item is protected by copyright and/or related rights. You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from the rights-holder(s).