Prof. Mojtaba Vaezi (Villanova University, and Visiting Fellow at Princeton University, USA)
Mojtaba Vaezi [Senior Member, IEEE] is an Assistant Professor of ECE at Villanova University and a Visiting Fellow at Princeton University. He obtained his Ph.D. in EE from McGill University, Montreal, Canada, in 2014. From 2014 to 2015, he was a researcher at Ericsson Research, Canada, and from 2015 to 2018, he was with Princeton University as a postdoc. He has also worked as a Radio Network Design and Optimization Engineer and Manager at Ericsson for a few years. His research interests include the broad areas of signal processing and machine learning for wireless communications with an emphasis on beyond fifth-generation (5G) radio access technologies and deep learning-based communication. Among his publications in these areas is the book “Multiple Access Techniques for 5G Wireless Networks and Beyond” (Springer, 2019).
Dr. Vaezi is a recipient of several research, academic, and leadership awards, including d the 2021 IEEE Philadelphia Section Delaware Valley Engineer of the Year Award, the 2020 IEEE Communications Society Fred W. Ellersick Prize, IEEE Communications Letters’ best editor award in 2018, NSERC Postdoctoral Fellowship in 2014, and the 2013 IEEE Larry K. Wilson Regional Student Activities Award. He is/was an Editor of IEEE Transactions on Communications (2019-present), IEEE Communications Letters (2017-present), and IEEE Communications Magazine (2014-2018), a TPC member of the IEEE ICC’16-22, Globecom’16-22, PIMRC’16,20-2, and the lead organizer of the 1st to 6th NOMA workshops at VTC’17, GC’17, ICC’18, GC’18, ICC’19, and ICC’20.
After covering the fundamentals and potentials of NOMA, the tutorial will 1) go over some common misconceptions in this field and debunk them, 2) discuss why NOMA has not get into practice and what steps should be taken toward this, and 3) provide concrete examples as well as a pragmatic vision for NOAM research towards 6G networks.
Motivation and Context
The research in the field of none-otrhogonal multiple access (NOMA) has been very hot for several years now. This is partly because the number of devices is projected to continue an exponential growth over the next several years. NOMA is a promising technique to address this challenge. However, the current research in this field is mostly about combining NOMA with other techniques rather than refining the assumptions and innovating new solution that embrace the practical challenges.
Unfortunately, despite extensive research on NOMA during the past decade, this promising technique has not found a way around the standardization. The objective of the tutorial is to bring new insights to the analysis and design of NOMA for pushing NOMA to become a technology with practical impacts, not just a theoretical concept. After covering the fundamentals and potentials of NOMA, the tutorial will 1) go over some common misconceptions in this field and debunk them, 2) discuss why NOMA has not get into practice and what steps should be taken toward this, and 3) provide concrete examples as well as a pragmatic vision for NOAM research towards 6G networks.
Structure and Content
This will be a half-day tutorial. The schedule is given in the table below.
|20 minutes||Introduction (What Drives NOMA and Why?)|
|40 minutes||Theory behind NOMA and Common Myths|
|30 minutes||State of NOMA Implementation and Standardization|
|30 minutes||Networking Break|
|20 minutes||What Are the Pivotal Questions When Moving to Practice?|
|40 minutes||Open Issues and Future Research Directions|
Summary: Non-orthogonal multiple access (NOMA) is a promising multiple access candidate to accommodate massive number of devices and keep pace with the exponential of data traffic in next generation of cellular networks. NOMA exploits the path loss differences among the users to allow them to access a common medium. It also can increase spectral efficiency and reduce communication latency. Owing to such promises, NOMA has become immensely popular in academia since the name was coined in 2013. Although NOMA has gained significant attention from the communications society, the research in this field has been subject to several widespread myths and misunderstandings that either hinder fully utilizing its capacity or direct the research in inappropriate directions. This is among the main reasons why NOMA has not entered 5G standards despite creating a sensation in academia. The main motivation for this tutorial is to create a better understanding of NOMA’s potential and limitations by explaining its fundamental limits using plain language and debunking several of such common myths about NOMA. Besides pruning the twigs of the NOMA tree, this tutorial will identify promising research directions in this field by posing critical questions that are important for the effective adoption of NOMA in real-world networks. Toward this goal, the tutorial will also provide a comprehensive overview of the state-of-the-art efforts on experimental NOMA. Further, several big research questions that can move NOMA from merely being interesting academic research to an impactful technology will be discussed. Among them are NOMA without successive interference cancellation, novel decoding techniques for resource-limited Internet of things (IoT) devices, and distributed and end-to-end learning for NOMA.
Detailed Outline of the Tutorial.
1. Introduction (What Drives NOMA and Why?) (20 min)
- Overview of 6G cellular networks requirements
- Multiple access methods in 1G-4G – key features/shortcomings
- Non-orthogonal multiple access (NOMA) for beyond 5G
- NOMA in power domain/code domain
- What makes NOMA popular?
This part of the tutorial will review multiple access methods in 1G-4G as well as the requirements of 5G/6G networks to motivate the need for NOMA. It will describe the basic concept of NOMA in the power and code domains.
2. Theory behind NOMA and Common Myths (40 min)
- NOMA in power domain – basic concept
- Multi-input multi-output (MIMO) NOMA
- Multi-cell and massive NOMA
- NOMA and rate splitting multiple
- Interplay between NOMA and other technologies
- Common myths and misunderstanding
After motivating the need for NOMA for the evolving 5G cellular networks, the theoretical basics of NOMA will be discussed in this part of the tutorial. Specifically, it will describe the theory behind NOMA in the single- and multi-antenna systems as well as single- and multi-cell networks. This will serve as a prelude for the upcoming parts, as well as our discussion on NOMA misconceptions.
3. State of NOMA Implementation and Standardization (30 min)
- NOMA in 3GPP and other standardization
- Implementation of NOMA in software defined radio (SDR) – latest results
- Practical challenges of NOMA (SIC error propagation, imperfect CSI, …)
- NOMA without successive interference cancellation (SIC)
In this part of the tutorial state of NOMA in 3GPP standardization will be discussed first. Next, important results in SDR-based NOMA implementation will be elaborated on and their findings will be highlighted. These works will be a stepping-stone to list practical challenges and ask critical questions in the next part that could help shape more pragmatic research in the field of NOMA.
4. What Are the Pivotal Questions When Moving Towards Practice? (20 min)
- How to push NOMA beyond just a theoretical topic?
- Is NOMA a reasonable access method for resource-limited devices like IoT devices?
- What role can machine learning/deep learning play in moving NOMA to practice?
- Will NOMA be part of 6G standards?
- What are the hindering issues?
The goal of this part is to pose crucial questions to push NOMA beyond just a theoretical topic. As such, it is important to know that NOMA is not a solution for ‘everything.’ Reasonable expectations along with problems hindering NOMA to enter standards will be discussed.
5. Promising Future Research Directions (40 min)
- Finite-alphabet NOMA. Why is it promising and what are the open issues?
- SIC-free decoding, bite error rates versus complexity
- Autoencoder-based constellation design and decoding
- NOMA for IoT: distributed learning versus end-to-end learning
This part will first take a critical view on the relation between Shannon-theory and practical digital modulation schemes. Following that, several novel, practical designs/solutions will be discussed. This includes modulation design using a deep autoencoder, SIC-free decoding, and end-to-end communication to reduce bit error rate which is the ultimate goal of digital communication.