Full Stack 5G Physical Layer Transceiver Design for NOMA in Mobile Heterogeneous Networks

Islam, Md Shantanu (2023) Full Stack 5G Physical Layer Transceiver Design for NOMA in Mobile Heterogeneous Networks. Doctoral thesis, Birmingham City University.

Preview	Text Md Shantanu Islam PhD Thesis published_Final version_Submitted Sep 2023_Final Award Dec 2023.pdf - Accepted Version Download (19MB)
Preview	Text (presentation) NOMA Fundamentals.pdf - Supplemental Material Download (7MB)

Abstract

The Fifth Generation (5G) and Beyond 5G (B5G) wireless networks are emerging with a variety of new capabilities, focusing on Massive Machine-Type Communications (mMTC), enabling new use cases and services. With this massive increment of mMTC along with increasing users, higher network capacity is a must for 5G and B5G. The integration of mMTC with traditional user traffic creates a heterogeneous network landscape. To address this challenge, future network designs must prioritize optimizing spectrum efficiency while meeting diverse service demands. Non-Orthogonal Multiple Access (NOMA) stands out as a promising technology for enhancing both system capacity and operational efficiency in such heterogeneous networks. Due to its non-orthogonal resource allocation, NOMA outperforms Orthogonal Multiple Access (OMA) in spectral efficiency, throughput, and user capacity, while also offering superior scalability and adaptability to network heterogeneity. Despite its promising advantages, large-scale implementation of NOMA in cellular systems remains elusive due to various challenges, making it a focal point of current research in cellular network technology.

While there has been considerable progress in implementing NOMA for broadcast and multicast services, notably with Layer Division Multiplexing (LDM) in next-generation digital TV, the challenges of unicast downlink transmission in NOMA remain largely unexplored. Unicast transmission requires a highly tailored network configuration adaptable to individual user requirements and dynamic channel conditions. Clustering users under a single NOMA channel must be both efficient and adaptive to ensure successful transmission, especially for mobile receiver. Besides, the interplay between NOMA and other 5G technologies remains insufficiently explored, in part due to the lack of an established NOMA-5G framework. Specifically, the collective impact of 5G physical layer technologies such as Low-Density Parity Check (LDPC) coding, Multiple-Input Multiple-Output (MIMO) Beamforming, and mmWave transmission on NOMA’s performance has not been comprehensively studied. Furthermore, in NOMA schemes involving more than two multiplexed users, known as Multilayer NOMA (N-NOMA), the system becomes increasingly complex and susceptible to noise. While N-NOMA holds considerable promise for scalability, its performance metrics are not yet fully characterized, due to challenges ranging from resource allocation complexities to transceiver design issues. Additionally, existing analytical models for performance evaluation are developed for orthogonal systems, are not fully applicable for assessing NOMA performance. Developing new models that incorporate the impact of non-orthogonality could provide more accurate performance assessments and offer valuable insights for future NOMA research.

Initially this thesis investigates the feasibility of LDM for unicast & multicast downlink transmission scenarios for Internet of Things (IoT)- user pairs. The findings indicate the Core Layer (CL) performance aligns with IoT requirements while Enhance Layer (EL) layer is suitable for users. A specialized Bit Error Rate (BER) expression is formulated to precisely predict CL performance, considering Lower Layer (LL) interference with predefined power ratio. Subsequently, the thesis employs a novel surface mobility model and adaptive power ratio allocation to evaluate LDM pair sustainability under various receiver mobility conditions.

Extending the LDM-Orthogonal Frequency Division Multiplexing (OFDM) model, this thesis presents a Third Generation Partnership Project (3GPP)-compliant 5G transceiver incorporating N-NOMA. This design incorporates a strategically-arranged set of NOMA functionalities and undergoes a rigorous performance evaluation. In particular, the transceiver provides a comprehensive assessment of N-NOMA performance, considering various transmission parameters such as LDPC code rate, MIMO order, modulation schemes, and channel specifications. These considerations not only provide new insights into non-orthogonal access technologies but also highlight dependencies on these factors for network configuration and optimization. To further advance this work, a one-shot N-NOMA multiplexing technique is developed and implemented, simplifying multi-layer standard sequential combiners to reduce transmission latency and transceiver complexity. A more accurate analytical BER expression is also formulated that considers the impact of both residual and non-residual Successive Interference Cancellation (SIC) errors across NOMA layers.

To build upon these advancements, an adaptive Power Allocation (PA) technique is introduced to optimize NOMA cluster sustainability and throughput. Employing a greedy algorithmic approach, this method uses real-time transmission feedback to dynamically allocate power across NOMA layers. In addition, a new Three Dimensional (3D) mobility model has been developed, consistent with existing 3GPP standards, capturing vehicular and pedestrian movement across urban and rural macro & micro-cell environments. When integrated with the PA technique, this model allows for real-time adjustments in the NOMA power ratio, effectively adapting to fluctuating receiver channel conditions.

Collectively, the findings from this research not only indicate significant physical layer performance improvements but also provide new insights into the potential of non-orthogonal access technologies. In the LDM-OFDM setup presented in Chapter 3, the EL layer needs 15 dB more Signal-to-Noise Ratio (SNR) than the CL to achieve the same BER, but allows for higher data rates. When it comes to mobility, IoT movement accounts for about 70% of link terminations in scenarios with similar mobility patterns. The N-NOMA-5G shows significant improvement in low SNR performance compared to existing literature. The 3 layer simulations shows on average a 60% reduction in the SNR requirements to achieve similar BER. The implementation of a one-shot multiplexer has demonstrated a substantial reduction in N-NOMA multiplexing time, particularly with the growing number of NOMA layers, as detailed in Chapter 4. Notably, the simulation outcomes spanning 2 to 10 layers of NOMA multiplexing indicate an remarkable 52% reduction in processing time. This underscores the effectiveness of the one-shot multiplexer in enhancing efficiency, particularly as the complexity of the NOMA setup intensifies. The developed analytical model also shows over 95% similarities with the simulation results. The impact of dynamic PA for both static and mobile receivers demonstrates on average, over 40% improvements in link sustainability time for mobile users and for static users, it achieves optimal PA and fast convergence within just 12 iterations, as detailed in Chapter 5.

Item Type:	Thesis (Doctoral)
Dates:	Date Event 4 September 2023 Submitted 19 December 2023 Accepted
Uncontrolled Keywords:	NOMA, 5G, B5G, BER, SIC, LDM, PA, Link Sustainability, User Mobility.
Subjects:	CAH11 - computing > CAH11-01 - computing > CAH11-01-01 - computer science
Divisions:	Doctoral Research College > Doctoral Theses Collection Faculty of Computing, Engineering and the Built Environment > School of Computing and Digital Technology
Depositing User:	Jaycie Carter
Date Deposited:	16 Feb 2024 16:17
Last Modified:	19 Feb 2024 10:54
URI:	https://www.open-access.bcu.ac.uk/id/eprint/15243

Actions (login required)

View Item