Sum Rate Optimization of IRS-Aided Uplink Muliantenna NOMA with Practical Reflection

Area of Science:

• Wireless Communication
• Signal Processing
• Information Theory

Background:

• Intelligent Reflecting Surfaces (IRSs) are emerging as a key technology for 6G networks.
• IRSs offer a cost-effective method to improve network performance metrics.
• Nonorthogonal Multiple Access (NOMA) is crucial for achieving massive connectivity and higher spectral efficiency.

Purpose of the Study:

• To maximize the sum rate of an IRS-aided uplink NOMA network.
• To optimize the IRS reflection pattern considering unit modulus and practical reflection constraints.
• To develop efficient algorithms for different IRS sizes.

Main Methods:

• Formulated a determinant maximization (max-det) problem for moderate IRS sizes after rank relaxation.
• Employed Gaussian randomization to obtain a feasible solution from the max-det problem.
• Applied gradient-based iterative algorithms (BFGS, L-BFGS) for large IRS element numbers, deriving an efficient gradient form.

Main Results:

• The max-det approach yields near-optimal sum rates for unit modulus IRS reflection.
• Gradient-based algorithms demonstrate superior performance and complexity trade-offs for large IRS sizes with practical reflection.
• The study provides a comprehensive analysis of IRS-aided NOMA network optimization.

Conclusions:

• IRSs significantly enhance sum rates in uplink NOMA networks.
• The choice of optimization method (max-det vs. gradient-based) depends on IRS size and reflection type.
• The proposed methods offer efficient solutions for future 6G wireless systems.