[Paper Review] Joint Beamforming Design and Power Splitting Optimization in IRS-Assisted SWIPT NOMA Networks

Zhendong Li, Wen Chen+3 more|arXiv (Cornell University)|Nov 30, 2020

Advanced Wireless Communication Technologies48 references4 citations

TL;DR

This paper proposes a joint optimization framework for IRS-assisted SWIPT-NOMA networks, jointly designing beamforming, power splitting ratios, and intelligent reflecting surface (IRS) phase shifts to minimize base station transmit power. By leveraging semidefinite relaxation and successive convex approximation, the algorithm reduces transmit power by 51.13% compared to non-IRS networks, significantly enhancing spectral and energy efficiency.

ABSTRACT

This paper proposes a novel network framework of intelligent reflecting surface (IRS)-assisted simultaneous wireless information and power transfer (SWIPT) non-orthogonal multiple access (NOMA) networks, where IRS is used to enhance the NOMA performance and the wireless power transfer (WPT) efficiency of SWIPT. We formulate a problem of minimizing base station (BS) transmit power by jointly optimizing successive interference cancellation (SIC) decoding order, BS transmit beamforming vector, power splitting (PS) ratio and IRS phase shift while taking into account the quality-of-service (QoS) requirement and energy harvested threshold of each user. The formulated problem is non-convex optimization problem, which is difficult to solve it directly. Hence, a two-stage algorithm is proposed to solve the above-mentioned problem by applying semidefinite relaxation (SDR), Gaussian randomization and successive convex approximation (SCA). Specifically, after determining SIC decoding order by designing IRS phase shift in the first stage, we alternately optimize BS transmit beamforming vector, PS ratio, and IRS phase shift to minimize the BS transmit power. Numerical results validate the effectiveness of our proposed optimization algorithm in reducing BS transmit power compared to other baseline algorithms. Meanwhile, compared with non-IRS-assisted network, the IRS-assisted SWIPT NOMA network can decrease BS transmit power by 51.13\%.

Motivation & Objective

• To address the challenge of low energy efficiency and poor coverage in massive IoT networks with energy-constrained devices.
• To enhance spectral efficiency and energy harvesting performance in NOMA systems by integrating intelligent reflecting surfaces (IRS).
• To jointly optimize beamforming, power splitting ratios, and IRS phase shifts under QoS and energy harvesting constraints.
• To minimize base station transmit power while satisfying user quality-of-service and minimum energy harvesting requirements.

Proposed method

• Formulates a non-convex optimization problem for minimizing BS transmit power under QoS and energy harvesting constraints.
• Applies a two-stage algorithm: first optimizing SIC decoding order via IRS phase shifts, then iteratively refining beamforming, power splitting, and phase shifts.
• Uses semidefinite relaxation (SDR) to handle rank-one constraints in beamforming optimization.
• Employs Gaussian randomization to recover feasible beamforming solutions from relaxed SDR problems.
• Applies successive convex approximation (SCA) to handle non-convex constraints involving power splitting ratios and SINR ratios.
• Transforms SINR and SIC constraints into convex approximations using first-order Taylor expansions around current estimates.

Experimental results

Research questions

• RQ1 How can IRS be leveraged to improve both spectral efficiency and energy harvesting in NOMA networks?
• RQ2 What is the optimal joint design of beamforming, power splitting, and IRS phase shifts to minimize BS transmit power?
• RQ3 How does the proposed algorithm compare to baseline schemes in terms of transmit power reduction?
• RQ4 To what extent can IRS reduce the required transmit power in SWIPT-NOMA systems?

Key findings

• The proposed two-stage algorithm based on SDR and SCA effectively solves the non-convex optimization problem with high convergence and feasibility.
• Numerical results show a 51.13% reduction in BS transmit power compared to non-IRS-assisted SWIPT-NOMA networks.
• The algorithm outperforms baseline schemes in terms of transmit power minimization and system energy efficiency.
• The integration of IRS significantly improves both information decoding and energy harvesting performance across all users.