Source-linked AI summary
Rate-Splitting Multiple Access for RIS-Enabled Full-Duplex Communications with User Relaying
Yuwen Cao, Hui Wang, Tomoaki Ohtsuki, Shahid Mumtaz
TL;DR
Prior RIS-enabled RSMA work mostly considered downlink design, leaving joint full-duplex cooperative uplink and downlink optimization as an open problem. This paper combines RIS-assisted user relaying with joint transmit and reflective beamforming optimization. Its simulations report gains over compared schemes, including up to 31.6% over HD-RSMA and 74.4% over no-RIS C-RSMA at the stated setting.
Problem
Most prior RIS-enabled RSMA studies considered downlink design, while joint uplink and downlink full-duplex cooperative RSMA remained to be explored.
Source: §I · p. 1Method
The paper proposes RIS-assisted cooperative RSMA with user relaying and a low-complexity alternative optimization of BS transmit and RIS reflective beamformers.
Source: §I · p. 1Results
With P_r=10 dBm and P_t=30 dBm, C-RSMA gains up to 31.6% over HD-RSMA and 74.4% over no-RIS C-RSMA.
Source: §V · p. 6Takeaways & Limitations
Concurrent reflection of BS and relay streams enables one RIS beamforming design, reducing complexity and overhead relative to OMA’s separate-slot designs.
Source: §I · p. 1
Abstract
from arXiv · showhide
Rate splitting multiple access (RSMA) has emerged as a promising non-orthogonal transmission mechanism to improve spectral- and energy efficiency, and user fairness in multi-antenna broadband channels. To further enhance its performance, in this paper we investigate a novel cooperative RSMA (C-RSMA) transmission framework assisted by reconfigurable intelligent surface (RIS) and user relaying. More specifically, we first formulate the common- and private-rate maximization problem satisfying the transmit power constraint, by jointly optimizing the transmit beamformers at the BS and the reflective beamformer at the RIS. The above formulated problem is an intractable and NP-hard problem as the optimization variables are coupled and the strong residual self-interference (SI) is involved. To conquer the challenge mentioned above, we propose a low-complexity alternative optimization approach to jointly optimize the transmit and reflective beamforming. Simulation results show that our approach is capable of enabling superior sum-rate performance than the RIS-assisted space division multiple access (SDMA), half-duplex RSMA, and SDMA without using RIS schemes.
I. INTRODUCTION
The paper addresses the predominance of downlink-only RIS-enabled RSMA studies by proposing a full-duplex cooperative RSMA framework for joint uplink and downlink transmission. It jointly optimizes BS transmit beamformers and the RIS reflective beamformer under a transmit-power constraint.
I. INTRODUCTION
- RSMA splits each user’s stream into common and private parts, combining common parts into a shared-codebook stream and encoding private parts for their associated users.The introduction describes RSMA as a multi-user MIMO transmission and interference-management strategy.Source: §I · p. 1
- Prior RIS-enabled RSMA studies mostly considered downlink design, motivating joint optimization for full-duplex cooperative uplink and downlink transmission.The stated aim is to enhance data rate with low complexity.Source: §I · p. 1
- The paper formulates common- and private-rate maximization under a transmit-power constraint and proposes alternating optimization of BS transmit and RIS reflective beamformers.The authors characterize the problem as intractable and NP-hard because the variables are coupled and residual self-interference is strong.Source: §I · p. 1
- The proposed framework lets the RIS simultaneously reflect common and private streams from the BS and relaying user, enabling one RIS beamforming design instead of OMA’s per-slot designs.The authors state that this reduces system complexity and overhead.Source: §I · p. 1
- At 30 dBm BS transmit power, C-RSMA yields up to 31.6% common- and private-rate performance gain over state-of-the-art HD-RSMA and, at a given information transmission rate, consumes less transmit power than RIS-assisted SDMA and HD-RSMA.Source: §I · p. 2
II. SYSTEM MODEL
The system model uses a two-user MU-MISO link assisted by a single-connected RIS, with user 1 acting as a full-duplex relay for user 2. BS and relay transmissions share the RIS, while user 1’s decoding is affected by residual self-interference.
II. SYSTEM MODEL
- The model has one BS with M_t transmit antennas, two single-antenna users, and an RIS with N passive elements; user 1 is assumed to have the better channel and be closer to the BS.The model is in TDD mode, and the multi-user, multi-antenna extension is left for future work.Source: §II · p. 2
- User 1 acts as a full-duplex relay, assisting the BS in sending the common stream to user 2 while both users are served with a single-connected RIS.User 1 also transmits to user 2 through direct and RIS-reflected D2D links.Source: §II · p. 2
- The BS uses one-layer RSMA with common and private streams, beamformed under a maximum transmit-power constraint.The model includes residual self-interference in user 1’s received signal during full-duplex operation.Source: §II · p. 2
- User 1 decodes the common stream while treating private streams as interference, then uses successive interference cancellation before decoding its private stream.The corresponding common- and private-stream rates are expressed as log2(1 + SINR).Source: §II · p. 2
- User 2 receives BS and user 1 signals through both direct and RIS-reflected paths, and the common message rate must be decodable by both users.The model defines the user-to-user, RIS-to-user, and user-to-RIS channels and their additive Gaussian noise.Source: §II · pp. 2–3
III. PROBLEM FORMULATION
The section formulates sum-rate maximization for RIS-enabled C-RSMA by jointly optimizing the transmit and reflective beamformers. The problem is NP-hard because the variables are coupled, motivating a low-complexity alternative optimization approach.
III. PROBLEM FORMULATION
- In the proposed C-RSMA transmission framework assisted by a single-connected RIS, user 1 serves as a full-duplex relay, and the authors express the framework’s sum rate.Source: §III · p. 3
- The objective is to maximize the framework's sum rate by jointly optimizing the RIS reflective beamformer Θ and C-RSMA transmit beamformer W.The formulation includes a transmit-power constraint and unit-modulus RIS phase variables.Source: §III · p. 3
- Coupled transmit and reflective beamformers make the optimization NP-hard, so the paper proposes low-complexity alternative optimization to approximate a solution.The variables are coupled for each time slot.Source: §III · p. 3
A. Optimizing Transmit Beamforming of C-RSMA
With the RIS beamformer fixed, the transmit-beamforming step optimizes the private-stream rates using successive convex approximations and semidefinite relaxation. The resulting relaxed problem is convex and can be handled by existing solvers.
A. Optimizing Transmit Beamforming of C-RSMA
- With the reflective beamformer fixed, the transmit-beamforming subproblem optimizes the private rates of users 1 and 2.The rate expressions are written using the cascade channel between each user and the BS.Source: A. Optimizing Transmit Beamforming of… · p. 3
- Semidefinite relaxation and successive convex approximation address the non-convex private-rate objective and constraint in the transmit-beamforming subproblem.The method uses first-order Taylor expansion to form concave lower bounds for the rate terms.Source: A. Optimizing Transmit Beamforming of… · p. 3
- The resulting relaxed transmit-beamforming problem is convex and can be solved with existing convex optimization tools such as CVX.The constraint approximations use the arithmetic-geometric mean inequality.Source: A. Optimizing Transmit Beamforming of… · p. 4
B. Optimizing Reflective Beamforming of RIS
With transmit beamformers fixed, the RIS step reformulates the sum-rate optimization over the reflective beamformer. Semidefinite relaxation and successive convex approximation yield a convex relaxed problem.
B. Optimizing Reflective Beamforming of RIS
- Given the C-RSMA transmit beamforming vectors, the RIS step reformulates the sum-rate optimization with respect to reflective beamforming Θ.The formulation is stated for a signal-connected RIS.Source: B. Optimizing Reflective Beamforming… · p. 4
- The RIS subproblem is NP-hard because its objective and inequality constraint are non-convex.Source: B. Optimizing Reflective Beamforming… · p. 4
- The method uses semidefinite relaxation and successive convex approximation to transform the RIS subproblem into a convex relaxed problem solvable with existing tools such as CVX.The phase variables are represented as vn = e^jθn, and non-convex constraints are approximated using the arithmetic-geometric mean inequality.Source: B. Optimizing Reflective Beamforming… · pp. 4–5
C. Convergence and Complexity Analysis
The alternative optimization algorithm alternates between transmit-beamforming and RIS updates until the common-stream rate gain between adjacent iterations reaches a minimum. Its complexity separates into transmit and RIS subproblems.
C. Convergence and Complexity Analysis
- The algorithm alternates between updating transmit beamforming with the RIS fixed and updating the RIS with transmit beamforming fixed.The process repeats until the common-stream rate gain between adjacent iterations reaches a minimum.Source: C. Convergence and Complexity Analysis · p. 5
- Vectorization reduces RIS reflection optimization complexity from O(N^3) to O(N^2).Source: C. Convergence and Complexity Analysis · p. 5
- Transmit-beamforming optimization uses hierarchical Taylor expansion to form quadratic programming with linear constraints, while the separated subproblems have complexities O(M_t^3) and O(N^2).The global method instead jointly optimizes N + M_t variables.Source: C. Convergence and Complexity Analysis · p. 5
V. SIMULATION RESULTS
Simulations compare RIS-assisted C-RSMA with alternative schemes and antenna and relay-power settings. C-RSMA gains up to 31.6% over HD-RSMA and 74.4% over no-RIS C-RSMA at a BS transmit power of 30 dBm.
V. SIMULATION RESULTS
- The simulations place the BS at (0, 0) m, the RIS at (30, 0) m, and the users at (25, -7) m and (35, -7) m.The system is evaluated at mmWave frequencies.Source: §V · p. 5
- Increasing the number of RIS reflecting elements improves C-RSMA achievable rate and lowers BS transmit power for a given rate.The simulations use the Saleh–Valenzuela channel model and compare C-RSMA against SDMA, HD-RSMA, and no-RIS SDMA.Source: §V · p. 5
- With Mt = 16, C-RSMA achieves higher performance than with Mt = 8, 4, or 2 antennas; increasing D2D transmit power degrades performance, particularly at high SNR.Fig. 3 uses N = 32 and compares Pr = 0 or 10 dBm.Source: §V · p. 6, Figure 3 · p. 6
- At Pt = 30 dBm, C-RSMA with Pr = 10 dBm gains up to 31.6% over HD-RSMA and 74.4% over no RIS C-RSMA.C-RSMA also provides superior rate performance to HD-RSMA, no RIS C-RSMA, and no RIS SDMA.Source: §V · p. 6
VI. CONCLUSION
The conclusion presents a RIS- and user-relaying-assisted C-RSMA framework and a low-complexity algorithm for jointly optimizing BS transmit and RIS reflective beamforming.
VI. CONCLUSION
- The proposed low-complexity algorithm jointly optimizes BS transmit and RIS reflective beamforming for a transmit-power-constrained common- and private-rate maximization problem.Simulations show better sum-rate performance than the selected counterpart methods.Source: §VI · p. 6