PaperlayerView exact revision on arXiv

Source-linked AI summary

Rate-Splitting Multiple Access: Fundamentals, Survey, and Future Research Trends

Yijie Mao, Onur Dizdar, Bruno Clerckx, Robert Schober, Petar Popovski, H. Vincent Poor

arXiv:2201.03192v3cs.ITeess.SP

TL;DR

Existing RSMA research is fragmented across narrow scenarios without a comprehensive, pedagogical review or consensus. This paper surveys RSMA’s architecture, taxonomy, applications, comparisons, and future challenges, finding explicit sum- and MMF-DoF gains over other schemes across tested user counts.

  • Problem

    RSMA research lacks a comprehensive pedagogical review spanning scenarios, benefits, and usage conditions, while the rapidly expanding literature lacks consensus.

    Source: D. Contributions · p. 6

  • Method

    The paper synthesizes state-of-the-art RSMA research, explains its architecture and taxonomy, compares multiple-access schemes, and discusses future challenges.

    Source: D. Contributions · p. 6, 3) 1-layer RS · p. 10

  • Results

    1-layer RS achieves explicit sum- and MMF-DoF gains over all compared schemes across tested user counts under imperfect CSIT.

    Source: 1) DoF region · p. 32

  • Takeaways & Limitations

    RSMA is presented as a promising 6G paradigm for interference management, non-orthogonal transmission, and multiple access.

    Source: §VIII · p. 48

    System-level RSMA evaluations incorporating higher-layer effects such as HARQ, scheduling, and application-layer QoS remain unavailable.

    Source: C. Standardization and Implementation… · p. 48

Abstract

from arXiv · show

Rate-splitting multiple access (RSMA) has emerged as a novel, general, and powerful framework for the design and optimization of non-orthogonal transmission, multiple access (MA), and interference management strategies for future wireless networks. Through information and communication theoretic analysis, RSMA has been shown to be optimal (from a Degrees-of-Freedom region perspective) in several transmission scenarios. Compared to the conventional MA strategies used in 5G, RSMA enables spectral efficiency (SE), energy efficiency (EE), coverage, user fairness, reliability, and quality of service (QoS) enhancements for a wide range of network loads (including both underloaded and overloaded regimes) and user channel conditions. Furthermore, it enjoys a higher robustness against imperfect channel state information at the transmitter (CSIT) and entails lower feedback overhead and complexity. Despite its great potential to fundamentally change the physical (PHY) layer and media access control (MAC) layer of wireless communication networks, RSMA is still confronted with many challenges on the road towards standardization. In this paper, we present the first comprehensive overview on RSMA by providing a survey of the pertinent state-of-the-art research, detailing its architecture, taxonomy, and various appealing applications, as well as comparing with existing MA schemes in terms of their overall frameworks, performance, and complexities. An in-depth discussion of future RSMA research challenges is also provided to inspire future research on RSMA-aided wireless communication for beyond 5G systems.

I. INTRODUCTION … 3) Non-Orthogonal Multiple Access:

The introduction motivates RSMA for 6G by contrasting existing multiple-access schemes’ interference-management limits with RSMA’s adaptive partial decoding and noise treatment. It presents RSMA as a flexible paradigm that bridges SDMA and NOMA across interference levels and network conditions.

I. INTRODUCTION

  • 6G demands more efficient wireless-resource use and stronger interference management to support high throughput, ultra reliability, heterogeneous QoS, massive connectivity, and converged services.Source: §I · p. 1

A. Rate-Splitting Multiple Access

  • RSMA splits each user message into common and private parts, enabling partial interference decoding and partial noise treatment instead of SDMA’s or NOMA’s extremes.By tuning common and private streams, RSMA reduces to SDMA under weak interference or NOMA under strong interference, thereby bridging both schemes.Source: A. Rate-Splitting Multiple Access · pp. 2–3

1) Orthogonal multiple access:

  • OMA uses orthogonal time, frequency, code, or combined resources, simplifying transceivers and avoiding multi-user interference.Its single-user resource allocation restricts simultaneous users, limits spectral efficiency, wastes resources for low-rate users, and requires high-overhead scheduling.Source: 1) Orthogonal multiple access · p. 3

2) Space division multiple access:

  • SDMA uses linear precoding and treats residual interference as noise, achieving maximum underloaded broadcast-channel DoFs with perfect CSIT and low receiver complexity.Its performance degrades under overload, unfavorable channel deployment, scheduler constraints, and imperfect CSIT, where residual interference remains after inaccurate precoding.Source: 2) Space division multiple access · p. 4

3) Non-Orthogonal Multiple Access:

  • NOMA combines SDMA across user groups with SC–SIC within groups, improving spectral efficiency in severely overloaded settings with aligned channels and diverse strengths.However, increasing SIC layers aggravates receiver complexity and error propagation, while grouping and scheduling can reduce performance and increase latency.Source: 3) Non-Orthogonal Multiple Access · pp. 4–5
  • For a perfect-CSIT MISO broadcast channel, one-group multi-antenna NOMA achieves sum-DoF 1, whereas SDMA achieves min{M, K}, exposing NOMA’s DoF loss and receiver-complexity tradeoff [11, Table II].Source: 3) Non-Orthogonal Multiple Access · p. 5

B. PHY Layer Challenges for the Multiple Access Schemes

  • Existing schemes are constrained by interference-management choices: OMA avoids interference, SDMA treats it as noise, and NOMA fully decodes and removes it.SDMA is suited to weak interference and NOMA to strong interference, while medium interference favors adaptively partial decoding.Source: 3) Non-Orthogonal Multiple Access · p. 5
  • Dynamically switching between SDMA and NOMA does not adequately handle medium interference because each system still supports only one interference-management mode.This limitation motivates a unified approach that can adapt the decoded and noise-treated interference portions within one transmission framework.Source: A. Rate-Splitting Multiple Access · p. 3, 3) Non-Orthogonal Multiple Access · p. 5

C. 5G/6G and the Need for RSMA … A. Definition and Design Principle

The paper motivates RSMA as a flexible interference-management paradigm for increasingly demanding 5G/6G services and imperfect CSIT, then defines its principle of partially decoding and partially treating interference as noise. It presents a holistic tutorial covering RSMA architectures, theory, comparisons, applications, standardization, and research challenges.

C. 5G/6G and the Need for RSMA

  • The motivation is intensified by 5G/6G requirements spanning eMBB Gbps rates, URLLC 10−5 BLER within 1 ms, and dense, energy-efficient mMTC connections.Interference is expected to become more severe and unpredictable in 6G, making inappropriate multi-user interference management a major performance threat.Source: C. 5G/6G and the Need for RSMA · p. 6
  • RSMA answers the need for an MA scheme that is robust to imperfect CSIT, adapts interference processing to its level, and encompasses existing MA schemes.These requirements arise because interference threatens tightened 5G/6G KPIs, while imperfect CSIT makes interference difficult to eliminate and can cause substantial throughput loss.Source: C. 5G/6G and the Need for RSMA · p. 6

D. Contributions

  • The treatise provides the first holistic tutorial on RSMA, consolidating rapidly expanding literature through design principles, state-of-the-art review, merits, research challenges, and future directions.Its contributions include downlink, uplink, and multi-cell frameworks; information-theoretic and communication perspectives; and quantitative complexity comparisons.Source: D. Contributions · p. 6
  • It compares RSMA with SDMA, NOMA, OMA, and multicasting in achieved DoFs, complexity, and throughput, while surveying applications, 3GPP pathways, standardization, and implementation issues.Emerging application areas include SAGIN, V2X communications, and 3D eMBB-URLLC-mMTC services.Source: D. Contributions · pp. 6–7

E. Organization and Notation

  • The paper is organized around RSMA principles and architectures, information-theoretic literature, resource allocation, PHY-layer design, comparisons with other MA schemes, applications, challenges, and future trends.Notation defines matrix, vector, operator, distribution, and trace conventions, with the main notation summarized in Table III.Source: E. Organization and Notation · p. 7

II. PRINCIPLES OF RSMA

  • RSMA is developed as a PHY-layer strategy whose transmission frameworks cover downlink, uplink, and multi-cell networks.The section starts from a rigorous definition and design principle before detailing these transmission frameworks.Source: §II · p. 7, A. Definition and Design Principle · p. 8

A. Definition and Design Principle

  • Existing MA schemes manage interference by orthogonalizing resources, treating interference as noise, or decoding interference, whereas future schemes should adapt between these extremes.When interference is medium, the desired behavior is to partially decode and partially treat it as noise rather than choose an extreme.Source: A. Definition and Design Principle · p. 7
  • RSMA implements this adaptive principle by partially decoding interference and partially treating it as noise, thereby softly bridging full decoding and full noise treatment.This design constitutes a new paradigm for multiple access, non-orthogonal transmission, and interference management in future wireless networks.Source: A. Definition and Design Principle · p. 8, Figure 3 · p. 8

B. Downlink RSMA · 1) Transmitter design: · 2) Receiver design:

Downlink RSMA provides a universal transmitter–receiver framework for multi-user broadcast channels, encompassing diverse message-splitting, combining, encoding, precoding, and detection strategies. Its schemes differ in how messages are split and combined, which streams users decode, and whether linear or nonlinear precoding is used.

B. Downlink RSMA

  • Downlink RSMA schemes include 1-layer RS, hierarchical RS, generalized RS, RS-CMD, THPRS, and DPCRS, which are compared through their transmitter and receiver designs.Source: B. Downlink RSMA · pp. 8–9

1) Transmitter design:

  • The universal framework serves K > 1 single-antenna users from an M-antenna transmitter, covering SISO broadcast channels when M = 1 and MISO broadcast channels when M > 1.Source: 1) Transmitter design · p. 9
  • RSMA transmitters split each user message into sub-messages, with L = 2 for several schemes, L = 3 for 2-layer HRS, and L = 2K−1 for generalized RS and M-DPCRS.The larger split in generalized RS and M-DPCRS supports encoding common streams into different layers for more flexible interference management.Source: 1) Transmitter design · p. 9
  • Message combiners merge sub-messages into N messages for several schemes, whereas RS-CMD disables combining and directly encodes all LK sub-messages into N = LK streams.For 1-layer RS, THPRS, and 1-DPCRS, N = K + 1; for 2-layer HRS, N = K + 2; and for generalized RS and M-DPCRS, N = 2K − 1.Source: 1) Transmitter design · p. 9
  • Common messages contain sub-messages from one or multiple users and are decoded by multiple users, while private messages contain one user’s sub-message and are decoded by one user.RSMA common messages differ from multicast messages because users decode them for interference management without necessarily needing their entire content.Source: 1) Transmitter design · p. 9
  • Common and private messages use shared and independent codebooks, respectively, and their N streams are mapped through precoders onto M antennas under a total transmit-power constraint.Source: 1) Transmitter design · p. 9
  • Linear RSMA includes 1-layer RS, 2-layer HRS, generalized RS, and RS-CMD, while nonlinear RSMA includes THPRS with THP and DPCRS with DPC precoding.The linear schemes can be extended to THP, DPC, or other nonlinear precoding methods by changing the precoder.Source: 1) Transmitter design · p. 9

2) Receiver design:

  • Receivers decode only selected common streams, treat remaining common streams as noise, extract intended sub-messages, and recombine them to recover each original user message.Decoded common messages need not contain the intended sub-messages of the user decoding them.Source: 2) Receiver design · p. 10

3) 1-layer RS: · 4) 2-layer HRS:

The paper defines 1-layer RS by splitting each user message into common and private parts, while 2-layer HRS adds inter-group and inner-group common streams for grouped users. Both schemes use linear precoding and successive interference cancellation, with HRS generalizing the decoding hierarchy and reducing to 1-layer RS when inner-group streams are disabled.

3) 1-layer RS:

  • 1-layer RS splits each user message into common and private sub-messages, combining all common parts into one stream decoded by every user and privately decoding each user’s stream.This creates K + 1 linearly precoded streams from K user messages.Source: 3) 1-layer RS · pp. 10–11
  • Each user in 1-layer RS decodes the common stream first, removes it through SIC, then decodes its private stream while treating remaining private-stream interference as noise.The user reconstructs its message by combining the allocated common and private rates.Source: 3) 1-layer RS · p. 11
  • 1-layer RS supports low-complexity or optimization-based precoder design, including ZFBF private beams, random common beams, WSR and worst-case-rate maximization, EE maximization, and transmit-power minimization.When fairness or individual QoS constraints are optimized, splitting all user messages is required for optimality; without QoS sum-rate maximization, splitting one message can suffice,,,.Source: 3) 1-layer RS · pp. 10–11

4) 2-layer HRS:

  • 2-layer HRS groups users and splits each message into inter-group common, inner-group common, and private sub-messages to improve robustness to imperfect CSIT and achievable rates.The scheme was initially proposed for FDD massive MIMO.Source: 4) 2-layer HRS · p. 11
  • HRS encodes one inter-group stream for all users, one inner-group stream per group, and one private stream per user, producing K + G + 1 linearly precoded streams.Inter-group common sub-messages are decoded by everyone, inner-group common messages by users within each group, and private messages by their intended users.Source: 4) 2-layer HRS · p. 11
  • Each HRS user applies two SIC layers, decoding the inter-group stream first, the inner-group stream second, and its private stream third.This ordering follows the scheme’s hierarchical stream structure and supports separate inter-group and inner-group common rates.Source: 4) 2-layer HRS · pp. 11–12
  • For four users divided into two groups, HRS uses one all-user stream and two group-specific streams; disabling the group-specific streams reduces it to 1-layer RS.In the example, s1234 is decoded by all users, while s12 and s34 are decoded only within their respective groups.Source: 4) 2-layer HRS · p. 12
  • Generalized RS increases message splits and SIC layers with user count, requiring L = 2K−1 splits and 2K−1 −1 SIC layers to create streams for different user subsets.This broader design targets higher achievable rate and QoS at the expense of greater transceiver complexity.Source: 4) 2-layer HRS · p. 12

5) Generalized RS: · 6) RS-CMD:

Generalized RS organizes messages into streams decoded by different user subsets and ordered by stream order, while RS-CMD splits each message into common and private streams without message combiners. Both schemes use linear precoding and successive interference cancellation to obtain achievable user rates.

5) Generalized RS:

  • Generalized RS creates l-order streams decoded by l users, while streams intended for other users are treated as noise.Stream order specifies the number of users decoding each stream.Source: 5) Generalized RS · p. 12
  • Generalized RS linearly precodes each l-order stream vector with a matrix containing the corresponding subset-specific precoders.The transmit signal is constructed from precoded stream vectors across stream orders.Source: 5) Generalized RS · p. 12
  • Each generalized-RS user decodes intended streams through 2^K−1−1 SIC layers, proceeding from the K-order stream to its 1-order private stream.Within each order, users follow an order π_l,k for decoding the streams containing user k.Source: 5) Generalized RS · p. 12
  • In the three-user generalized-RS example, each message is split into four sub-messages and encoded into one 3-order, three 2-order, and three 1-order streams.Each user decodes four intended streams using three SIC layers.Source: 5) Generalized RS · p. 13

6) RS-CMD:

  • RS-CMD splits each user message into common and private sub-messages, independently encodes them into 2K streams, and linearly precodes the resulting streams.Common streams are decoded by all users in the stated configuration, whereas each private stream is decoded only by its corresponding user.Source: 6) RS-CMD · p. 13
  • RS-CMD differs from conventional RSMA message processing by turning off message combiners at both the transmitter and receiver.The scheme was initially proposed for cloud-radio access networks (C-RAN).Source: 6) RS-CMD · p. 13
  • RS-CMD can further optimize the user group decoding each common stream to enhance performance.The supplied configuration assumes all K common streams are decoded by all users.Source: 6) RS-CMD · p. 13
  • Each RS-CMD user applies K SIC layers to decode the common streams in an order π_k before decoding its intended private stream.The achievable user rate combines the common-stream rate guaranteed across users with the private-stream rate.Source: 6) RS-CMD · p. 13

7) DPCRS: … 1) Coordinated transmission:

The section extends RSMA through DPC- and THP-based private-stream encoding, uplink capacity-region access, and multi-cell interference coordination. These designs improve robustness, practicality, or spectral efficiency while introducing trade-offs in complexity, overhead, and implementation losses.

7) DPCRS:

  • DPCRS combines linearly precoded common streams with non-linear DPC private streams to compensate DPC’s vulnerability to imperfect CSIT.The framework includes 1-DPCRS and M-DPCRS, while extending other linearly precoded strategies to DPC-based counterparts.Source: 7) DPCRS · pp. 13–14

8) THPRS:

  • THPRS, uses THP for private messages and beamforming to manage interference, offering a less-complex but suboptimal DPCRS implementation.Modulo operations at the transmitter and receiver create power and modulo losses.Source: 8) THPRS · p. 14

C. Uplink RSMA

  • Uplink RSMA achieves every point of the Gaussian MAC capacity region without time sharing, including rate points between SIC’s corner points.Rate splitting avoids joint encoding’s complexity and time sharing’s coordination overhead, supporting grant-free access services such as URLLC and mMTC.Source: C. Uplink RSMA · p. 14
  • In K-user uplink RSMA, splitting K −1 users’ messages suffices for capacity-region access, but decoding 2K −1 streams requires 2K −2 SIC layers.Power-allocation optimization for sum rate has been investigated in, with the capacity benefit requiring higher receiver complexity.Source: C. Uplink RSMA · pp. 14–15

D. Multi-cell RSMA

  • Multi-cell RSMA improves spectral efficiency in coordinated and cooperative networks by managing both intra-cell and inter-cell interference.Coordination avoids data sharing, whereas cooperation relies on data sharing; both generally require CSI sharing among base stations.Source: D. Multi-cell RSMA · p. 15

1) Coordinated transmission:

  • In coordinated multi-cell transmission, inter-cell interference is partially decoded and partially treated as noise through message splitting, linear precoding, and successive common-stream cancellation.Each user decodes relevant common streams before its desired private stream, following RS-CMD principles.Source: 1) Coordinated transmission · p. 15
  • Splitting each user’s message into more than two parts, with groups decoding different parts according to instantaneous CSIT quality, can further boost MISO-BC degrees of freedom.This extends coordinated RS beyond the two-part common/private split.Source: 1) Coordinated transmission · p. 16

2) Cooperative transmission: … B. RSMA in Multi-Antenna Networks

RSMA’s information-theoretic foundations span interference channels, multiple-access channels, and multi-antenna networks, where message splitting and partial interference decoding yield broad capacity, DoF, fairness, and GDoF gains. In cooperative transmission, RSMA extends to jointly served multi-cell systems subject to per-BS power and fronthaul constraints.

2) Cooperative transmission:

  • In cooperative transmission, multi-cell transmitters share user data and CSI through a central processor, operate as a virtual joint transmitter, and apply downlink RSMA under per-BS power and fronthaul constraints,.With one antenna per transmitter, the cooperative network is equivalent to a single-cell MIMO BC under per-antenna power constraints.Source: 2) Cooperative transmission · p. 16

A. RSMA in Single-Antenna Networks

  • In single-antenna networks, Han–Kobayashi rate-splitting partially decodes interference and partially treats it as noise, bridging full decoding and full interference-as-noise strategies,,.This mechanism is identified as the fundamental principle underlying RS applications in wireless networks.Source: A. RSMA in Single-Antenna Networks · p. 16
  • For the SISO MAC, RSMA achieves every boundary point of the capacity region without time sharing or user synchronization.With suitable decoding order and power allocation, uplink RSMA attains the entire two-user SISO MAC capacity region.Source: A. RSMA in Single-Antenna Networks · pp. 16–17
  • RSMA’s multi-antenna development includes optimal sum-DoF under imperfect CSIT, a larger rate region than SDMA, and linearly precoded RS outperforming DPC when CSIT is imperfect,.These results established RSMA’s robustness and performance advantages in multi-antenna broadcast channels.Source: A. RSMA in Single-Antenna Networks · p. 18

III. INFORMATION-THEORETIC BACKGROUND AND MILESTONES OF RSMA

  • The information-theoretic literature on RSMA covers four main multi-antenna metrics: sum-DoF and DoF regions, max-min fairness or symmetric-DoF, GDoF, and capacity regions.These metrics organize the survey’s analysis of RSMA in multi-antenna networks.Source: B. RSMA in Multi-Antenna Networks · p. 19

B. RSMA in Multi-Antenna Networks

  • RS achieves optimal DoF regions in several multi-antenna settings, including underloaded MISO BCs with partial CSIT, asymmetric MIMO BCs, and heterogeneous-CSIT overloaded MISO BCs.For underloaded MISO BCs, RS matches the sum-DoF upper bound and achieves the entire DoF region; in overloaded systems, it combines rate-splitting with superposition coding for no-CSIT users.Source: B. RSMA in Multi-Antenna Networks · p. 19
  • RS outperforms SDMA and NOMA in sum-DoF and max-min fairness across underloaded and overloaded MISO BCs with perfect or imperfect CSIT.The surveyed results show that RS exploits multi-antenna DoF and successive-interference-cancellation benefits across these regimes.Source: B. RSMA in Multi-Antenna Networks · p. 19
  • RS-based schemes also achieve the GDoF region of the two-user underloaded MISO BC with imperfect CSIT and the two-user MIMO BC capacity region within a constant gap.The capacity results include sum-capacity achievement and then entire-region achievement within a constant gap for perfect CSIT.Source: B. RSMA in Multi-Antenna Networks · p. 19

C. Milestones of RSMA … 1) Joint Resource Optimization:

RSMA has progressed from information-theoretic foundations toward communication-theoretic resource-allocation frameworks for practical wireless systems. Joint optimization of precoders, power, and common rates is formulated for perfect, bounded-error, and statistical CSIT, with algorithms balancing optimality and tractability.

C. Milestones of RSMA

  • RSMA milestones motivate next-generation multiple access focused on improving SE, EE, throughput, latency, reliability, robustness, coverage, and cost.The paper summarizes these milestones in Fig. 16 and identifies continued evolution as a driver of future research.Source: C. Milestones of RSMA · p. 19

IV. COMMUNICATION-THEORETIC BACKGROUND OF RSMA

  • Communication-theoretic RSMA research in the moderate-SNR regime centers on resource allocation, including precoding, power control, common-rate allocation, user scheduling, and subcarrier allocation.The literature is organized around conventional multi-antenna broadcast channels, first for single-carrier and then for multi-carrier scenarios.Source: §IV · p. 19

A. Resource Allocation for Single-Carrier RSMA

  • Single-carrier RSMA resource-allocation research divides into joint optimization, which targets maximum achievable performance, and low-complexity schemes, which trade performance against implementation complexity.The main single-carrier variables are precoders, power, and common-rate allocation, particularly in the finite-SNR regime.Source: A. Resource Allocation for… · pp. 19–20

1) Joint Resource Optimization:

  • Imperfect-CSIT optimization uses worst-case designs for bounded estimation errors and long-term designs for unbounded errors, respectively guaranteeing performance over uncertainty regions or optimizing ergodic rates over fading states.When estimation error vanishes, both imperfect-CSIT formulations reduce to the perfect-CSIT problem; RSMA therefore defines a more general class of MIMO optimization and transmission strategies.Source: 1) Joint Resource Optimization · pp. 20–22
  • For perfect CSIT, joint RSMA optimization maximizes WSR, minimum user rate, or EE, or minimizes sum power by jointly designing precoders, power allocation, and common-rate allocation.Setting all common-rate allocations to zero reduces the formulation to MU-LP design for SDMA, showing that RSMA contains SDMA as a special case and enlarges the optimization space.Source: 1) Joint Resource Optimization · pp. 20–21
  • Joint RSMA optimization is non-convex, so research combines globally optimal SIT branch-and-bound with tractable WMMSE, SCA, ADMM, and semidefinite-relaxation algorithms.These methods target KKT points, while numerical results report WMMSE WSR and SCA EE performance nearly coinciding with corresponding globally optimal results.Source: 1) Joint Resource Optimization · p. 22
  • Although suboptimal algorithms have tractable complexity, real-world deployment remains challenging because methods such as SCA repeatedly solve convex subproblems using locally tight approximations.This computational burden motivates the separate research line on low-complexity resource allocation.Source: A. Resource Allocation for… · p. 20, 1) Joint Resource Optimization · p. 22

2) Low-Complexity Resource Allocation:

Low-complexity RSMA resource allocation separates precoder-direction design from power optimization, reducing computational burden while supporting several closed-form or tractable designs. These methods cover common-stream beamforming, private-stream precoding, and power allocation under perfect and imperfect CSIT.

Low-Complexity Resource Allocation

  • Separating precoder directions from powers reduces complexity; 1-layer RS uses one common and K private streams, with the common direction typically designed multicast-style.The common stream must be decoded by all users, whereas private streams are decoded only by their intended users.Source: 2) Low-Complexity Resource Allocation · p. 22
  • Common-stream directions can use random precoding, weighted matched beamforming, or SVD, trading DoF sufficiency, common-rate maximization, and tractable initialization,,,,,, –.Random precoding suffices for the entire DoF region but is inefficient for improving spectral efficiency; equal-weighted MBF with k = 1 is commonly employed.Source: 2) Low-Complexity Resource Allocation · p. 23
  • Private-stream directions use linear methods including ZFBF, RZF/MMSE, and regularized block diagonalization, extending low-complexity designs from MISO to multi-antenna MIMO broadcast channels [27], [30],,,,,,.RZF reduces to ZFBF when κ = 0, while regularized block diagonalization supports users with Q receive antennas.Source: 2) Low-Complexity Resource Allocation · p. 23
  • ZFBF for private streams with random common-stream beamforming achieves the optimal DoF region for the MISO broadcast channel with imperfect CSIT,,.ZFBF is limited to the underloaded MISO broadcast channel, where M ≥ K.Source: 2) Low-Complexity Resource Allocation · p. 23
  • Power allocation is governed by the private-stream fraction τ, with closed-form solutions reported for perfect and imperfect CSIT and equal-power or water-filling allocation across private streams,,.For total transmit power P, Pc = (1−τ)P and Pc + P1 + … + PK = P.Source: 2) Low-Complexity Resource Allocation · p. 23

3) Resource Allocation by Machine Learning: · B. Resource Allocation for Multicarrier RSMA

Machine learning has optimized RSMA power allocation, while multicarrier RSMA additionally requires user scheduling and subcarrier assignment alongside precoding, power, and common-rate allocation. Existing studies report RSMA gains over SDMA but leave imperfect-CSIT scheduling and low-complexity allocation as open problems.

3) Resource Allocation by Machine Learning:

  • Deep reinforcement learning designs transmit-stream power allocation and yields significant RSMA gains over SDMA under imperfect CSIT.The method optimizes power allocation for given precoders.Source: 3) Resource Allocation by Machine… · p. 23

B. Resource Allocation for Multicarrier RSMA

  • Multicarrier RSMA requires joint precoder design, power control, common-rate allocation, user scheduling, and subcarrier assignment because users share and span multiple subcarriers.These variables extend the resource-allocation problem beyond single-carrier RSMA.Source: B. Resource Allocation for… · p. 23
  • The first multicarrier RSMA study jointly optimizes subcarrier allocation, precoding, power, and common rates to maximize minimum user rate in multigroup multicast systems.The resulting problem is mixed-integer and nonlinear because of binary subcarrier-allocation indicators.Source: B. Resource Allocation for… · p. 23
  • Continuous relaxation, semidefinite programming, and successive convex approximation solve the mixed-integer multicarrier allocation problem, while exhaustive searches evaluate optimal MMF-rate performance,.The exhaustive approach searches all subcarrier assignments and optimizes precoding, power, and common-rate allocation for each.Source: B. Resource Allocation for… · pp. 23–24
  • RSMA without subcarrier optimization can outperform SDMA with subcarrier optimization, while serving users with arbitrary channel conditions on each subcarrier,.Unlike SDMA, RSMA does not require semi-orthogonal users, potentially reducing transmitter-side subcarrier-allocation and scheduling overhead.Source: B. Resource Allocation for… · p. 24
  • Multicarrier RSMA studies, assume perfect CSIT, whereas imperfect-CSIT work, serves all legitimate users on every subcarrier.Thus, imperfect-CSIT subcarrier allocation and user scheduling remain open problems requiring lower-complexity strategies.Source: B. Resource Allocation for… · p. 24

V. RSMA PHY LAYER DESIGN … 1) Transmitter:

RSMA’s practical PHY architecture combines message splitting and combining with finite-blocklength coding, finite-constellation modulation, adaptive modulation and coding, precoding, and successive interference cancellation. Its receiver partially decodes interference through the common stream while treating residual interference as noise on private streams.

V. RSMA PHY LAYER DESIGN

  • RSMA’s practical PHY design extends the abstract framework with finite-constellation modulation, finite-length polar codes, and AMC, with extensions to delayed-CSIT, mobility, V-BLAST, satellite, and joint sensing scenarios,,,, [132].The section uses the architecture in to explain bit-level transmitter and receiver processing.Source: §V · p. 24, A. A Quick Introduction to RSMA PHY… · p. 24

A. A Quick Introduction to RSMA PHY Layer

1) RSMA transmitter operation:

  • Transmission rates determine message-bit counts, modulation, coding parameters, and common-rate allocations, using rates computed from CSIT or obtained through user feedback under FBL and BICM assumptions.For four channel uses, the number of message bits per stream equals four times its transmission rate.Source: 1) RSMA transmitter operation · pp. 24–25
  • In the examples, independent common parts may be split and concatenated using different valid patterns, then separately encoded, interleaved, and modulated according to each stream’s rate.Example 3 uses 16-QAM and coding rate 5/8 for user-1’s 2.5 bps/Hz private stream, while lower-rate streams use QPSK.Source: 1) RSMA transmitter operation · p. 25

1) RSMA receiver operation:

  • Each receiver equalizes, computes LLRs, deinterleaves, and decodes the common stream, decomposes its message, keeps its own common part, and discards the other user’s part.After reconstruction and interference cancellation, it decodes the corresponding private stream and desplits the recovered common and private parts into the intended message.Source: 1) RSMA receiver operation · pp. 26–27, Figure 19 · p. 26
  • RSMA receivers bridge decoding interference and treating interference as noise: common-stream processing partially decodes interference, whereas private-stream processing treats part of it as noise.This differs from SDMA’s all-interference-as-noise reception and NOMA’s extreme decoding or noise-treatment modes.Source: 1) RSMA receiver operation · pp. 26–27

B. RSMA PHY Layer Architecture

  • The two-user 1-layer RS architecture integrates message splitting, finite-alphabet modulation, finite-length polar coding, AMC, precoding, and SIC, and can extend to the RSMA schemes discussed earlier [36].The transmitter independently encodes the common and private messages into codewords, then interleaves their bits before modulation.Source: B. RSMA PHY Layer Architecture · p. 27, 1) Transmitter · p. 27

1) Transmitter:

  • AMC selects modulation and coding rates from C1, C2, R1, and R2 as link-quality metrics, after which the streams are mapped to antennas through precoding; alternative LDPC and PSK implementations are possible,.The reference architecture uses polar codes and 4-QAM, 16-QAM, 64-QAM, or 256-QAM, while other coding and modulation techniques remain applicable.Source: 1) Transmitter · p. 27

2) Receiver: … 2) Scheduler complexity:

The section describes RSMA receiver processing, compares RSMA with existing multiple-access schemes, and analyzes their encoder and scheduler complexity. It emphasizes RSMA’s generality, flexible interference management, and differing scheduling burdens across schemes.

2) Receiver:

  • The receiver equalizes the signal, decodes common-message bits using soft decisions, reconstructs and removes the common symbol through hard-decision SIC, then decodes and combines private and common parts.MMSE equalization is used before LLR calculation and channel decoding; the recovered common component is subtracted before private-stream detection.Source: 2) Receiver · p. 28

VI. COMPARISON OF MULTIPLE ACCESS SCHEMES

  • The comparison covers multiple-access design principles, DoF, SE, LLS, and complexity, summarizing RSMA’s advantages and disadvantages relative to existing schemes.Source: §VI · p. 28

A. Framework Comparison

  • RSMA is more general than OMA and NOMA, while generalized RS, hierarchical RS, 1-layer RS, and RS-CMD form related downlink frameworks with explicit reduction relationships.Generalized RS contains the other linearly precoded schemes as subschemes except RS-CMD, whose common streams serve individual users without jointly encoding their messages.Source: A. Framework Comparison · p. 28
  • RSMA flexibly decodes part of interference and treats the remainder as noise, thereby encompassing multicasting, OMA, SDMA, and NOMA without hard switching between operating modes.In the two-user case, 1-layer RS specializes to SDMA, NOMA, OMA, or multicasting through appropriate common/private stream and power assignments.Source: A. Framework Comparison · pp. 28–29

B. Complexity Comparison

  • Message splitting makes RSMA encoder complexity higher than other multiple-access schemes; among linearly precoded RSMA schemes, 1-layer RS has the lowest complexity, adding one stream in the K-user case.For 1-layer RS, K messages are encoded into K streams in SDMA and NOMA, whereas 2-layer HRS adds G group-specific common streams plus the all-user common stream.Source: 1) Encoder complexity · p. 29

1) Encoder complexity:

  • 1-layer RS has the lowest encoding complexity among linearly precoded RSMA schemes, requiring only one additional encoded stream compared with SDMA and NOMA.The encoder comparison counts encoded streams; RS-CMD and generalized RS require increasingly more sub-messages as K grows.Source: 1) Encoder complexity · p. 29

2) Scheduler complexity:

  • SDMA minimizes encoder and receiver complexity but can require complex scheduling because user pairing, accurate CSIT, and channel conditions affect performance; SIC-based schemes additionally require grouping and decoding-order optimization.Multi-antenna NOMA, 2-layer HRS, and RS-CMD have high grouping complexity, while multi-antenna NOMA, generalized RS, and RS-CMD also face many decoding-order choices, including K! possibilities for single-group NOMA.Source: 2) Scheduler complexity · pp. 29–30
  • 1-layer RS has relatively the lowest scheduler complexity because it needs neither user grouping nor decoding ordering, while generalized RS and 1-layer RS can outperform scheduled SDMA when CSIT is sufficiently inaccurate.All RSMA schemes are described as suitable for arbitrary channel-strength disparities and channel directions; multi-antenna NOMA and RS-CMD have relatively the highest scheduler complexity.Source: 2) Scheduler complexity · p. 30

3) Receiver complexity: · C. Performance Comparison · 1) DoF region:

The comparison finds that RSMA combines low receiver complexity with optimal DoF performance under imperfect CSIT. In particular, 1-layer RS achieves broad DoF gains and supports more users with fewer SIC layers than competing schemes.

3) Receiver complexity:

  • 1-layer RS and 2-layer HRS require only one and two SIC operations per user, respectively, independent of K, limiting complexity and error propagation.By contrast, NOMA, generalized RS, and RS-CMD require SIC layers that increase with the number of users.Source: 3) Receiver complexity · p. 30
  • Generalized RS and RS-CMD incur higher receiver complexity, motivating low-complexity designs that control decoded common streams and simplify decoding orders.Generalized RS also explores performance gains and identifies common streams that may be disabled to reduce encoding and decoding complexity.Source: 3) Receiver complexity · pp. 30–31

C. Performance Comparison

  • The performance comparison evaluates multiple-access schemes using DoF, rate, operational-region, user-fairness, and LLS metrics, with the operational region describing favorable user-channel conditions.The DoF analysis interprets DoF as the high-SNR rate pre-log or number of simultaneously transmitted interference-free streams.Source: C. Performance Comparison · p. 31, 1) DoF region · p. 31

1) DoF region:

  • 1-layer RS achieves the optimal DoF region for the multi-antenna BC with imperfect CSIT, so other multiple-access strategies can achieve only equal or lower DoF.The two-user MISO BC comparison in Fig. 23 confirms an explicit DoF gain for RSMA under imperfect CSIT with α = 0.6.Source: 1) DoF region · p. 31, Figure 23 · p. 31
  • With imperfect CSIT, SDMA gives each user DoF α, whereas 1-layer RS provides corner points (1, α) and (α, 1) by allocating the common stream between users.The line segment between these corner points is achieved by adjusting the common-stream allocation.Source: 1) DoF region · p. 31
  • At M = 4, K = 6, and α = 0.5, 1-layer RS achieves the highest sum-DoF and MMF-DoF while requiring much lower complexity than NOMA.Multi-antenna NOMA with G = 3 has better sum-DoF but worse MMF-DoF than NOMA with G = 1.Source: Figure 24 · p. 31, 1) DoF region · p. 32
  • With M = 6 and α = 0.5, 1-layer RS provides explicit sum- and MMF-DoF gains over all schemes for every tested K.SDMA and multi-antenna NOMA with G = 1 have MMF-DoF zero beyond K > 6 and K > 7, respectively.Source: 1) DoF region · p. 32, Figure 25 · p. 32
  • At an MMF-DoF threshold of 0.1, 1-layer RS serves around 15 users with one SIC layer, compared with at most 10 users for NOMA using nine layers.This demonstrates greater user-support efficiency for RSMA than NOMA under the stated imperfect-CSIT setting.Source: 1) DoF region · p. 32

2) Rate region: · 3) Operational region: · 4) User fairness:

Across rate, operational, and fairness analyses, RSMA enlarges achievable regions, adapts to channel conditions and user priorities, and improves fairness while reducing interference-management and receiver complexity. Its common streams flexibly combine transmitter- and receiver-side interference cancellation, providing robustness to imperfect CSIT and overloaded deployments.

2) Rate region:

  • DPCRS enlarges the ergodic-rate region of all compared schemes, while linearly precoded RS can outperform DPC under poor CSIT.DPC and SDMA regions shrink rapidly as CSIT quality decreases, whereas RSMA bridges transmitter- and receiver-side interference cancellation and remains more robust to CSIT uncertainty.Source: 2) Rate region · p. 33
  • Generalized RS outperforms other schemes in three-user ergodic sum rate, especially in overloaded networks, while 1-layer RS improves spectral efficiency with lower transceiver complexity.Generalized RS gains most when interference is strong and common-stream power is sufficient; 1-layer RS needs one SIC layer per user, versus six decoding orders and two SIC layers for multi-antenna NOMA.Source: 2) Rate region · p. 33, Figure 27 · p. 34
  • 1-layer RS achieves explicit weighted-sum-rate gains over multi-antenna NOMA, SDMA, and multicast in an overloaded ten-user deployment.RS uses the maximum DoF of 2 by packing eight users into a common stream and serving two users with private streams, while MU–LP and multi-antenna NOMA achieve DoF 1.Source: 2) Rate region · p. 34, Figure 28 · p. 34

3) Operational region:

  • RSMA’s operational region changes with channel alignment and user weights: it reduces to SDMA for sufficiently orthogonal channels, to NOMA under fairness-driven aligned deployments, and otherwise can outperform all alternatives.NOMA has no benefit when the stronger-channel user receives equal or higher weight, while RSMA retains an explicit superiority region over SDMA, NOMA, and OMA for different user weights.Source: 3) Operational region · pp. 34–35, Figure 29 · p. 34

4) User fairness:

  • 1-layer RS provides superior max-min rates to multi-antenna NOMA and SDMA for a six-user MISO broadcast channel with imperfect CSIT.With α = 0.5 and M = K = 6, this performance is achieved with one SIC layer per user, whereas multi-antenna NOMA requires substantially greater scheduling and receiver complexity.Source: 4) User fairness · p. 35, Figure 30 · p. 35
  • 1-layer RS allocates a large portion of transmit power to the common stream, with common-stream power increasing with SNR as interference strengthens.The common stream improves fairness by serving users jointly, while message splitting and power allocation adjust interference cancellation between receivers and the transmitter.Source: 4) User fairness · pp. 35–36

5) LLS Performance: · D. Advantages of RSMA · E. Disadvantages of RSMA

LLS experiments show that RSMA achieves substantial throughput gains over SDMA and NOMA under imperfect CSIT and user mobility. Across broader comparisons, RSMA offers flexible, robust, efficient, and lower-complexity transmission, while introducing SIC, encoding, signaling, and optimization burdens.

5) LLS Performance:

  • RSMA achieves significant throughput gains over SDMA and NOMA in two-user MISO BCs with imperfect CSIT, with SDMA saturating at high SNR.The LLS throughput trend follows Shannon bounds, while RSMA’s gain over SDMA exceeds the corresponding bound-based expectation.Source: 5) LLS Performance · p. 36
  • With an 8 bps/Hz QoS constraint, RSMA supports 40 km/h user speed versus 5 km/h for SDMA, demonstrating greater robustness to mobility.RSMA’s robustness follows from partially decoding interference while partially treating it as noise; SDMA can incur AMC throughput losses after switching to single-user transmission.Source: 5) LLS Performance · pp. 36–37, Figure 32 · p. 36

D. Advantages of RSMA

  • RSMA is universal because it subsumes SDMA, NOMA, OMA, and multicasting, thereby achieving equal or better performance within a unified framework.Its common-stream design flexibly combines transmitter-side and receiver-side interference mitigation across underloaded, overloaded, and intermediate channel conditions.Source: D. Advantages of RSMA · p. 37
  • RSMA remains robust to imperfect CSIT from quantization, pilot contamination, estimation errors, and mobility, whereas conventional schemes are more vulnerable to such impairments.It achieves robust connectivity and significant sum-rate and throughput gains under mobility, unlike SDMA’s rapid saturation and speed-related degradation.Source: D. Advantages of RSMA · p. 37
  • RSMA provides higher spectral and energy efficiency, achieving equal-or-better SE and outperforming existing MA schemes in EE across diverse loads and deployments.Its rate-region and sum-DoF advantages extend across perfect and imperfect CSIT, including cases where linearly precoded RSMA exceeds DPC under imperfect CSIT.Source: D. Advantages of RSMA · pp. 37–38
  • RSMA enhances QoS, fairness, coverage, and latency through MMF-DoF and max-min-rate gains, cooperative rate-splitting, and shorter finite-blocklength transmissions.CRS with relaying improves cell-edge performance for users with large channel-strength disparities, while RSMA supports lower-latency URLLC transmission.Source: D. Advantages of RSMA · p. 38
  • RSMA can reduce transceiver, scheduling, receiver, and CSI-feedback complexity relative to several schemes, despite requiring message splitting and at least one SIC layer.Unlike multi-antenna NOMA, 1-layer RS avoids user ordering and grouping and requires only one SIC layer per user.Source: D. Advantages of RSMA · p. 38, E. Disadvantages of RSMA · p. 38

E. Disadvantages of RSMA

  • RSMA incurs higher encoding, signaling, and optimization burdens because message splitting requires additional streams, coordination, joint precoder-rate optimization, and scenario-dependent power allocation.For generalized RS and M-DPCRS, SIC layers also grow exponentially with user count, increasing receiver complexity.Source: E. Disadvantages of RSMA · p. 38

VII. EMERGING APPLICATIONS, CHALLENGES, AND FUTURE RESEARCH TRENDS OF RSMA … 4) Non-orthogonal unicast and multicast:

RSMA research spans emerging 5G/6G technologies and addresses unresolved PHY, cross-layer, performance, standardization, and implementation challenges. Across these applications, RSMA improves interference management, spectral and energy efficiency, robustness, and fairness, while several theoretical and practical limits remain open.

VII. EMERGING APPLICATIONS, CHALLENGES, AND FUTURE RESEARCH TRENDS OF RSMA

  • RSMA applications are expanding across massive MIMO, IRS, VLC, UAV, joint communication and sensing, and satellite networks, but the broader research field remains in its infancy.Open problems include coding and modulation, precoding and receiver design, joint scheduling and precoding, subcarrier allocation, large-scale networks, stochastic geometry, throughput, error rate, and outage analysis.Source: §VII · pp. 38, 40

1) MIMO BC:

  • In MIMO BC, RSMA achieves substantial ESR, sum-DoF, and throughput gains over MU–MIMO and MIMO NOMA, with gains increasing with common-stream count and imperfect CSIT.Theoretical results establish optimal or achievable DoF regions in several MIMO BC settings, but the capacity region with imperfect CSIT and practical stream-count tradeoffs remain open.Source: 1) MIMO BC · p. 40

2) MIMO IC:

  • In MIMO interference channels, RSMA achieves the best known DoF region in several imperfect-CSIT settings and improves SE–EE tradeoffs, but capacity limits and practical transceiver designs remain unknown.Combining RSMA with interference alignment may unite RSMA’s robustness with interference alignment’s dimensional confinement, motivating cross-layer resource allocation research.Source: 2) MIMO IC · p. 40

3) Alternative Receiver Design:

  • Under imperfect CSIR, SIC error propagation degrades RSMA reliability, whereas a deep-learning receiver improves symbol detection and error-rate performance before and after cancellation.These results identify imperfect-CSIR error-rate performance as an important direction for alternative receiver design.Source: 3) Alternative Receiver Design · p. 41

4) Hybrid Automatic Repeat Request (HARQ) Design:

  • RSMA has not yet been integrated with HARQ, because message splitting and common/private decoding create unresolved encoding and retransmission choices; waveform impairments likewise remain underexplored.RSMA may improve waveform robustness and spectral efficiency, including against vulnerabilities such as OFDM intercarrier interference under Doppler shift.Source: 4) Hybrid Automatic Repeat Request… · p. 41, 5) Waveform Design · p. 41

1) Massive MIMO:

  • RSMA addresses massive-MIMO CSIT bottlenecks through hierarchical common streams, pilot-contamination mitigation, and robustness to impairments, while cell-free architectures and low-resolution converters remain research priorities.In FDD massive MIMO, 2-layer HRS manages inter- and intra-group interference and achieves sum-rate gains over two-tier precoding; low-resolution ADC/DAC robustness is unresolved.Source: 1) Massive MIMO · pp. 41–42

3) Multigroup multicasting:

  • RSMA improves multigroup multicast fairness, max-min rate, energy efficiency, and DoF under perfect and poor CSIT, motivating integrated terrestrial-satellite network designs.It enables partial decoding and partial noise treatment of inter-group interference, but multicast rates remain constrained by each group’s worst-channel user.Source: 3) Multigroup multicasting · p. 42

4) Non-orthogonal unicast and multicast:

  • RSMA-aided NOUM improves SE and EE without increasing receiver complexity, while dirty-paper-coded RS enlarges the rate region under imperfect CSIT; multigroup extensions must address three-dimensional interference.Unicast messages are split into common and private parts, with common parts combined with multicast content into a super-common stream decoded using one SIC layer.Source: 4) Non-orthogonal unicast and multicast · p. 43

5) Multi-cell MIMO including coordinated multi-point (CoMP), cloud-radio access network (C-RAN), and fog-radio access network (F-RAN): … 13) Cognitive radio:

Across multi-cell, energy-transfer, relaying, caching, aerial, satellite, sensing, and cognitive-radio settings, RSMA improves interference management and system performance while exposing application-specific design challenges. These challenges include fronthaul and cache optimization, imperfect CSIT and mobility, secrecy, cooperation, and spectrum-sharing constraints.

5) Multi-cell MIMO including coordinated multi-point (CoMP), cloud-radio access network (C-RAN), and fog-radio access network (F-RAN):

  • In coordinated and cooperative multi-cell networks, RSMA improves SE/EE, outperforms interference-as-noise strategies, and can reduce fronthaul or caching requirements, but inter-cell decoding increases receiver burden.Cooperative RSMA also simplifies receiver design through shared common streams and achieves gains over MU–LP and NOMA under diverse channel disparities.Source: 5) Multi-cell MIMO including… · p. 43

6) Simultaneous wireless information and power transfer (SWIPT):

  • For SWIPT, 1-layer RS enlarges information-receiver rate regions under harvested-power constraints, can avoid dedicated energy signals, and saves more transmit power than SDMA for co-located receivers.RSMA’s common and private streams jointly manage information interference and energy delivery, while limited range and secrecy against energy-receiver eavesdropping remain open issues.Source: 6) Simultaneous wireless information… · pp. 43–44

7) Cooperative user relaying:

  • Cooperative rate-splitting (CRS) forwards decoded common streams in a second phase and achieves substantial rate-region and max-min-fairness gains over cooperative NOMA, 1-layer RS, and MU–LP, even at low SNR.CRS also enhances PHY-layer security, but imperfect-CSIT design and joint precoding–relay selection remain underexplored.Source: 7) Cooperative user relaying · p. 44

8) Wireless caching:

  • RSMA-aided caching splits messages into cached and uncached parts, boosting DoF and coded-caching gain while reducing CSIT requirements; optimizing placement, delivery, and fronthaul-constrained F-RAN operation remains open.System-level co-design further exploits gains from jointly designing wireless caching and RSMA.Source: 8) Wireless caching · p. 44

9) Unmanned aerial vehicles-aided communications:

  • In UAV-aided networks, RSMA improves sum rate, weighted sum rate, and energy efficiency over SDMA or NOMA across replacement, mmWave, placement, and satellite–aerial scenarios.Trajectory design, mobility effects, channel uncertainty, and UAV lifetime implications remain insufficiently studied.Source: 9) Unmanned aerial vehicles-aided… · pp. 44–45

10) Physical layer security:

  • RSMA enhances secrecy through common-stream interference and achieves max-min secrecy or secrecy-sum-rate gains over MU–LP and NOMA, while internal eavesdroppers and pilot-contamination attacks remain unresolved.The cited security studies also report higher information-user sum rate than SDMA under equal jamming power.Source: 10) Physical layer security · p. 45

11) Satellite Communications:

  • In multibeam and integrated terrestrial-satellite systems, RSMA manages inter-beam and inter-gateway interference and provides MMF-DoF, rate, throughput, secrecy-rate, and EE gains over MU–LP or non-cooperative strategies.Current integrated designs remain limited in beam scope and terrestrial–satellite cooperation.Source: 11) Satellite Communications · p. 45

12) Integrated Radar/Sensing and Communication; 13) Cognitive radio:

  • RSMA improves the WSR–MSE tradeoff in DFRC, handles communication–radar interference, and retains sum-rate gains over SDMA with imperfect CSIT and low-resolution DACs; in MIMO cognitive radio, it reduces transmit power under SU-rate and PU-interference constraints.Security in RSMA-aided DFRC and cognitive-radio secrecy under SU–PU spectrum sharing remain open problems.Source: 12) Integrated Radar/Sensing and… · p. 46, 13) Cognitive radio · p. 46

14) Massive machine-type communication: … VIII. CONCLUSION

The surveyed applications show RSMA’s promise for massive IoT, VLC, IRS-aided networks, and other 6G scenarios, while standardization still requires solutions for signaling, HARQ, and system-level evaluation. Overall, the paper presents RSMA as a promising 6G framework for interference management, non-orthogonal transmission, and multiple access.

14) Massive machine-type communication:

  • RSMA is appealing for mMTC because it can improve spectral efficiency, reduce receiver complexity, and remain robust to CSIT inaccuracies and network loads.RSMA-aided power partitioning achieves the optimal DoF region in an overloaded MISO BC with heterogeneous CSIT, while also providing finite-SNR spectral-efficiency gains,.Source: 14) Massive machine-type communication · p. 46

15) Visible light communication:

  • RSMA delivers clear sum-rate gains over conventional multiple-access schemes in MU–MIMO VLC, including highly correlated user channels, but achievable-rate modeling and imperfect CSIT remain open issues.The initial VLC studies used the Shannon formula, whereas [157] derived achievable rates and first studied RSMA-aided VLC with imperfect CSIT.Source: 15) Visible light communication · p. 46

16) Intelligent reconfigurable surface-aided communications:

  • In IRS-aided MISO BCs, jointly optimizing precoders, message splits, and IRS phase shifts gives RSMA better energy efficiency than NOMA and OFDMA, while hierarchical RS improves WSR and outage performance.These gains are reported for 1-layer and 2-layer RS transmission strategies assisted by one or multiple IRSs,, [192],.Source: 16) Intelligent reconfigurable… · p. 47
  • IRS integration may compensate for imperfect CSIT caused by limited channel training, but joint passive-active beamforming, user scheduling, and IRS allocation remain unstudied for RSMA networks.The absence of RF chains makes CSI acquisition at IRSs challenging,.Source: 16) Intelligent reconfigurable… · p. 47

17) Other open issues:

  • RSMA opens research directions across cooperative relaying, SAGIN, V2X, and hybrid 3D eMBB-URLLC-mMTC services by addressing interference, mobility, latency, reliability, and mixed service demands.Combining CRS with NOUM can improve multicast and unicast interference management, while RSMA is identified as promising for SAGIN, V2X, and hybrid 6G services, [106].Source: 17) Other open issues · p. 47

C. Standardization and Implementation of RSMA

  • RSMA standardization requires low-complexity HARQ, additional signaling for message-splitting synchronization, and system-level evaluations that include higher-layer effects.Existing MU–MIMO, CoMP, MBS, NAICS, and MUST mechanisms could reduce the additional standardization and implementation burden,,,,.Source: C. Standardization and Implementation… · p. 48

VIII. CONCLUSION

  • The paper provides a holistic RSMA tutorial and survey covering theory, transmission frameworks, precoder design, comparisons with existing multiple-access schemes, simulations, applications, challenges, and standardization.It surveys downlink, uplink, and multi-cell RSMA; compares DoF, complexity, and performance against SDMA, NOMA, OMA, and multicasting; and discusses PHY-layer design and link-level simulation.Source: §VIII · p. 48
  • RSMA partially decodes interference and partially treats it as noise, offering prospects for higher data rates, reliability, energy efficiency, lower latency, and robustness to CSIT uncertainty in 6G.The conclusion identifies RSMA as a promising PHY-layer paradigm that could reform PHY and lower-MAC design and motivate new research across academia and industry.Source: §VIII · p. 48

Related papers

Loading 2201.03192v3…