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

Wireless networks need broader interference-management and multiple-access strategies as 6G requirements intensify, while RSMA research remains fragmented and lacks a comprehensive tutorial. This paper surveys RSMA's architecture, taxonomy, applications, comparisons, and future challenges, concluding that partial interference decoding and noise treatment make it a promising 6G paradigm, though system-level evaluation and standardization remain open issues.

  • Problem

    Existing RSMA studies focus mainly on narrow scenarios, and no review explains comprehensively why, how, or when RSMA is beneficial.

  • Method

    The paper provides a comprehensive survey of RSMA architecture, receiver operation, scheme complexity, theoretical results, applications, and standardization challenges.

  • Results

    RSMA shows sum- and MMF-DoF gains over other schemes for all tested K in the cited imperfect-CSIT setting, while bridging transmitter- and receiver-side interference cancellation.

  • Takeaways & Limitations

    RSMA is a promising PHY-layer paradigm for interference management, non-orthogonal transmission, and multiple access in 6G.

  • Takeaways & Limitations

    System-level evaluations incorporating higher-layer design effects such as HARQ, scheduling, and QoS provisioning are still unavailable.

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

6G's demands for efficient resource use and stronger interference management motivate rethinking multiple-access designs. RSMA addresses this by combining common and private message streams, enabling partial interference decoding and partial noise treatment across interference conditions.

  • 6G requires more efficient wireless-resource use and stronger interference management to support higher throughput, ultra reliability, heterogeneous QoS, and massive connectivity.
  • OMA evolved by assigning orthogonal resources, while SDMA uses spatial processing to serve multiple users in one time-frequency resource.
  • RSMA splits each user message into common and private parts, then partially decodes interference while partially treating it as noise.
  • By tuning common-stream powers and contents, RSMA reduces to SDMA under weak interference or NOMA under strong interference, bridging the two schemes.
  • OMA has simple transceivers and avoids multi-user interference, but dedicating each resource to one user restricts simultaneous support and limits spectral efficiency.
  • SDMA reaches maximum underloaded-BC DoFs with perfect CSIT but is unsuitable for overloaded systems and degrades sharply with imperfect CSIT.
  • NOMA can improve spectral efficiency in severely overloaded scenarios, but multi-antenna NOMA with G = 1 achieves sum-DoF 1 versus SDMA's optimal min{M, K}.
  • SDMA and NOMA suit weak and strong interference, respectively, whereas switching between them does not work well for medium interference.

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

5G/6G demands efficient resource use and stronger interference management under diverse service requirements and imperfect CSIT. RSMA is presented as a flexible, comprehensive framework that addresses these needs and consolidates existing research.

  • Motivation: 5G services require high throughput, ultra-reliable low-latency transmission, massive connectivity, and high energy efficiency.eMBB targets Gbps-level rates, URLLC targets 10^-5 BLER within 1 ms, and mMTC requires high connection density and EE.
  • Motivation: Imperfect CSIT makes transmitter-side interference management difficult because inaccurate channels prevent reliable interference elimination.The paper identifies CSIT acquisition impairments as a major bottleneck in modern MIMO networks.
  • Motivation: The paper asks whether an MA scheme can robustly handle imperfect CSIT, adapt interference processing to interference levels, and encompass existing MA schemes.These questions motivate the search for a new PHY-layer strategy.
  • Paper scope: The survey targets a fragmented literature in which existing RSMA studies mainly focus on narrow scenarios and lack a pedagogical consensus-building review.The authors present consolidation as timely because of rapidly increasing RSMA research activity.
  • Paper scope: The paper provides a holistic tutorial covering RSMA principles, frameworks, literature, precoding, comparisons, applications, research challenges, and standardization issues.Its contributions include downlink, uplink, and multi-cell frameworks plus complexity comparisons and comparisons with OMA, SDMA, NOMA, and multicasting.
  • RSMA rationale: RSMA answers these questions by partially decoding interference and partially treating it as noise, rather than using only the two conventional extremes.Its design principle is framed as a flexible bridge between fully decoding interference and fully treating interference as noise.

B. Downlink RSMA

Downlink RSMA comprises multiple transmitter and receiver architectures built around message splitting, common and private streams, and flexible interference decoding. These schemes differ in message combination, precoding, decoding, and complexity.

  • Scheme taxonomy: Downlink RSMA includes 1-layer RS, hierarchical RS, generalized RS, RS-CMD, THPRS, and DPCRS schemes.The schemes are compared through their transmitter and receiver designs.
  • System architecture: A universal downlink architecture serves K single-antenna users with M antennas, covering SISO BCs when M = 1 and MISO BCs when M > 1.RSMA is formulated for K > 1 to manage co-channel interference in multi-user transmission.
  • Transmitter design: RSMA transmitters split each user message into sub-messages, combine selected sub-messages into messages, encode streams, and map them onto precoders.The number of splits and streams depends on the specific RSMA scheme.
  • Scheme taxonomy: Linearly precoded RSMA includes 1-layer RS, 2-layer HRS, generalized RS, and RS-CMD, while THPRS and DPCRS use nonlinear THP and DPC precoding.The taxonomy distinguishes schemes by their precoding approach.
  • Message structure: A common RSMA message carries parts of different users’ unicast messages for interference management, unlike a multicast message whose full content is wanted by multiple users.Both message types may be decoded by multiple users, but their intended content differs.
  • Receiver design: Receivers decode only selected common streams, treat remaining common streams as noise, and may use SIC, joint, or turbo detectors.Different channel decoders, including V-BLAST and Polar decoding, can be paired with SIC receivers.

3) 1-layer RS:

The section presents 1-layer RS as the practical building block of RSMA and contrasts it with hierarchical and generalized extensions. These designs increase message-layer flexibility at differing complexity levels.

  • 1-layer RS: 1-layer RS splits each user message into common and private sub-messages, creating one common stream and one private stream per user.The common sub-messages are combined and decoded by all users, while private sub-messages are independently encoded for their corresponding users.
  • 1-layer RS: Each user first decodes the common stream, removes it using SIC, decodes its private stream, and recombines the recovered parts into the original message.The common stream is decoded while private-stream interference is treated as noise; the private stream is then decoded after cancellation.
  • 1-layer RS: 1-layer RS supports low-complexity or optimization-based precoder design, including ZFBF, WSR maximization, worst-case rate maximization, EE maximization, and transmit-power minimization.The precoding matrix contains one common and K private precoders.
  • 2-layer HRS: 2-layer HRS groups users and adds inter-group and inner-group common streams before private-stream decoding.In the four-user example, disabling inner-group streams reduces HRS to 1-layer RS.
  • Generalized RS: Generalized RS uses L = 2^K−1 message splits and 2^K−1−1 SIC layers per user to create common streams for different user subsets.This design aims to improve achievable rate and QoS at the expense of higher transceiver complexity.

5) Generalized RS:

Generalized RS creates streams intended for every relevant user subset and decodes them through multiple SIC layers. RS-CMD instead uses independently encoded common streams without message combiners.

  • Generalized RS: Generalized RS assigns streams by order, where an l-order stream is decoded by l users in a selected subset.The transmitter linearly precodes the stream vectors associated with different user subsets.
  • Generalized RS: Each user decodes intended common streams from K-order down to 1-order private streams using 2^K−1−1 SIC layers.Within each order, users follow an order-specific decoding permutation.
  • Generalized RS: A three-user generalized RS example splits every message into four sub-messages and forms one 3-order, three 2-order, and three 1-order streams.Each user uses three SIC layers to decode its four intended streams.
  • RS-CMD: RS-CMD splits each user message into common and private parts and independently encodes the resulting 2K sub-messages without message combiners.Its common streams are decoded by all users in the considered configuration, while private streams are decoded only by their corresponding users.
  • RS-CMD: RS-CMD users decode K common streams in an order-specific K-layer SIC process before decoding their intended private streams.The common-stream decoding order may differ across users.

7) DPCRS:

This section surveys RSMA across uplink, multi-cell, and dirty-paper-coded settings, emphasizing flexible interference management and capacity-region coverage. It also notes receiver-complexity and application-study boundaries.

  • 7) DPCRS:: DPCRS combines linearly precoded common streams with non-linearly dirty-paper-coded private streams to address imperfect-CSIT losses.The surveyed variants are 1-DPCRS and M-DPCRS; their private streams use DPC while common streams remain linearly precoded.
  • C. Uplink RSMA: RSMA achieves every point of the two-user Gaussian MAC capacity region, including the line segment between SIC corner points, without time sharing.It uses message splitting, successive cancellation, and suitable rate selection; time sharing otherwise incurs coordination and synchronization overhead.
  • C. Uplink RSMA: Avoiding time sharing in uplink RSMA requires decoding all 2K − 1 streams with 2K − 2 SIC layers, increasing receiver complexity.Power allocation and decoding order are optimized to support this construction.
  • C. Uplink RSMA: Uplink RSMA can support homogeneous services such as URLLC or mMTC and heterogeneous non-orthogonal multiplexing, but both applications require further study.The generic uplink model is described as adaptable to both service settings.
  • D. Multi-cell RSMA: In multi-cell networks, RSMA addresses both intra-cell and inter-cell interference and has been shown to enhance spectral efficiency in coordinated and cooperative deployments.Coordination does not require data sharing, whereas cooperation relies on data sharing; both normally require CSI sharing among base stations.

2) Cooperative transmission:

The paper traces RSMA from information-theoretic rate splitting to multi-antenna and practical communication-system designs. Across these settings, the surveyed results include capacity, DoF, and throughput gains, while broader deployment potential remains under study.

  • A. RSMA in Single-Antenna Networks: RS originated in interference channels by splitting each message into common and private parts, enabling partial interference decoding while treating the remainder as noise.The Han–Kobayashi construction exemplifies this principle and was shown within one bit of the two-user SISO IC capacity in a special case.
  • A. RSMA in Single-Antenna Networks: Uplink RSMA achieves all boundary points of the SISO MAC capacity region without time sharing or synchronization among users.Users split messages at transmitters and the base station applies successive cancellation.
  • A. RSMA in Single-Antenna Networks: Generalized RSMA includes SDMA and NOMA as subschemes and was compared with both NOMA and DPC for multi-antenna broadcast channels.The framework was introduced with the RSMA terminology for the multi-antenna BC.
  • A. RSMA in Single-Antenna Networks: Practical studies introduced finite constellations, finite-length polar codes, adaptive modulation and coding, OFDM, and realistic 3GPP channels.Link-level simulation reported significant throughput gains over existing multiple-access schemes.
  • B. RSMA in Multi-Antenna Networks: Across overloaded and underloaded MISO BCs with perfect and imperfect CSIT, summarized studies report sum-DoF and symmetric-DoF gains over SDMA and multi-antenna NOMA.These results motivate RSMA for future-generation multiple access and 6G-oriented research.
  • A. RSMA in Single-Antenna Networks: Information-theoretic studies established RSMA results spanning SISO channels, multi-antenna broadcast channels, imperfect CSIT, and generalized message demands.Reported results include inner bounds, capacity-region characterizations, DoF optimality, and constant-gap capacity results.

B. RSMA in Multi-Antenna Networks

In multi-antenna networks, RSMA research spans information-theoretic optimality and finite-SNR resource allocation. The surveyed framework enlarges the design space, while practical optimization remains computationally demanding.

  • B. RSMA in Multi-Antenna Networks: Information-theoretic RSMA work studies sum-DoF, DoF regions, fairness, generalized DoF, and capacity regions in multi-antenna networks.The capacity region remains open for the K-user MISO BC with partial CSIT, so much work focuses on DoF and GDoF.
  • B. RSMA in Multi-Antenna Networks: RSMA outperforms SDMA and NOMA in summarized DoF and max-min fairness results across underloaded and overloaded MISO BCs with perfect and imperfect CSIT.The survey attributes this to combining multi-antenna DoF exploitation with SIC receivers.
  • B. RSMA in Multi-Antenna Networks: RS-based schemes achieve the entire GDoF region in the two-user underloaded MISO BC with imperfect CSIT and capacity results within a constant gap in two-user MIMO BC settings.The cited capacity results include both sum capacity and the entire capacity region.
  • B. RSMA in Multi-Antenna Networks: Finite-SNR research optimizes precoders, powers, common rates, scheduling, and subcarrier allocation through joint optimization or low-complexity approximations.The two research lines target maximum achievable performance versus favorable performance–complexity trade-offs.
  • B. RSMA in Multi-Antenna Networks: Turning off common-stream rate allocation reduces the RSMA optimization problem to SDMA design, revealing that RSMA enlarges the optimization space.This reduction follows by setting Ck = 0 for every user.
  • B. RSMA in Multi-Antenna Networks: RSMA under imperfect CSIT produces a more general class of MIMO optimization problems, but existing suboptimal algorithms remain unfavorable for real-world applications because of computational complexity.Sequential convex approximation requires solving a series of convex subproblems with locally tight approximations.

2) Low-Complexity Resource Allocation:

Low-complexity RSMA resource allocation separates precoder-direction design from power optimization and uses tractable beamforming choices. These designs trade optimization simplicity against performance, especially for the common stream.

  • 2) Low-Complexity Resource Allocation:: Low-complexity RSMA allocation separates precoder directions from power allocation across the common and private streams.For 1-layer RS, there are K + 1 streams, and the common/private power split is especially important.
  • 2) Low-Complexity Resource Allocation:: Random common-stream precoding suffices for RSMA to achieve the entire DoF region but is inefficient for improving spectral efficiency.It is therefore mainly used in DoF analysis rather than finite-SNR performance optimization.
  • 2) Low-Complexity Resource Allocation:: Weighted matched beamforming designs the common-stream direction to maximize its achievable rate, with equal weighting commonly used for tractability.The cited optimal or asymptotically optimal weights depend on user weights, while equal weighted MBF sets each weight to 1.
  • 2) Low-Complexity Resource Allocation:: ZFBF steers each private-stream precoder orthogonally to the other users’ channels and is limited to underloaded MISO BCs with M ≥ K.Together with random common-stream precoding, it achieves the optimal DoF region for imperfect-CSIT MISO BCs.
  • 2) Low-Complexity Resource Allocation:: RZF/MMSE regularizes channel inversion, while regularized block diagonalization extends the approach to MIMO broadcast channels with multiple receive antennas.RZF reduces to ZFBF when the regularization parameter κ equals zero.

3) Resource Allocation by Machine Learning:

Machine learning has been applied to optimize RSMA power allocation when precoders are given, including deep reinforcement learning for transmit-stream power design.

  • Deep reinforcement learning designs power allocation for each transmit stream and yields a significant performance gain over SDMA under imperfect CSIT.

B. Resource Allocation for Multicarrier RSMA

Multicarrier RSMA requires joint resource allocation across precoding, power, common-rate allocation, scheduling, and subcarrier assignment. Existing studies show performance benefits but leave imperfect-CSIT allocation and low-complexity scheduling open.

  • Multicarrier RSMA jointly considers precoder design, power control, common-rate allocation, user scheduling, and subcarrier assignment.Each subcarrier may serve multiple users, while each user may occupy multiple subcarriers.
  • Mixed-integer non-linear multicarrier RSMA allocation is addressed by relaxing binary subcarrier indicators and applying semidefinite programming and SCA.
  • RSMA without subcarrier optimization can outperform SDMA with subcarrier optimization because it serves users with arbitrary channel conditions on each subcarrier.This can reduce transmitter overhead from subcarrier allocation and user scheduling.
  • Multicarrier RSMA studies with imperfect CSIT assume all legitimate users are served on every subcarrier, leaving allocation and scheduling under imperfect CSIT open.
  • Practical RSMA architectures extend abstract Gaussian-signaling models with finite constellations, finite-length polar codes, and adaptive modulation and coding.
  • A two-user RSMA transmitter forms one common and two private messages, with rates determining message splitting, combining, modulation, and coding.The example uses four consecutive symbols, and common-rate portions are allocated to the users.
  • At the receiver, users decode the common stream first, extract their common parts, remove the reconstructed stream, and then decode their private streams.Processing includes equalization, LLR calculation, deinterleaving, decoding, and SIC.

2) Receiver:

RSMA generalizes several multiple-access schemes through flexible message-to-stream mappings and partial interference decoding, but its broader flexibility can increase encoder and scheduler complexity.

  • One-layer RS reduces to SDMA when common-stream power is zero, while generalized RS contains other linearly precoded schemes except RS-CMD.
  • RSMA manages interference by partially decoding it through common streams and treating residual interference as noise, unlike schemes using more extreme strategies.
  • RSMA encompasses OMA, SDMA, NOMA, and multicasting as special cases through different common-stream and private-stream power or message assignments.
  • Message splitting increases RSMA encoder complexity, although one-layer RS adds only one stream relative to SDMA and NOMA in the K-user case.
  • SDMA has low encoder and receiver complexity but requires channel-aware user pairing and accurate CSIT for effective scheduling.User scheduling also introduces signaling overhead.
  • Multi-antenna NOMA, generalized RS, and RS-CMD have high decoding-order complexity because the transmitter must select among many possible orders.
  • One-layer RS has relatively low scheduler complexity among the compared schemes, and without user scheduling can outperform scheduled SDMA when CSIT is sufficiently inaccurate.

3) Receiver complexity:

Receiver complexity varies substantially across multiple-access schemes: OMA and SDMA avoid SIC, while one-layer RS and two-layer HRS limit SIC depth and broader schemes require more layers.

  • 3) Receiver complexity: OMA and SDMA have the lowest receiver complexity because neither requires SIC.
  • 3) Receiver complexity: One-layer RS and two-layer HRS require one and two SIC operations per user, respectively, independent of K.Their limited SIC depth reduces susceptibility to error propagation relative to schemes requiring user-dependent layers.
  • 3) Receiver complexity: Multi-antenna NOMA requires K−1 SIC layers per user for G = 1, while generalized RS and RS-CMD require 2^K−1 and K layers, respectively.
  • 1) DoF region: Under imperfect CSIT, one-layer RS achieves the optimal DoF region for the multi-antenna broadcast channel.
  • 1) DoF region: With M = 4, K = 6, and α = 0.5, one-layer RS achieves the highest sum-DoF and MMF-DoF with lower complexity than NOMA.
  • 1) DoF region: For M = 6 and α = 0.5, one-layer RS outperforms other schemes in sum-DoF and MMF-DoF for all tested K.
  • 1) DoF region: At an MMF-DoF threshold of 0.1, one-layer RS serves around 15 users with one SIC layer, versus at most 10 users for NOMA with nine layers.

2) Rate region:

Across rate-region, overloaded, operational-region, and fairness evaluations, RSMA flexibly combines common and private streams to manage interference and improve performance across diverse conditions.

  • Rate-region comparisons: RSMA bridges transmitter-side and receiver-side interference cancellation, outperforming SDMA, DPC, and NOMA under practical partial CSIT.Common streams dynamically determine how much interference is canceled at each side.
  • Overloaded deployments: 2-layer RS achieves the maximum DoF of 2 in an overloaded ten-user, two-antenna deployment, whereas MU–LP and multi-antenna NOMA achieve DoF 1.RS packs messages from eight users into the common stream and serves two users with private streams.
  • Operational regions: RSMA adapts its preferred operating mode to channel alignment, channel-strength disparity, and user-weighting conditions.It can reduce to NOMA for fairness-oriented aligned deployments and to SDMA in other channel conditions.
  • User fairness: 1-layer RS provides superior max-min rate performance to multi-antenna NOMA and SDMA with imperfect CSIT, while using one SIC layer per receiver.For K = 6, multi-antenna NOMA requires five SIC layers per user, whereas 1-layer RS uses one.
  • User fairness: Allocating substantial and increasing power to the common stream as SNR rises helps 1-layer RS improve user fairness under stronger interference.The common stream enables interference management through message splitting and power allocation.

5) LLS Performance:

Link-level simulations compare RSMA with existing multiple-access schemes across throughput, mobility, complexity, and broader system objectives. RSMA generally delivers higher throughput and robustness while retaining lower complexity in the reported settings.

  • LLS throughput: RSMA achieves significant throughput gains over SDMA and NOMA in link-level simulations with imperfect CSIT.The throughput trend follows Shannon bounds, and the gain over SDMA is larger than expected from the Shannon bound.
  • Mobility robustness: At the same 8 bps/Hz QoS constraint, RSMA supports 40 km/h user speed versus 5 km/h for SDMA.The comparison uses outdated CSIT caused by user mobility and feedback delay.
  • System-level properties: RSMA subsumes SDMA, NOMA, OMA, and multicasting, supporting flexible operation across underloaded and overloaded networks and diverse user deployments.Power allocation and message splitting allow RSMA to simplify to existing schemes when appropriate.
  • System-level properties: RSMA is reported to improve spectral efficiency, energy efficiency, coverage, latency, QoS, and user fairness across varied network conditions.Coverage extension can use cooperative relaying, UAVs, IRSs, or relay stations; RSMA also supports shorter blocklengths for some latency settings.
  • Complexity: RSMA combines lower scheduling and receiver complexity with sum-rate and MMF-DoF gains over multi-antenna NOMA.Multi-antenna NOMA requires joint grouping, decoding-order, and precoder optimization, while 1-layer RS uses one SIC layer.

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

This section surveys RSMA applications across emerging wireless technologies, reports performance gains in several settings, and identifies unresolved PHY-, cross-layer, and system-design challenges.

  • Emerging applications: RSMA studies span massive MIMO, IRS, visible light communication, UAV, joint communication and sensing, and satellite communications.The survey notes that these applications have generated substantial research interest, although RSMA remains in its infancy.
  • Performance evidence: RSMA achieves significant throughput gains over MU–MIMO and MIMO NOMA with realistic finite constellations, finite-length polar codes, and AMC.These practical-layer results are consistent with RSMA’s Shannon ergodic sum-rate gain.
  • Open theoretical problems: RSMA’s role in MIMO broadcast channels with imperfect CSIT remains bounded by the unresolved capacity region of the K-user MIMO BC.Existing results establish optimal DoF regions for the 2-user MIMO BC and K-user MISO BC under imperfect CSIT, but not the general K-user MIMO BC.
  • Open PHY-layer problems: Under imperfect CSIR, SIC can cause performance degradation even when the cancelled message is decoded correctly, motivating further RSMA receiver research.A deep-learning receiver was reported to mitigate interference-cancellation effects and improve modulated-symbol detection and error-rate performance.
  • Open MAC-layer problems: RSMA HARQ design has not yet been considered, despite HARQ’s role in recovering missing packets through ACK-NACK feedback and channel coding.The survey identifies HARQ as an important mechanism in modern wireless standards.
  • Emerging applications: RSMA reduces mmWave massive-MIMO training and feedback complexity through one-stage feedback while achieving sum-rate comparable to MU–LP with two-stage feedback.The comparison addresses the high complexity and overhead of conventional mmWave feedback procedures.
  • Emerging applications: In multigroup multicasting, RSMA provides DoF, max-min rate, energy-efficiency, and user-fairness gains through partial decoding and partial interference-as-noise treatment.The reported gains extend to imperfect CSIT and multibeam satellite communication settings.
  • Emerging applications: RSMA-aided non-orthogonal unicast and multicast improves spectral and energy efficiency over conventional NOUM without increasing receiver complexity.Its SIC layer manages both multicast–unicast and inter-unicast interference; dirty-paper-coded RS can also enlarge the achievable rate region under imperfect CSIT.

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

The survey examines RSMA in SWIPT, cooperative relaying, caching, UAV, and secure communications. Reported results include improved rate regions, power savings, caching gains, and sum-rate performance, alongside unresolved range, CSIT, trajectory, and secrecy questions.

  • SWIPT: In multi-antenna SWIPT, 1-layer RS outperforms MU–LP in information-receiver rate regions for a given harvested-power lower bound.Dedicated energy-carrying signals achieve nearly the same rate-region performance as configurations without them.
  • SWIPT: When information and energy receivers are co-located, RSMA can save more transmit power than SDMA.
  • SWIPT challenges: SWIPT remains constrained by limited transmission range to energy receivers, while energy receivers may also eavesdrop information at relatively high received strengths.The survey identifies coverage extension and information security as open issues in these scenarios.
  • Cooperative relaying: Cooperative rate-splitting uses two transmission phases and leverages users’ decoding of a common stream to combine RSMA with user relaying.Existing CRS studies assume perfect CSIT, leaving imperfect-CSIT design and joint precoding–relay-selection optimization open.
  • Wireless caching: RSMA-aided caching splits messages into cached and uncached parts, boosting DoF and coded-caching gain while reducing transmitter CSIT requirements.Further work is needed on placement and delivery optimization for spectral efficiency, energy efficiency, delay, and fronthaul constraints.
  • UAV communications: In UAV-aided communications, RSMA improves sum-rate in several applications and is further shown to improve sum-rate in satellite–aerial integrated networks.Its potential robustness to CSIT inaccuracy is relevant because UAV mobility makes perfect channel tracking impossible.
  • Security: RSMA secrecy remains challenging against internal legitimate-user eavesdroppers and pilot-contamination attacks in TDD massive MIMO.Further study is needed to determine whether RSMA’s robustness to pilot contamination translates into secrecy-rate benefits.

14) Massive machine-type communication:

RSMA is presented as an appealing multiple-access scheme for massive IoT and as a framework spanning diverse 6G applications and research directions. The paper surveys its architectures, comparisons, applications, challenges, and standardization considerations.

  • Massive machine-type communication: RSMA can support massive IoT with higher spectral efficiency, lower receiver complexity, and greater robustness to CSIT inaccuracy and network loads.Its information-theoretic power partitioning achieves the optimal DoF region in an overloaded MISO broadcast channel under heterogeneous CSIT conditions.
  • Massive machine-type communication: Whether RSMA can accommodate hybrid throughput, latency, and massive-connectivity requirements in 6G IoT remains unresolved.The paper identifies enhanced eMBB-URLLC-mMTC as a core 6G service whose suitability for RSMA is not yet known.
  • Visible-light communications: In MU-MIMO VLC, RSMA shows clear sum-rate gains over conventional multiple-access schemes even with highly correlated user channels.The paper motivates this application through RSMA’s interference-management capability and VLC’s high operating SNR.
  • Intelligent reconfigurable surface-aided communications: In IRS-aided communication, jointly optimizing precoders, message splits, and IRS phase shifts enables RSMA to achieve better energy-efficiency performance than NOMA and OFDMA.The cited study considers a MISO broadcast channel with multiple IRSs cooperatively assisting downlink transmission.
  • Intelligent reconfigurable surface-aided communications: Imperfect CSIT at IRSs and unstudied joint passive-active beamforming, user scheduling, and IRS allocation remain important RSMA research challenges.The absence of RF chains makes CSI acquisition difficult, while limited channel training can produce imperfect CSIT.
Loading 2201.03192v3…