Source-linked AI summary
Smart and Secure Wireless Communications via Reflecting Intelligent Surfaces: A Short Survey
Abdullateef Almohamad, Anas M. Tahir, Ayman Al-Kababji, Haji M. Furqan, Tamer Khattab, Mazen O. Hasna, Huseyin Arslan
TL;DR
Future wireless security faces limits that conventional cryptography alone cannot address, while RIS-assisted PLS offers a way to involve propagation in the design. This survey reviews and classifies RIS-assisted PLS studies, then identifies practical challenges and research directions.
Problem
RIS-assisted PLS research faces practical design challenges, including imperfect CSI in practice and unavailable CSI for passive eavesdroppers.
Source: C. Effect of CSI availability and… · p. 8Method
The survey classifies RIS-assisted PLS studies by system model and reviews their objectives and optimization methodologies.
Source: §III · p. 4Results
The survey identifies CSI availability, RIS design and deployment, and computational complexity as key open issues in RIS-assisted PLS.
Source: §V · p. 12Takeaways & Limitations
RIS deployment and orientation are key to system secrecy, and flying RISs are a promising direction for optimizing 3D location, orientation, and trajectory.
Source: §V · p. 12Most surveyed techniques assume perfect CSI at the transmitter and/or RIS, although practical systems may have only imperfect CSI and passive-eavesdropper CSI is unavailable.
Source: C. Effect of CSI availability and… · p. 8
Abstract
from arXiv · showhide
With the emergence of the internet of things (IoT) technology, wireless connectivity should be more ubiquitous than ever. In fact, the availability of wireless connection everywhere comes with security threats that, unfortunately, cannot be handled by conventional cryptographic solutions alone, especially in heterogeneous and decentralized future wireless networks. In general, physical layer security (PLS) helps in bridging this gap by taking advantage of the fading propagation channel. Moreover, the adoption of reconfigurable intelligent surfaces (RIS) in wireless networks makes the PLS techniques more efficient by involving the channel into the design loop. In this paper, we conduct a comprehensive literature review on the RIS-assisted PLS for future wireless communications. We start by introducing the basic concepts of RISs and their different applications in wireless communication networks and the most common PLS performance metrics. Then, we focus on the review and classification of RIS-assisted PLS applications, exhibiting multiple scenarios, system models, objectives, and methodologies. In fact, most of the works in this field formulate an optimization problem to maximize the secrecy rate (SR) or secrecy capacity (SC) at a legitimate user by jointly optimizing the beamformer at the transmitter and the RIS's coefficients, while the differences are in the adopted methodology to optimally/sub-optimally approach the solution. We finalize this survey by presenting some insightful recommendations and suggesting open problems for future research extensions.
I. INTRODUCTION
Future wireless systems need stronger security and more controllable propagation, motivating RIS-assisted physical layer security (PLS). The survey reviews this area, its practical challenges, and open research directions.
I. INTRODUCTION
- Future wireless networks must support efficient, secure, reliable, and flexible communications for emerging applications.The introduction notes that conventional designs treat the propagation environment as uncontrollable, often viewing it as a negative factor to compensate.Source: §I · p. 1
- An RIS uses controllable passive elements to adjust incoming-wave amplitude and/or phase, shaping signal strength and direction at receivers.This controllability can support SNR, data-rate, security, and coverage improvements; one cited MISO study reports a squared power gain in the number of reflecting elements through active and passive beamforming.Source: §I · p. 1
- RIS-assisted PLS can tune signal strength toward legitimate users and away from eavesdroppers, including when both lie in the same direction,.The introduction identifies passive beamforming as an additional design degree of freedom and notes that reflected-signal noise is not amplified as with conventional relays.Source: §I · p. 2
- RIS-assisted PLS increases design complexity because transmitter beamforming must be jointly optimized with RIS coefficients using RIS-side CSI.CSI acquisition is challenging with many passive elements; under far-field propagation, RIS links also incur double path loss that must be addressed in the link budget or with more elements [14].Source: §I · p. 2
- The survey systematically reviews RIS-assisted PLS studies, highlights literature shortcomings, and identifies open research extensions.It focuses on information-theoretic secrecy rather than covert communications, and previews metrics, study categorization, recommendations, and open directions.Source: §I · p. 2
II. SECRECY PERFORMANCE METRICS
This section introduces common first- and second-order physical layer security metrics. It explains secrecy rate and secrecy capacity, including their channel-based interpretations.
II. SECRECY PERFORMANCE METRICS
- The section reviews commonly used physical layer security metrics, covering both first- and second-order measures.The supplied subsection material focuses on secrecy rate and secrecy capacity.Source: §II · p. 2
A. Secrecy Rate/Capacity (SR/SC)
- Secrecy rate is the securely delivered information rate, defined as the legitimate user's achievable rate minus the eavesdropper's rate, floored at zero.Positive secrecy rate can be obtained through transmitter and/or RIS beamforming that degrades the eavesdropper channel while improving the legitimate channel.Source: A. Secrecy Rate/Capacity (SR/SC) · pp. 2–3
- Secrecy capacity is the upper bound of secrecy rate; for a channel realization, it is expressed using the legitimate and eavesdropper channel capacities.The ergodic capacity is obtained by averaging the channel-realization expression according to available CSI.Source: A. Secrecy Rate/Capacity (SR/SC) · p. 3
B. Secrecy Outage Probability/Capacity (SOP/SOC)
Secrecy outage probability (SOP) measures the chance that instantaneous secrecy capacity falls below a target secrecy rate. Secrecy outage capacity (SOC) instead gives the maximum secrecy capacity for a specified outage threshold.
B. Secrecy Outage Probability/Capacity (SOP/SOC)
- SOP is the probability that instantaneous secrecy capacity falls below a target secrecy rate Rth.The definition can also distinguish outages involving unreliable legitimate transmission from those involving information leakage.Source: B. Secrecy Outage Probability/Capacity… · p. 3
- SOC is the maximum achievable secrecy capacity Cout that guarantees SOP below a threshold ϵ.This makes SOC an outage-constrained capacity metric.Source: B. Secrecy Outage Probability/Capacity… · p. 3
C. Average Secrecy Outage Rate/Duration (ASOR/ASOD)
ASOR and ASOD use second-order secrecy statistics to characterize threshold-crossing dynamics. ASOR measures the average rate of crossing a secrecy-capacity threshold, while ASOD measures the average time spent in outage.
C. Average Secrecy Outage Rate/Duration (ASOR/ASOD)
- ASOR measures the average rate at which secrecy capacity crosses a threshold Rth.It is denoted R(Rth).Source: C. Average Secrecy Outage… · p. 3
- ASOD measures, in seconds, the average duration the system remains in secrecy outage.At a given Rth, it is expressed in terms of SOP and ASOR.Source: C. Average Secrecy Outage… · pp. 3–4
D. Amount of Secrecy Loss (ASL)
The survey introduces ASL as a second-order-statistics measure of information leakage and classifies RIS-assisted PLS studies by system model and optimization methodology. Its SISO coverage spans single-link, vehicular, and indoor multiuser settings.
D. Amount of Secrecy Loss (ASL)
- ASL measures the amount of information leaked to the eavesdropper using second-order statistics of secrecy capacity.The expectation operator E{·} denotes statistical expectation.Source: D. Amount of Secrecy Loss (ASL) · p. 4
III. CATEGORIZING RECENT STUDIES ON WIRELESS RIS-REINFORCED SECRECY
- The survey classifies RIS-reinforced PLS studies by transmitter and receiver antenna counts, system model, and optimization methodology.The reviewed works and their assumptions, models, methods, and performance metrics are summarized in Tables I and II.Source: §III · p. 4
A. SISO System Model
- The basic SISO setup has single-antenna Alice, Bob, and Eve; Fig. 2 contrasts SISO and MIMO models.In, RIS phases optimized using legitimate-user CSI improve SR, but secrecy drops slightly when Eve also has an RIS LoS link.Source: A. SISO System Model · p. 4, Figure 2 · p. 4
- For VANETs, RIS-assisted SC improves with more reflecting elements and degrades as source–RIS distance increases, [30].The reviewed analysis uses two RIS placements and a double-Rayleigh model; a reliable LoS link dominates SC, while optimized phases can improve it, and NLoS links can achieve comparable SC to dominant LoS when RIS–Eve CSI is unknown.Source: A. SISO System Model · pp. 4–5
- In an indoor SISO system with multiple Bobs and Eves, the optimal TAaPSA strategy depends significantly on the number of Eves.With few Eves, it improves Bob rates while degrading Eve rates; with more Eves, it can focus on boosting Bob’s average rate.Source: A. SISO System Model · p. 5
B. MIMO System Model
The MIMO studies jointly optimize RIS phase shifts and transmitter variables to improve secrecy rate, using iterative methods for non-convex problems. Their reported tradeoffs include optimization complexity and the effects of phase quantization and competing RISs.
B. MIMO System Model
- In, alternating optimization monotonically converges for a MIMO NLoS design, while [34] reports higher SR with more RIS elements but a larger phase-shift optimization burden.The convergence iteration count depends on antenna numbers; [34] uses block coordinate descent aided by majorization minimization.Source: B. MIMO System Model · p. 5
- Three-bit quantized RIS phases yield acceptable SR with negligible loss versus continuous phases.The alternating-optimization approach can be computationally costly at large scale and is suboptimal for full-rank BS–RIS channels [36].Source: B. MIMO System Model · p. 5
- A legitimate RIS with L reflecting elements can safeguard communication against an eavesdropping RIS with more than 3L elements [37].Without a legitimate RIS, the studied system cannot achieve non-zero SR with artificial noise and precoding.Source: B. MIMO System Model · p. 5
C. MISO System Model
The MISO literature is commonly organized around single- or multiple-eavesdropper models, with joint transmitter beamforming and RIS control often used to optimize secrecy performance. The reviewed works vary in system assumptions, objectives, and solution methods, and the survey classifies those differences.
Single Bob and Eve
- Single-Bob, single-Eve studies commonly jointly optimize BS beamforming and RIS phase shifts to maximize secrecy rate, using methods such as AO-MM, BCD, and AO.AO-MM is reported as favorable for large-scale RISs, while BCD is superior for small-scale RISs; a large RIS gives better secrecy-rate enhancement and energy efficiency than enlarging the transmit antenna array.Source: 1) Single Bob and Eve · p. 5
- For the single-Bob, single-Eve model, an AO-based solution also extends to multiple-antenna eavesdroppers, with secrecy rate degrading as Eve's antenna count increases.The cited work optimizes the BS transmit covariance and RIS phase-shift matrix.Source: 1) Single Bob and Eve · p. 6
Scope and classification
- The survey notes that distance-dependent simulations generally use planar 1D movements, leaving 2D/3D movement unaddressed, and it tabulates reviewed assumptions, models, methods, and metrics.The review also classifies studies by system model and optimization methodology.Source: 1) Single Bob and Eve · p. 6, 3) Multiple Bobs and Eves · p. 8
Other single-user settings
- Other single-user extensions include THz operation with sensing and reflection modes at the RIS, and RIS-based jamming that can outperform conventional active jamming when RIS-to-Bob distance is below 10m.The sensing mode uses active antennas with RF chains, which the survey notes is not power efficient given the RIS's many elements.Source: 1) Single Bob and Eve · pp. 6–7
Energy efficiency and SWIPT
- In energy-efficiency and SWIPT settings, RIS-assisted designs optimize secrecy alongside power or harvested energy, with SDR, PGD, AO, and SCA among the reported methods.For SWIPT, harvested power can double under a secrecy-rate constraint compared with the no-RIS case; the SDR method converges faster than PGD in the cited energy-consumption study.Source: 1) Single Bob and Eve · p. 6
CSI and learning-based methods
- When eavesdropper CSI is unavailable, joint beamforming and jamming can still secure communication by relaxing Bob's QoS threshold or increasing RIS element count; a DNN offers comparable secrecy-rate results with simpler, faster implementation.The DNN is proposed for real-time optimization; artificial noise is projected onto Bob's channel null space to jam Eve.Source: 1) Single Bob and Eve · p. 7
Multiple Bobs or Eves
- For one legitimate user facing multiple Eves, cooperative-jamming and transmitter artificial-noise designs jointly optimize RIS control and can maintain secrecy efficiently or require fewer RIS elements for a target secrecy rate.The cooperative-jamming scheme is reported as energy-efficient even with high jamming power; a separate RIS-assisted approach mitigates malicious jamming and improves transmission protection and system rate.Source: 2) Multiple Bobs or Eves · p. 7, 3) Multiple Bobs and Eves · p. 7
- With multiple legitimate users and one Eve, AO combined with SCA, SDR, and manifold optimization increases average sum secrecy rate with RIS support and BS artificial noise, but the analysis assumes NLoS.For multiple Bobs and Eves, reviewed approaches include two-RIS designs, minimum-secrecy-rate optimization, and DRL policies that improve average per-user secrecy rate and QoS satisfaction.Source: 2) Multiple Bobs or Eves · p. 7, 3) Multiple Bobs and Eves · pp. 7–8
IV. CHALLENGES, RECOMMENDATIONS, AND FUTURE RESEARCH DIRECTIONS
The survey identifies simultaneous BS transmission control and RIS reflection control as an efficient approach to confidentiality. It reports simulated secrecy-rate gains over conventional systems and sets out challenges and open research directions.
Challenges and future directions
- Jointly controlling BS transmission and RIS reflections is presented as an efficient solution for confidentiality in wireless communication.The survey frames this as a finding from its review of RIS-assisted PLS.Source: §IV · p. 8
- Several simulation results verify improved overall secrecy rate for RIS-assisted systems compared with conventional systems.Source: §IV · p. 8
- The survey identifies unresolved challenges and open directions for designing practical, efficient, and secure future RIS-assisted wireless systems.Source: §IV · p. 8
A. Effect of RIS physical design and deployment
The survey identifies RIS physical design and deployment as open PLS problems, including element configuration, collaboration, and mobility. It also highlights the difficulty of securing LoS links when an eavesdropper shares the legitimate user's direction.
A. Effect of RIS physical design and deployment
- RIS physical design and deployment remain underexplored for PLS, including RIS number, distribution, orientation, size, shape, element count, and element distribution.The survey also identifies optimal design, deployment, collaboration, and association as open challenges; stochastic-geometry and machine-learning approaches are suggested as options.Source: A. Effect of RIS physical design and… · p. 8
- Imperfect RIS elements, including discrete phase shifts and non-unit-modulus attenuating elements, need consideration in RIS-assisted PLS design.Source: A. Effect of RIS physical design and… · p. 8
B. PLS in LoS Environments
- LoS security is challenging when Eve lies in the legitimate user's direction; RIS can provide additional channel paths, but further study is needed, especially with imperfect or partial CSI.The survey notes that conventional beamforming and AN-based MIMO techniques can fail to secure these cases.Source: B. PLS in LoS Environments · p. 8
C. Effect of CSI availability and imperfect CSI
The survey finds that RIS-assisted PLS studies commonly assume perfect CSI, although RIS channel estimation and passive-eavesdropper CSI create practical constraints. It also calls for more secure-communication experiments and highlights higher-order secrecy metrics for assessing vulnerability over time.
C. Effect of CSI availability and imperfect CSI
- Most reviewed RIS-assisted PLS techniques assume perfect CSI, although RIS estimation is challenging and passive eavesdropper CSI is unavailable.The survey describes active RIS elements paired with compressive sensing or deep learning as estimation approaches, while noting assumptions and active-element trade-offs.Source: C. Effect of CSI availability and… · p. 8
- RIS-assisted security designs should account for imperfect or incomplete CSI and its availability at the transmitter to improve robustness across channel models.The survey notes that TDD channel reciprocity may no longer hold with an RIS.Source: C. Effect of CSI availability and… · p. 8
D. Practical Realization and Higher-Order Metrics Assessment
- Practical work on RIS-assisted secure communication remains scarce, so experiments are needed to assess RIS effects in real environments.The survey says current practical RIS work is insufficient to determine actual potential under practical conditions.Source: D. Practical Realization and… · pp. 8, 10
- Higher-order metrics complement SR, SC, and SOP: ASOR and ASOD quantify how often and how long secrecy-threshold vulnerability occurs, while ASL quantifies information lost during vulnerability durations.Source: D. Practical Realization and… · pp. 10–11
E. Flying RIS-Systems for PLS Enhancement
The survey identifies flying RISs as a flexible way to improve PLS, while noting that a single platform may be insufficient in challenging scenarios. It also highlights emerging application settings, unmet multi-objective and cross-layer design needs, and complexity-motivated use of learning methods.
E. Flying RIS-Systems for PLS Enhancement
- A flying RIS can be positioned and reoriented in 3D, with trajectory and phase shifts jointly adjusted to enhance PLS [25], while multiple flying RISs with active UAVs are motivated when one platform cannot meet secure-communication needs.Flying RISs can also act as mobile cooperative jammers with active UAVs or ground BSs.Source: E. Flying RIS-Systems for PLS… · p. 11
F. Integration of RISs with Emerging Technologies and Future Applications
- RIS-assisted PLS is proposed for emerging settings including millimeter-wave, massive MIMO, visible-light, IoT, THz, free-space-optics, full-duplex, NOMA [65], and VANET.Source: F. Integration of RISs with Emerging… · p. 11
- Future designs should jointly address security, reliability, latency, complexity, and throughput according to URLLC, eMBB, and mMTC quality-of-service requirements.RIS-assisted cross-layer security spanning physical, MAC, network, and application layers remains unstudied from the physical-layer perspective.Source: F. Integration of RISs with Emerging… · pp. 11–12
G. Optimization Problems to Enhance PLS
- RIS deployment increases design and analysis complexity, and only a few works [44], [54] have reported using ML-based approaches for RIS-assisted PLS.Data-driven ML, deep learning, and reinforcement learning are proposed as ways to support network flexibility and self-optimization.Source: G. Optimization Problems to Enhance PLS · p. 12
H. RIS-Assisted Secure SWIPT
The survey identifies secure RIS-assisted SWIPT and spectrum sensing as research directions, alongside a synthesis of PLS limitations, surveyed work, and recommendations. It emphasizes CSI, RIS deployment and design, and learning-based approaches to computational complexity.
H. RIS-Assisted Secure SWIPT
- Secure RIS-assisted SWIPT remains an open design problem because an untrusted energy-harvesting receiver may eavesdrop on information intended for legitimate users.The problem should be examined with discrete RIS phase shifts, phase-reflection-gain coupling, and imperfect or unavailable eavesdropper CSI.Source: H. RIS-Assisted Secure SWIPT · p. 12
I. RIS Enhanced PLS for Intelligent Spectrum Sensing and CR Realization
- RISs are proposed as a way to create a secure environment for reliable spectrum sensing and to support secure cognitive-radio communications in targeted primary or secondary networks.Source: §I · p. 12
V. CONCLUSIONS
- The survey discusses PLS metrics and classifies RIS-assisted PLS works by system model and methodology, noting that PLS has limits when legitimate and eavesdropping channels are highly correlated.RISs are presented as a possible response through propagation-channel manipulation that adds degrees of freedom.Source: §V · p. 12
- The survey recommends attention to CSI availability, RIS design and deployment, and ML-based approaches to computational complexity; RIS placement and orientation matter for secrecy.Flying RISs are identified as a promising direction because optimizing 3D location, orientation, and trajectory may boost secrecy and energy efficiency.Source: §V · p. 12