Source-linked AI summary
Reconfigurable Intelligent Surfaces for Wireless Communications: Principles, Challenges, and Opportunities
Mohamed A. ElMossallamy, Hongliang Zhang, Lingyang Song, Karim G. Seddik, Zhu Han, Geoffrey Ye Li
TL;DR
Wireless propagation is traditionally treated as an uncontrollable source of unreliability, while higher-frequency channels can suffer weak received power, outages, and spatial sparsity. The paper presents a tutorial on RIS implementations, channel models, optimization, and channel shaping, reporting that RIS assistance can improve channel eigenstructure and simplify transceiver processing, while practical channel estimation remains difficult.
Problem
Wireless systems face weak received power, outage-prone propagation, and low-rank channels, while RIS-assisted designs must estimate many channel parameters without passive-element receive chains.
Method
The paper surveys reflectarray and metasurface RIS implementations, implementation-dependent channel models and path loss, RIS optimization, and numerical MIMO-channel shaping.
Results
RIS assistance can equalize channel eigenmodes, improving zero-forcing error performance and enabling simpler transceiver designs through orthogonal channel matrices.
Takeaways & Limitations
RISs can manipulate propagation to increase beamforming gain, improve channel conditioning, facilitate spatial multiplexing, and simplify transmitter and receiver processing.
Takeaways & Limitations
RIS channel gains principally rely on the ability to estimate channels despite passive surfaces and their massive number of channel parameters.
Abstract
from arXiv · showhide
Recently there has been a flurry of research on the use of reconfigurable intelligent surfaces (RIS) in wireless networks to create smart radio environments. In a smart radio environment, surfaces are capable of manipulating the propagation of incident electromagnetic waves in a programmable manner to actively alter the channel realization, which turns the wireless channel into a controllable system block that can be optimized to improve overall system performance. In this article, we provide a tutorial overview of reconfigurable intelligent surfaces (RIS) for wireless communications. We describe the working principles of reconfigurable intelligent surfaces (RIS) and elaborate on different candidate implementations using metasurfaces and reflectarrays. We discuss the channel models suitable for both implementations and examine the feasibility of obtaining accurate channel estimates. Furthermore, we discuss the aspects that differentiate RIS optimization from precoding for traditional MIMO arrays highlighting both the arising challenges and the potential opportunities associated with this emerging technology. Finally, we present numerical results to illustrate the power of an RIS in shaping the key properties of a MIMO channel.
I. INTRODUCTION
Wireless channels are traditionally treated as uncontrollable, while higher-frequency systems increasingly face weak received power, outages, and spatially sparse low-rank channels. RISs address these conditions by making propagation an optimizable part of the system and by shaping channel gain and eigenstructure.
- I. INTRODUCTION: RISs turn environmental reflection and scattering into system parameters that can be optimized rather than modeled only as uncontrollable stochastic effects.This reframes the propagation environment as a controllable component of wireless-system design.
- I. INTRODUCTION: Wireless-link rates depend on modulation order and spatial streams, so low channel gain or low-rank eigenstructure constrains achievable throughput.Cell-edge users may require lower-order modulation, whereas line-of-sight links can have high gain but support few spatial streams.
- I. INTRODUCTION: Higher-frequency bands can produce severe shadowing, outage events, and spatially sparse channels that limit received power and supported spatial streams.Millimeter-wave channels rely more on line-of-sight, first-order reflections, and scattering, while sparsity limits the number of spatial data streams.
- I. INTRODUCTION: RISs can increase channel gains or create paths around obstacles, and can emulate rich scattering to improve channel condition and spatial multiplexing.These functions target both low-received-power scenarios and line-of-sight spatial sparsity.
- I. INTRODUCTION: The tutorial covers RIS implementations, channel modeling, optimization challenges, and numerical demonstrations of MIMO-channel shaping.It also discusses implementation-dependent wave manipulation, link budgets, and differences from traditional MIMO precoding.
II. RECONFIGURABLE INTELLIGENT SURFACES
RISs are programmable passive surfaces that manipulate incident waves to improve wireless propagation, with reflectarrays and metasurfaces representing two principal implementation paths. Their collective operation can provide beamforming or alter channel rank and conditioning.
- II. RECONFIGURABLE INTELLIGENT SURFACES: An RIS is a dynamically reconfigurable passive surface that manipulates incident electromagnetic waves to change channel conditions.The definition applies regardless of the implementation technology.
- II. RECONFIGURABLE INTELLIGENT SURFACES: RISs can increase received power through beamforming or influence channel rank and condition number to facilitate spatial multiplexing.These functions allow surfaces to assist the ongoing transmission without communicating information of their own.
- A. Reflectarray-based Implementation: A reflectarray-based RIS electronically controls antenna terminations to backscatter and phase-shift incident signals.Each element has limited effect individually, so effective manipulation requires probably thousands of antenna elements.
- A. Reflectarray-based Implementation: RIS elements resemble backscatter tags, but RISs assist an ongoing transmission rather than communicating information from the reflector.RIS elements also operate collectively over a very large area to produce stronger effects on incident waves.
- A. Reflectarray-based Implementation: Reflectarray-based RISs provide centralized analog beamforming that can shift complexity from communication endpoints to the RIS and its controllers.Their elements have dimensions comparable to the wavelength and individually act as diffuse scatterers.
B. Metasurface-based Implementation
Metasurface-based RISs use densely packed, deeply subwavelength meta-atoms and reconfigurable tiles to offer flexible control of incident wavefronts. Their theoretical flexibility exceeds simpler reflectarrays, but practical gains require real-world validation.
- B. Metasurface-based Implementation: A metasurface contains many closely spaced, deeply subwavelength resonating structures whose size and density provide many wave-manipulation degrees of freedom.These structures are called pixels or meta-atoms, and both the atoms and their spacing are much smaller than the wavelength.
- B. Metasurface-based Implementation: Later metasurface designs use electrically, mechanically, or thermally tunable components to reconfigure electromagnetic behavior in real time.Electrically tunable designs are especially attractive because they can be cheaply manufactured using established semiconductor technology.
- B. Metasurface-based Implementation: A metasurface-based RIS comprises individually reconfigurable tiles much larger than the wavelength, each functioning similarly to a reflectarray.A tile can apply an approximately continuous amplitude/phase profile and reflect an incident wavefront in a chosen direction.
- B. Metasurface-based Implementation: Metasurface-based RISs theoretically offer greater flexibility than reflectarray-based RISs, but empirical validation is needed to determine whether that sophistication yields practical gains.Most empirical RIS studies cited in the passage use reflectarray-based implementations.
III. CHANNEL MODEL
RIS channel modeling incorporates both uncontrollable propagation and a controllable RIS contribution into the effective channel. Unlike traditional precoding, the RIS contribution is additive, environment-dependent, and only partially controllable, creating both optimization difficulty and greater channel-shaping potential.
- III. CHANNEL MODEL: Accurate RIS channel models are needed for analytical studies, simulations, utility evaluations, and comparisons with technologies such as relaying.The appropriate model may depend on whether the RIS uses reflectarrays or metasurfaces.
- III. CHANNEL MODEL: The channel model considers an M-antenna transmitter, an N-antenna receiver, and an L-element RIS under narrowband flat-fading assumptions.The received signal is expressed for this transmitter–receiver–RIS configuration.
- III. CHANNEL MODEL: H_env denotes the uncontrollable transmitter–receiver channel excluding RIS effects, while H_RIS denotes the controllable channel associated with the RIS.The effective RIS-augmented channel is the channel observed by the transceivers after combining the relevant components and gains.
- III. CHANNEL MODEL: Traditional precoding produces H_eff = H_envP multiplicatively, whereas RIS assistance contributes an additive, propagation-dependent term that is only partially controllable.This makes RIS optimization more challenging, while the additive effect can provide more control over the effective channel than traditional mixing and steering.
A. Dyadic Backscatter Channel Model
The dyadic backscatter model represents the RIS channel as cascaded transmitter-to-RIS and RIS-to-receiver channels connected by an interaction matrix. It captures configurable element responses but has important applicability limits for metasurface-based RISs.
- The model treats each RIS element as a regular omnidirectional antenna affected by fading and uses a dyadic backscatter channel.
- The RIS-mediated channel is formed from the transmitter-to-RIS channel, RIS interaction matrix, and RIS-to-receiver channel.F denotes the RIS-to-receiver channel, G the transmitter-to-RIS channel, and Q the RIS interaction matrix.
- Element responses are controlled through amplitude and phase parameters, with continuous or discrete phase shifts depending on implementation.The parameters β_i and θ_i are controlled by changing the complex antenna load.
- The cascaded channel has a product distribution and generally causes more detrimental fading than regular fading, although increasing L improves fading characteristics.
- This model may be inadequate for metasurface RISs because their elements are planar surfaces much larger than the wavelength rather than typical antennas.
B. Spatial Scattering Channel Model
The spatial scattering model represents sufficiently large RIS elements as reflectors creating distinct propagation paths, with controllable complex gains and potentially reflection angles. This implementation distinction also changes the predicted large-scale path loss and its validity conditions.
- B. Spatial Scattering Channel Model: The model treats each RIS element as a reflector creating a distinct propagation path when its dimensions are much larger than the wavelength.
- B. Spatial Scattering Channel Model: Each path combines a non-RIS path gain, a controllable RIS-element effect, and transmitter and receiver array steering vectors.α_ℓ represents path gain, q_ℓ the controllable element effect, and a_R and a_T the steering vectors.
- B. Spatial Scattering Channel Model: A metasurface can control reflection angles through the phase gradient applied across the surface, changing the receiver’s array response.
- C. Large-scale Path Loss: Accurate RIS path-loss modeling is critical for evaluating RIS links and can differentiate reflectarray and metasurface implementations.
- C. Large-scale Path Loss: Reflectarray elements diffusely scatter incident energy, whereas sufficiently large metasurface elements redirect incident wavefront sections at programmable reflection angles.
- C. Large-scale Path Loss: For metasurfaces, path loss scales with the sum of transmitter-RIS and RIS-receiver distances; for reflectarrays, it scales with their product.
- C. Large-scale Path Loss: The two implementation models can produce immense path-loss differences.
- C. Large-scale Path Loss: The metasurface sum-distance scaling is valid when the RIS is sufficiently large relative to endpoint distances, placing endpoints in its near field.For the cited 1.5m × 1.5m example, transceivers within 420m are in the near field and the scaling is well approximated by (5).
IV. RIS-ASSISTED OPTIMIZATION
RISs give system designers the ability to alter wireless channel realizations for different objectives. The section frames RIS configuration as distinct from traditional MIMO precoding.
- RISs can alter the wireless channel realization to achieve different objectives in various scenarios.
- The section reviews RIS-assisted optimization and discusses how it differs from precoding for traditional MIMO arrays.
- RIS optimization can target system objectives by configuring the propagation environment rather than only processing signals at conventional array endpoints.
A. State-of-the-Art Review
RIS optimization spans single-user gain maximization, multi-user interference management, and strategies for limited channel information. Practical performance depends strongly on acquiring or bypassing difficult RIS-related channel estimates.
- A. State-of-the-Art Review: In single-user single-antenna scenarios, RIS configuration simplifies to maximizing effective channel gain, received power, or minimizing transmit power for a target SNR.
- A. State-of-the-Art Review: Co-phasing all paths is optimal when relevant channels are known; continuous phases admit analytical or SDP solutions, while discrete phases can use quantization or greedy iteration.
- A. State-of-the-Art Review: Multi-user RIS configuration must account for interference, so maximizing channel gains alone is insufficient.
- B. Limited Channel State Information: Effective RIS optimization requires at least partial propagation-channel knowledge, yet RIS systems face many parameters and passive elements cannot directly send or process pilots.
- B. Limited Channel State Information: Receiver feedback can replace explicit channel estimation, using beam-searching or robust optimization to configure the RIS without measuring each element individually.
- B. Limited Channel State Information: A small subset of RIS elements can receive pilots and estimate partial channels, but those measurements alone are insufficiently accurate for beamforming or other channel manipulation.
- B. Limited Channel State Information: Prototype testbeds use receiver-side signal-strength or received-power feedback, often piggybacking RIS training on normal data transmissions without modifying the wireless standard.
- B. Limited Channel State Information: The RIS does not prevent ordinary transmitter-receiver channel estimation, but estimating the channels involving the RIS remains the difficult task.
C. Optimization Objectives
RISs enable wireless systems to pursue channel objectives beyond conventional endpoint precoding, including increasing channel rank, improving conditioning, diagonalizing interference, and mitigating Doppler spread. These capabilities come with substantial optimization and channel-information challenges, especially in dynamic multi-user, multi-RIS settings.
- Channel structure: Multiple RISs can create additional viable propagation paths, increasing channel rank and facilitating spatial multiplexing in millimeter-wave systems.This addresses cases where sparse millimeter-wave channels may offer only one viable propagation path.
- Channel structure: An RIS can keep the effective channel well conditioned, reducing noise enhancement and improving simple linear MIMO receiver performance.The motivation is that ill-conditioned channels limit spatial multiplexing and degrade MIMO detector performance.
- Interference management: An RIS can configure an interference-channel matrix toward a diagonal form, enabling spectrum sharing among multiple communicating pairs without interference.The diagonal terms represent desired links, while off-diagonal terms represent interference; empirical results exist, but theoretical results were not yet reported.
- Multiple objectives: Multiple RISs can be deployed to engineer propagation for diverse objectives, including beamforming, channel-rank enhancement, interference diagonalization, and increased channel coherence time.The RIS may focus power to a single-antenna receiver, support spatial multiplexing, enable spectrum sharing, or mitigate Doppler spread.
- Optimization challenges: RIS optimization is difficult even for a single RIS and single-user transmission because accurate channel information is hard to acquire.Multi-user, multi-RIS scenarios are expected to be less tractable, motivating machine-learning approaches using environmental sensing.
- Optimization challenges: Learning-based RIS control must operate within channel coherence times and adapt to changes caused by transceiver movement or environmental objects.The wireless channel is highly dynamic, so sensed propagation information can quickly become outdated.
V. NUMERICAL RESULTS
The numerical results show that RIS phase optimization can reshape MIMO channels toward orthogonality, improving conditioning and simplifying receiver and transmitter processing. The study also examines spectral-entropy optimization and its nonconvex computational challenge.
- RIS phase adjustment can orthogonalize the MIMO channel and achieve a near-unity condition number.This shifts complexity from transmitter precoding and receiver equalization to RIS controller optimization.
- The RIS-assisted channel is compared with a Rayleigh fading channel after normalizing average channel gain.This isolates gains associated with the effective channel’s improved eigenstructure.
- A near-unity condition number implies higher capacity because maximum capacity at a given channel gain occurs when singular values are equal.Orthogonal channels with equal singular values also support simpler processing.
- Spectral entropy of the effective channel is used to optimize the RIS configuration toward desirable singular-value properties.The objective reaches its maximum only when the channel is orthogonal.
- The spectral-entropy optimization is nonconvex, so sub-optimal heuristics may be necessary for reasonable computation.Gradient-based interior-point methods typically reach a local solution with SE = ln (min (M, N)) within a few tens of steps.
- For the simulated 4 × 4 link, the RIS has 100 elements, QPSK is used, and half the received power is assumed to arrive through the RIS.The simulation also assumes continuous phase shifts, while discrete shifts can remain adequate for large RISs.
- RIS assistance drastically improves zero-forcing error performance by eliminating noise enhancement when the channel is orthogonalized.In the RIS-assisted case, the effective channel is unitary, making zero-forcing equivalent to maximum-likelihood and matched-filter reception.
VI. FUTURE RESEARCH DIRECTIONS
The paper identifies future directions spanning RIS-enabled beamforming, experimental validation of path-loss scaling, and RF sensing and localization. These directions target practical deployment constraints and expanded sensing capabilities.
- Centralized Beamforming for IoT Devices: RISs can provide large beamforming gains to small, energy-constrained IoT devices that cannot support sufficiently large antenna arrays.Fixed RIS and base-station placement with limited scatterers can simplify beamforming optimization.
- Experimental Validation of Path Loss Scaling: Experimental work is needed to validate RIS path-loss scaling, especially whether metasurfaces outperform reflectarrays for the same physical size.The RIS is passive, so only co-phasing gains are available, making path loss central to practicality.
- RF Sensing and Localization: RIS-assisted RF sensing and localization can use a large aperture and programmable propagation environment to enhance sensing conditions.The channel can be altered for favorable sensing and then monitored with high accuracy.
VII. CONCLUSION
RIS-assisted wireless communications remain an early-stage field, but programmable propagation offers substantial potential across wireless scenarios. The paper surveys implementations, modeling, optimization challenges, and numerical evidence for shaping MIMO channels.
- RIS research remains in its infancy, with many practical aspects still requiring thorough investigation.
- The paper covers reflectarray- and metasurface-based RIS implementations, channel modeling, path loss, and RIS-assisted network optimization.
- Practical optimization techniques are essential for obtaining system performance gains from shaping wireless propagation.