PaperlayerView exact revision on arXiv

Source-linked AI summary

Intelligent Reflecting Surface Aided Wireless Communications: A Tutorial

Qingqing Wu, Shuowen Zhang, Beixiong Zheng, Changsheng You, Rui Zhang

arXiv:2007.02759v2cs.ITcs.LGcs.NI

TL;DR

The paper addresses how IRS can be integrated into wireless networks despite challenges in reflection optimization, channel estimation, and deployment. It provides a tutorial synthesis of IRS models, architectures, constraints, applications, and communication-design approaches, concluding that IRS-aided systems represent a paradigm shift toward networks combining active and passive components.

  • Problem

    Future wireless networks require sustainable capacity growth with low cost, complexity, and energy consumption, while IRS integration introduces challenges including reflection optimization and accurate channel estimation.

  • Method

    The paper provides a comprehensive tutorial covering IRS fundamentals, reflection and channel models, hardware architecture, practical constraints, applications, and approaches to major communication challenges.

  • Results

    The tutorial shows that IRS-aided wireless communication produces a fundamental shift from wireless designs with active components only toward systems combining active and passive components.

  • Takeaways & Limitations

    IRS is presented as a promising technology for engineering wireless propagation and supporting cost-effective sustainable capacity growth through hybrid wireless networks.

Abstract

from arXiv · show

Intelligent reflecting surface (IRS) is an enabling technology to engineer the radio signal prorogation in wireless networks. By smartly tuning the signal reflection via a large number of low-cost passive reflecting elements, IRS is capable of dynamically altering wireless channels to enhance the communication performance. It is thus expected that the new IRS-aided hybrid wireless network comprising both active and passive components will be highly promising to achieve a sustainable capacity growth cost-effectively in the future. Despite its great potential, IRS faces new challenges to be efficiently integrated into wireless networks, such as reflection optimization, channel estimation, and deployment from communication design perspectives. In this paper, we provide a tutorial overview of IRS-aided wireless communication to address the above issues, and elaborate its reflection and channel models, hardware architecture and practical constraints, as well as various appealing applications in wireless networks. Moreover, we highlight important directions worthy of further investigation in future work.

I. INTRODUCTION

Future wireless networks face stringent 6G requirements that existing active-node, massive-antenna, and higher-frequency trends may not meet cost-effectively. IRS is proposed as a reconfigurable, predominantly passive means to reshape propagation and support hybrid networks with active and passive components.

  • 6G targets ultra-high data rate, energy efficiency, global coverage, connectivity, reliability, and low latency beyond existing 5G-oriented technology trends.
  • Conventional expansion through more active nodes, antennas, and higher-frequency bands increases energy, hardware, processing, deployment, backhaul, maintenance, and interference costs.
  • IRS uses passive reflecting elements to reconfigure signal propagation, offering a new means to address channel fading and interference while requiring no transmit RF chains.
  • IRS-aided MIMO can use substantially fewer active antennas while exploiting a large IRS aperture to create fine-grained reflected beams without compromising users’ QoS.
  • IRS integration shifts heterogeneous active-only networks toward hybrid architectures whose passive components can be densely deployed at lower cost and with reduced coordination needs.
  • Coordinated IRS reflections are envisioned for coverage extension, cell-edge signal enhancement and interference suppression, long-distance loss compensation, and Doppler mitigation.

C. What’s New?

This tutorial presents IRS as a reconfigurable propagation technology and systematically addresses its communication-design challenges, including reflection optimization, channel estimation, and deployment.

  • What’s New?: IRS extends reconfigurable metasurfaces toward wireless communication by digitally controlling electromagnetic-wave reflections.The paper contrasts this role with applications such as invisibility cloaking, imaging, radar sensing, and holograms.
  • What’s New?: IRS-aided networks require jointly designing passive reflections and transmissions to focus signals, cancel interference, and optimize end-to-end communications.The reflection design must support users whether or not an IRS is nearby.
  • What’s New?: Channel-state-information acquisition is practically difficult because IRSs generally lack RF chains and contain many reflecting elements with associated coefficients.The resulting CSI is essential for reflection optimization.
  • What’s New?: IRS deployment requires renewed study because passive-reflection architectures and operating mechanisms differ from active BSs/APs and relays.The paper identifies deployment at both link and network levels as a dedicated design issue.
  • What’s New?: The paper offers a tutorial treatment covering IRS fundamentals, reflection optimization, channel estimation, deployment, practical constraints, and related applications.Its organization spans signal and channel models, hardware architecture, system setups from single-user to multi-cell, and broader IRS topics.

B. IRS Architecture, Hardware, and Practical Constraints

IRS combines reconfigurable reflecting elements, layered hardware, and smart control to adapt wireless signal reflections in real time. Practical designs offer fast electronic tuning and amplitude/phase flexibility, but independently controlling both remains hardware-intensive.

  • IRS architecture: IRS uses a layered structure in which tunable metallic patches manipulate incident signals, a copper plate reduces leakage, and a control board tunes reflection amplitude or phase.A smart controller coordinates the elements and communicates with network components through wired or wireless control links.
  • Control and sensing: IRS controllers can use attached FPGAs, backhaul or control links, and dedicated sensors to learn the radio environment and design reflection coefficients.The architecture therefore connects physical reflecting elements with network coordination and environmental sensing.
  • Reconfiguration approaches: IRS reflection adaptation can use mechanical actuation, functional materials, or electronic devices such as PIN diodes, FETs, and MEMS switches.Electronic control is widely adopted because of its fast response, low reflection loss, relatively low energy consumption, and hardware cost.
  • Electronic tuning: PIN-diode biasing switches an element between ON and OFF states, producing a phase-shift difference of π in the incident signal.The cited switching frequency reaches 5 MHz, corresponding to a 0.2 µs switching time, smaller than typical millisecond-scale channel coherence times.
  • Practical constraints: Although ideal models independently and continuously tune amplitude and phase, practical independent control requires more sophisticated hardware designs.Wireless channels vary with transmitter, receiver, and surrounding-object mobility, requiring real-time tunable IRS responses based on channel variation.

1) Discrete reflection amplitude and phase shift:

Practical IRS elements use discrete amplitude and phase-shift levels to reduce hardware cost and complexity, but quantization coarsens control and complicates optimization. Moreover, amplitude and phase are physically coupled, so their joint design must balance reflected signal strength and phase alignment.

  • Discrete reflection amplitude and phase shift: Discrete amplitude and phase-shift levels require fewer control bits and can reduce IRS hardware cost and design complexity.Examples include two-level amplitude control and two-level phase-shift control.
  • Discrete reflection amplitude and phase shift: With bβ and bθ control bits, the corresponding amplitude and phase-shift levels are Kβ = 2^bβ and Kθ = 2^bθ.The discrete levels are represented by ordered amplitude and phase-shift sets.
  • Discrete reflection amplitude and phase shift: Quantized reflection control provides coarser amplitude and phase adjustment than continuous models, degrading communication performance and making optimization more difficult.The resulting optimization variables are discrete rather than continuous.
  • Discrete reflection amplitude and phase shift: Phase-shift control costs more than amplitude control but achieves better passive beamforming performance at the same number of bits or discrete levels.Higher-resolution phase control also increases implementation complexity, including additional PIN diodes and controller pins.
  • Coupled reflection amplitude and phase shift: Practical reflecting elements nonlinearly couple reflection amplitude with phase shift, with minimum amplitude near zero phase and maximum amplitude near ±π.The maximum reflection amplitude asymptotically approaches one, and the model agrees with reported experimental results.
  • Coupled reflection amplitude and phase shift: This coupling requires balancing reflected amplitude and phase so signals from all IRS elements combine with maximum received power or SNR.Near zero phase, energy dissipation is highest and reflection amplitude is lowest; near ±π, energy loss is minimized and amplitude is highest.
  • Practical energy consumption: IRS elements and controllers consume power in practice, although their consumption is substantially lower than that of active relays.Reported examples include about 0.33 mW or 50 µW per element and about 0.72 W for a 256-element controller.

C. Other Related Work and Future Direction

IRS reflection optimization must account for hardware and channel-model limitations, including mutual coupling, angle-dependent responses, possible channel nonreciprocity, and frequency-dependent phase shifts. The tutorial studies passive beamforming across diverse system setups and derives SISO design principles under idealized assumptions.

  • Reflection modeling and hardware constraints: Mutual coupling among closely spaced IRS elements can invalidate the linear channel model and require more complex nonlinear modeling.The tutorial identifies decoupling and isolation techniques as future directions.
  • Reflection modeling and hardware constraints: IRS reflection coefficients, particularly phase shifts, can depend strongly on signal incident angle, complicating optimization in multipath propagation.Each signal path may arrive at the IRS with a different angle of arrival.
  • Channel estimation limitations: When IRS is involved, channel reciprocity in TDD systems may no longer hold, so uplink-training-based channel estimation may not apply to downlink communication.The tutorial calls for further study of this effect.
  • Broadband modeling limitations: The constant-over-bandwidth phase-shift assumption is suitable for narrowband systems with B ≪ fc but can become inaccurate when broadband B is comparable to fc.Frequency-dependent delay creates phase drift and can cause phase errors in OFDM signals, requiring hardware or signal-processing compensation.
  • Passive reflection optimization: Passive reflection optimization is examined from single-user to multi-user, multi-antenna, broadband, and multi-cell communication settings under perfect channel knowledge.A single-cell multi-user example contains an N-element IRS assisting downlink transmission from an AP or BS to K users.
  • SISO passive beamforming: For SISO links, optimal phase shifts align IRS-reflected signals with the direct signal to achieve coherent combining and maximize received power.If the direct link is negligible, a common phase shift can be chosen arbitrarily without changing the optimal value.
  • SISO passive beamforming: The optimal SISO reflection design depends on the cascaded AP–IRS–user channels rather than requiring the individual component channels separately.This property can simplify IRS channel-estimation design.

2. Assume i.i.d. Rayleigh fading

Under the considered channel models, IRS performance improves with the number of reflecting elements, with receive power scaling quadratically in N at large N. IRS can eventually approach or exceed active-array and relay systems, although equal-element comparisons ignore hardware-cost differences.

  • With sufficiently large N, user receive power scales as O(N^2), enabling transmit-power reduction by a factor of 1/N^2 without compromising receive SNR.The gain combines O(N) reflect beamforming and O(N) aperture gain.
  • The achievable rate of the IRS-aided SISO system increases about 2 bps/Hz by doubling N from 200 to 400.
  • Maintaining the same average receive power over a larger IRS coverage range requires increasing the number of reflecting elements, for example linearly with d^2.
  • Comparison with M-MIMO and MIMO relay: For small N, IRS performs worst because of product-distance path loss and insufficient passive beamforming gain, but its rate gaps versus M-MIMO and relays narrow as N increases.
  • Comparison with M-MIMO and MIMO relay: At large N, IRS can outperform HD MIMO relay because its beamforming gain scales as O(N^2) rather than O(N), while passive elements avoid transmit RF chains.
  • Comparison with M-MIMO and MIMO relay: Equal passive-versus-active-element comparisons may be unfair because IRS elements cost less, allowing more reflecting elements at a given hardware cost.

C. IRS-aided MIMO and OFDM Systems

IRS-aided MIMO reflection design jointly optimizes IRS phases and transmit covariance under a non-convex capacity objective. Alternating optimization provides locally optimal solutions and can improve channel rank, spatial multiplexing, and achievable rate.

  • IRS-aided MIMO optimization jointly designs the reflection matrix and transmit covariance because capacity is a non-concave log-determinant function of both.
  • An alternating-optimization algorithm iteratively updates one IRS phase shift or the transmit covariance matrix while fixing the other variables.
  • The AO-based algorithm converges to at least a locally optimal solution with polynomial complexity in N, Mt, and Mr.
  • Proper IRS reflection design can improve channel power, condition number, and rank, especially when the direct channel is low-rank or rank-one.
  • Flexible amplitude design may improve capacity, but unit-amplitude reflection is used because jointly tuning amplitude and phase is practically difficult.
  • In the illustrated 4×4 MIMO setup, AO achieves substantially higher rate than both the system without IRS and random IRS phase shifts.

2) IRS-aided OFDM System:

In IRS-aided OFDM, frequency-flat passive phase shifts affect all subcarriers identically, so reflection design must jointly handle frequency-selective channels and power allocation. SCA and strongest-CIR methods improve rate, but frequency selectivity remains a fundamental limitation.

  • IRS phase shifts impact every OFDM subcarrier identically, because practical passive reflection is frequency-flat rather than frequency-selective.
  • The OFDM optimization jointly designs IRS phase shifts and transmit powers across Q subcarriers under a more difficult non-concave rate objective.
  • SCA-based optimization converges to a stationary point with polynomial complexity, while strongest-CIR maximization offers a lower-complexity alternative.
  • The SCA algorithm achieves higher rate than OFDM without IRS and random IRS phases, while strongest-CIR maximization performs closely to SCA.
  • A frequency-selective reflection upper bound substantially exceeds practical frequency-flat reflection, with the gap increasing as the number of subcarriers grows.
  • The lack of frequency-selective IRS reflection due to passive operation is identified as a fundamental limitation of IRS-aided OFDM.
  • IRS-aided MIMO-OFDM: Despite lacking frequency selectivity, properly designed IRS coefficients still improve MIMO-OFDM rate over conventional systems without IRS.

1) OMA versus NOMA:

IRS changes the trade-offs among multiple-access schemes and can improve multi-user interference management through controllable reflected paths. However, multi-user beamforming remains tightly coupled, and AO methods can become inefficient as QoS constraints increase.

  • OMA versus NOMA: For symmetric rate targets, NOMA requires more transmit power than TDMA, whereas TDMA can outperform NOMA for near-IRS users.
  • OMA versus NOMA: In the two-user setup, TDMA and NOMA require less minimum transmit power than FDMA, while NOMA is more robust to rate disparity.
  • OMA versus NOMA: Serving multiple users simultaneously makes IRS reflection design more challenging because the passive beamformer must accommodate all users’ channels.
  • OMA versus NOMA: Dynamic IRS beamforming over OFDMA time slots can exploit time selectivity and multi-user channel diversity, improving performance over fixed reflections.
  • SDMA: IRS can reduce channel correlation through controllable signal paths, improving all users’ SINR even when some users are not directly IRS-aided.
  • SDMA: Active and passive beamforming are closely coupled in multi-user systems, and AO may become inefficient or settle at undesired suboptimal solutions as QoS constraints increase.
  • Discrete reflection control: With discrete phase shifts, the asymptotic receive-power scaling remains O(N^2) independently of N, and 2-bit phase shifters achieve performance within 1 dB.
  • Discrete reflection control: One-bit phase control generally outperforms one-bit amplitude control because it uses both reflected signal components rather than simple ON/OFF reflection.

F. Other Related Work and Future Direction

Future IRS research must address scalable reflection design, broader network settings, non-ideal hardware, and channel acquisition with imperfect CSI. The tutorial also identifies channel-estimation overhead and practical deployment constraints as central open issues.

  • IRS reflection coefficients must be jointly designed with multiple access-point transmissions to enhance desired signals and mitigate intra-cell and inter-cell interference.
  • The ideal O(N^2) power scaling can persist with phase-shift-dependent non-uniform reflection amplitudes, but practical reflection constraints remain worth investigating.
  • More computationally efficient methods, including machine-learning-based reflection design, are needed for practically large IRSs.
  • Accurate IRS-reflected-link CSI is difficult because passive elements cannot transmit pilot signals, and broadband, FDD, and large-array settings increase acquisition demands.
  • IRS channel estimation involves many more coefficients than conventional systems, including K × NMu + MBN for IRS links and K×MBMu for direct links.
  • TDD can exploit uplink-downlink reciprocity to estimate either uplink or downlink coefficients, whereas FDD generally requires twice the coefficients because links are non-symmetric.

B. Semi-Passive IRS Channel Estimation

Semi-passive IRSs add sensing devices and alternate between sensing and reflection, enabling channel reconstruction from low-dimensional measurements. Their accuracy depends on sensing resources, while training and reflection design expose practical trade-offs.

  • A semi-passive protocol estimates direct channels, senses BS/users-to-IRS CSI, exchanges CSI with the controller, jointly designs active/passive beamforming, and then transmits data.
  • For semi-passive IRSs, reverse-link CSI is available through reciprocity in TDD but unavailable in FDD, making FDD channel estimation infeasible in this configuration.
  • Semi-passive estimation reconstructs high-dimensional channels G and Hr,k from low-dimensional sensor channels using spatial correlation.
  • Estimation accuracy is limited by sensor count, ADC resolution, and sensing time; more sensors, higher-resolution ADCs, and longer sensing can reduce different error sources.
  • Element grouping and training design create a trade-off between CSI accuracy, passive beamforming gain, training overhead, and data-transmission time.

D. Other Related Work and Future Direction

The tutorial identifies major gaps in IRS channel-estimation research, particularly for semi-passive versus passive cost-performance comparisons and for broadband, FDD, and discrete-phase settings.

  • Few studies address semi-passive IRS channel estimation, and its performance and cost relative to passive estimation remain unclear.
  • Most existing work considers narrow-band channels, leaving broadband IRS channel estimation in need of further investigation.
  • FDD channel estimation and discrete phase-shift designs remain practically important and challenging because most studies assume TDD and continuous phase shifts.

V. IRS DEPLOYMENT

IRS deployment differs from active-node placement because passive reflection suffers product-distance loss, while low cost permits dense deployment and weak mutual interference. The tutorial studies deployment by optimizing achievable rates across link- and network-level settings.

  • Deployment optimization must account for reflected-channel distributions, operational cost, user demand, space constraints, and propagation environment.
  • IRSs should be placed near the transmitter or receiver to reduce severe product-distance or double path-loss from passive reflection without amplification.
  • Compared with active nodes, IRSs can be deployed more densely because their substantially lower cost supports many reflecting elements.
  • Sufficient separation makes mutual interference between passive IRSs practically negligible, simplifying deployment design.
  • The tutorial characterizes maximum achievable rates by optimizing IRS placement in basic single-user and multi-user communication setups.
  • In a point-to-point link, receive SNR is maximized when the IRS is directly above the user or access point, not midway between them.

2) Single IRS versus Multiple Cooperative IRSs:

IRS deployment balances path loss against cooperative beamforming and spatial multiplexing. The paper compares single, cooperative, centralized, and distributed deployments, showing that benefits depend on channel conditions, user count, and reflecting-element allocation.

  • Single IRS versus Multiple Cooperative IRSs: Splitting reflecting elements across cooperative IRSs increases inter-IRS path loss but can unlock larger multiplicative passive beamforming gains.The deployment choice is therefore not straightforward because multiple IRSs introduce additional reflections while enabling cooperation.
  • Single IRS versus Multiple Cooperative IRSs: Under a rank-one inter-IRS channel and far-field spacing, two cooperative IRSs achieve passive beamforming gain of order O((N/2)^4).This is obtained by properly aligning the passive beamforming directions of the two IRSs.
  • Single IRS versus Multiple Cooperative IRSs: Two cooperative IRSs can exceed one single IRS in receive SNR when the total number of reflecting elements is sufficiently large.Their gain grows more significantly with N, in order O((N/2)^4) versus O(N^2), while the additional path loss remains fixed for given H.
  • IRS Deployment at the Network Level: IRS locations should balance LoS and NLoS propagation against passive beamforming and spatial multiplexing gains to maximize system capacity.Low-rank LoS channels limit multi-stream capacity, whereas comparable NLoS and LoS components can improve channel rank and conditioning.
  • IRS Deployment at the Network Level: For multi-user MACs, centralized deployment generally outperforms distributed deployment under TDMA, NOMA, and FDMA, especially for asymmetric user rates.Centralized deployment also improves the farther user's rate and helps mitigate the near-far issue; with K = 1, the two strategies are equivalent.

C. Other Related Work and Future Direction

The tutorial surveys IRS applications including multi-cell deployment, physical-layer security, wireless power transfer, and SWIPT, while identifying unresolved deployment, CSI, and training problems.

  • Deployment and future directions: Optimal IRS deployment and association remain unknown for general multi-cell networks, while massively deployed IRSs make optimization computationally formidable.The paper points to stochastic geometry and machine learning for system-level analysis and efficient deployment design.
  • Physical-layer security: IRS passive beamforming can increase legitimate-user rates, reduce eavesdropper rates, and jointly use artificial noise without compromising legitimate-user performance.These capabilities are presented as mechanisms for enhancing secrecy rates in physical-layer security systems.
  • Wireless power transfer: IRS-assisted WPT and SWIPT can enhance received power and charging efficiency, improving both rate and energy performance.The benefits depend critically on available CSI and on balancing channel-estimation training against downlink power transfer and energy harvesting.
  • Wireless power transfer: Limited channel coherence time creates a training trade-off: insufficient training reduces beamforming gain, whereas excessive training consumes energy and reduces harvesting time.IRS deployment choices, including single versus multiple and centralized versus distributed IRSs, also affect WPT/SWIPT performance.

C. UAV Communications

The paper discusses UAV-mounted IRSs, mmWave communications, and mobile edge computing as application directions. These systems expand coverage or improve links but introduce placement, tracking, beam-training, and scheduling challenges.

  • C. UAV Communications: Aerial IRSs mounted on UAVs can establish stronger LoS links, reduce blockage, provide 360° reflection, and serve geographically separated users.Their mobility allows sequential movement toward multiple users while exploiting short-range LoS channels and reducing reflected-link product distance.
  • C. UAV Communications: Multiple AIRSs require joint 3D placement, trajectories, passive beamforming, and user association, alongside low-complexity channel estimation over moving trajectories.These requirements are more challenging than for terrestrial IRSs because UAV-ground channels and positions vary in three dimensions.
  • C. UAV Communications: IRSs can create virtual LoS channels in mmWave systems, mitigating blockage and propagation loss while supporting high spectral and energy efficiencies.However, narrow pencil-like IRS beams make beam scanning costly, and straightforward successive training overhead scales linearly with the number of users.
  • C. UAV Communications: IRS passive beamforming can reshape MEC offloading loads across edge servers, improving computational resource utilization and reducing computation latency.For a single user, improved edge-server links enable computation-intensive offloading without high transmission energy consumption.
  • Conclusion: The tutorial concludes that IRS research is still in its infancy while covering fundamentals, technical challenges, applications, and future research directions.It presents IRS as a hybrid active-passive architecture and a resource for further work on unlocking its potential.
Loading 2007.02759v2…