1. What is a Reconfigurable Intelligent Surface (RIS)?

The Basic Idea

An RIS is a programmable surface that changes how an incoming radio wave leaves it.

• What it controls: reflection, scattering, focusing, and, in some designs, transmission.

• Physical idea: use part of the environment to shape a propagation path.

• What it does not do: a passive RIS does not create new signal energy.

• Ordinary path: walls and objects reflect, scatter, or block the base-station signal.

• RIS path: one reflection becomes adjustable, as the figure below shows.

• Smart radio environment: a wall, ceiling, or facade helps control the radio channel.

• Uses: coverage, communication, sensing, and localization.

• Base station: still transmits the original signal.

• Surface: adjusts its response to redirect that signal toward the user.

• Outcome: position, orientation, and phase settings determine whether the reflected path helps.

A programmed surface opens a path around the blockage

1. Formula and symbols

$$\Delta\phi=\frac{2\pi d}{\lambda}(\sin\theta_{\rm out}-\sin\theta_{\rm in})$$

• $d$: cell spacing
• $\lambda$: wavelength
• $\Delta \phi$: required adjacent phase step

angles are measured from the surface normal in the shown geometry. A plain mirror instead fixes equal incident/reflected angles.

Surface on the facade

Plain wallProgrammed RIS

2. Work with the current inputs

$$\theta_{\rm in}=55.2^\circ,\quad\theta_{\rm out}=34.7^\circ$$

$$\Delta\phi=180^\circ[\sin(34.7^\circ)-\sin(55.2^\circ)]=-45.2^\circ$$

Here d = λ/2. The programmed phase step applies in RIS mode; mirror mode uses equal input/output angles.

3. Result

Signal at the userVia RIS path

Angle in55°

Angle out35°chosen by the phases

References for this section3

• E. Basar et al., "Wireless Communications Through Reconfigurable Intelligent Surfaces", IEEE Access, vol. 7, pp. 116753-116773, 2019.
• Q. Wu and R. Zhang, "Towards Smart and Reconfigurable Environment: Intelligent Reflecting Surface Aided Wireless Network", IEEE Communications Magazine, vol. 58, no. 1, pp. 106-112, 2020.
• M. Di Renzo et al., "Smart Radio Environments Empowered by Reconfigurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead", IEEE Journal on Selected Areas in Communications, vol. 38, no. 11, pp. 2450-2525, 2020.

How RIS Differs from a Relay

• Relay: receives, processes, and retransmits; it needs active RF and processing hardware.

• Passive RIS: changes the incident field without decoding or regenerating it.

• Benefit: lower power use than a typical active relay.

• Limit: it cannot fully rebuild a weak or corrupted signal.

Feature	Relay	Passive RIS
Receives and retransmits	Yes	No, mainly reflects/scatters
RF chains	Usually yes	Usually no
Power consumption	Higher	Lower
Noise amplification	Possible	No active receiver noise, but double path loss exists
Main strength	Signal regeneration	Environment control

Mirror, RIS or relay: same spot, three answers

1. Formula and symbols

$$\gamma_{\rm DF}=\min(\gamma_1,\gamma_2)$$

$$R_{\rm DF}=\tfrac12B\log_2(1+\gamma_{\rm DF})$$

$$R_{\rm surface}=B\log_2(1+\gamma_{\rm surface})$$

• $\gamma _{1}$: Linear SNR of the first decode-forward relay hop
• $\gamma _{2}$: Linear SNR of the second decode-forward relay hop
• $B$: bandwidth

Half-duplex relay rate uses half the time

mirror/RIS use their modeled reflected-path SNR. All SNRs inside logarithms are linear.

User angle from the wall's normal

10°

BS → wall distance

20 m

Wall → user distance

15 m

Plate size (elements)

2561,0244,096

2. Work with the current inputs

	Mirror	RIS	Relay
Where the wave goes	Fixed: angle out = angle in	Programmable	Anywhere (re-transmits)
RF power at the device	None	None (only control electronics)	Amplifiers and radio chains
Noise added	None	None (passive)	Receiver noise; amplify-and-forward also amplifies it
Duplex	Full	Full	Usually half: listen, then talk
Extra delay	None	None	Processing delay (decode, re-encode)
Path loss	Double (two hops)	Double (two hops)	Two separate hops, each recovered
SNR here	6.0 dB	32.1 dB	35.1 dB
Rate here	231 Mbit/s	1068 Mbit/s	583 Mbit/s

$$\gamma_{\text{mirror}} = \gamma_{\text{plate}}\,\operatorname{sinc}^{2}\!\left(\tfrac{\pi D}{\lambda}(\sin\theta_u - \sin\theta_s)\right), \quad \gamma_{\text{RIS}} = \frac{\gamma_{\text{plate}}}{L_{\text{RIS}}}, \quad C_{\text{relay}} = \tfrac{1}{2}B\log_2\!\left(1 + \min(\gamma_1, \gamma_2)\right)$$

3. Result

Mirror231 Mbit/sSNR 6.0 dB

RIS1068 Mbit/sSNR 32.1 dB

Relay (DF)583 Mbit/sSNR 35.1 dB

Best hereRIS

References for this section3

• E. Basar et al., "Wireless Communications Through Reconfigurable Intelligent Surfaces", IEEE Access, vol. 7, pp. 116753-116773, 2019.
• Q. Wu and R. Zhang, "Towards Smart and Reconfigurable Environment: Intelligent Reflecting Surface Aided Wireless Network", IEEE Communications Magazine, vol. 58, no. 1, pp. 106-112, 2020.
• M. Di Renzo et al., "Smart Radio Environments Empowered by Reconfigurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead", IEEE Journal on Selected Areas in Communications, vol. 38, no. 11, pp. 2450-2525, 2020.

What is Inside an RIS?

An RIS contains many small unit cells. Each unit cell can impose a phase shift, amplitude change, polarization change, or transmission/reflection coefficient on an incident wave.

Typical hardware components include:

• Metallic patches or metasurface elements.
• PIN diodes, varactors, MEMS switches, or tunable materials.
• A biasing/control network.
• A controller that receives configuration commands from the network.

The surface works collectively. One unit cell alone does little. Hundreds or thousands of cells together can shape a wavefront.

Inside an RIS: one unit cell, layer by layer

1. Formula and symbols

$$d\approx\lambda/2=c/(2f_c),\qquad\phi\in\{0,\pi\}$$

• $d$: illustrative unit-cell pitch
• $\lambda$: wavelength
• $f_c$: carrier
• $c$: wave speed
• $\phi$: selected ideal 1-bit reflection phase

Physical cell dimensions and achievable phases are design dependent.

Layer

PatchDiodeSubstrateGroundBias & drivers

Diode state

Off → 0°On → 180°

2. Work with the current inputs

$$d\approx\frac{299792458}{2\times28\times10^9}\ \mathrm{m}\approx5.4\ \mathrm{mm},\quad\phi=0^\circ$$

PIN diode (or varactor). Bridges a small gap in the patch. A PIN diode is either on or off, which changes the patch's electrical size and flips the reflection phase by about 180°: one bit. A varactor instead tunes its capacitance smoothly for a wide, continuous phase range.

3. Result

SelectedPIN diode (or varactor)

Reflection phase0°design target, 1-bit cell

Cell size at 28 GHz≈ 5.4 mmhalf a wavelength

The controller: how a new phase pattern reaches the surface

1. Formula and symbols

$$T_{\rm update}=T_{\rm latency}+Nb/R_{\rm link}+Nb/(k f_{\rm clk})$$

• $N$: elements
• $b$: bits per element
• $R_{\rm link}$: link bit rate
• $k$: parallel chains
• $f_{\rm clk}$: chain clock
• $T_{\rm latency}$: fixed link latency

The two serialization times are added in this model.

Control link

Wired EthernetWi-FiCellular side link

Elements

1,0244,09616,384

Bits per element

1-bit2-bit3-bit

Parallel shift-register chains

16

Shift clock

50 MHz

2. Work with the current inputs

$$T_{\text{update}} = T_{\text{latency}} + \frac{N b}{R_{\text{link}}} + \frac{N b}{k\, f_{\text{clk}}} = 100.0\ \mu\mathrm{s} + \frac{4096}{1000\times10^{6}} + \frac{4096}{16\cdot 50\times10^{6}} = 109.2\ \mu\mathrm{s}$$

3. Result

Pattern size0.50 kB4,096 bits

Link time104.1 µs

Shift-in time5.1 µs

Total update109.2 µs9,156 patterns/s; coherence at 30 km/h ≈ 543.5 µs

References for this section3

• E. Basar et al., "Wireless Communications Through Reconfigurable Intelligent Surfaces", IEEE Access, vol. 7, pp. 116753-116773, 2019.
• Q. Wu and R. Zhang, "Towards Smart and Reconfigurable Environment: Intelligent Reflecting Surface Aided Wireless Network", IEEE Communications Magazine, vol. 58, no. 1, pp. 106-112, 2020.
• M. Di Renzo et al., "Smart Radio Environments Empowered by Reconfigurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead", IEEE Journal on Selected Areas in Communications, vol. 38, no. 11, pp. 2450-2525, 2020.

Why Phase Control Matters

• Aligned RIS phases at a user: reflected waves reinforce.
• Opposing phases at an interferer: a designed pattern may reduce unwanted energy there.
• Constraint: improving one direction can affect other users and sidelobes.

The simplified design goal is:

$$\boldsymbol{\phi}^{\star} = \arg\max_{\boldsymbol{\phi}} \Big| h_d + \sum_{n=1}^{N} g_n\, e^{j\phi_n} f_n \Big|^{2}$$

• $h_d$: direct BS-to-user channel.
• $f_n$: BS-to-element-$n$ channel.
• $g_n$: element-$n$-to-user channel.

For one user the answer is simple: each element cancels its own path phase, $\phi_n = \arg h_d - \arg(f_n g_n)$, so every reflected term lines up with the direct path.

This is why RIS is not just a mirror. A mirror follows the law of reflection. An RIS can implement anomalous reflection, focusing, splitting, beam steering, and in advanced designs simultaneous transmission and reflection.

Aligned reflections add up; random ones cancel

1. Formula and symbols

$$E=\sum_{n=1}^Ne^{j\phi_n},\quad q=|E|/N$$

$$L_{\rm relative}=20\log_{10}q$$

• $N$: equal-amplitude elements
• $\phi_n$: each reflected phase
• $E$: summed field
• $q$: amplitude fraction of perfect alignment
• $L_{\rm relative}$: relative received-power change in dB

Phase misalignment (± spread)

150°

2. Work with the current inputs

$$q=\frac{\sqrt{(2.651)^2+(-0.806)^2}}{12}=0.2309$$

$$L_{\rm relative}=20\log_{10}(0.2309)=-12.7\ \mathrm{dB}$$

The displayed dB calculation floors the amplitude fraction at 0.001 for numerical readability.

3. Result

Elements12

Sum vs perfect alignment23%amplitude

Received power penalty-12.7 dB

References for this section3

• E. Basar et al., "Wireless Communications Through Reconfigurable Intelligent Surfaces", IEEE Access, vol. 7, pp. 116753-116773, 2019.
• Q. Wu and R. Zhang, "Towards Smart and Reconfigurable Environment: Intelligent Reflecting Surface Aided Wireless Network", IEEE Communications Magazine, vol. 58, no. 1, pp. 106-112, 2020.
• M. Di Renzo et al., "Smart Radio Environments Empowered by Reconfigurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead", IEEE Journal on Selected Areas in Communications, vol. 38, no. 11, pp. 2450-2525, 2020.

Why RIS Matters for 6G

RIS is attractive for 6G because future networks are expected to use:

• Dense deployments.
• mmWave and sub-THz bands.
• Massive MIMO and large apertures.
• Sensing and localization.
• Energy-efficient infrastructure.
• AI-driven radio control.

High-frequency signals provide large bandwidth but are easily blocked. RIS can add controllable non-line-of-sight paths and improve coverage without installing a full active base station everywhere.

More elements, more coherent gain

1. Formula and symbols

$$G=N^2,\quad G_{\rm dB}=20\log_{10}N,\quad D\approx\sqrt N\,\lambda/2$$

• $N$: equal-strength, coherently aligned cells
• $G$: power gain over a single cell under the same per-cell model
• $D$: illustrative square-panel side at half-wavelength pitch

It is not an absolute end-to-end link gain.

Elements on the surface (N)

16642561,0244,096

2. Work with the current inputs

$$G_{\rm dB}=20\log_{10}(1024)=60.2\ \mathrm{dB}$$

$$D\approx\sqrt{1024}\times0.535\ \mathrm{cm}=17.1\ \mathrm{cm}$$

3. Result

Elements1,02432 × 32 grid

Ideal coherent gain60.2 dB

Panel side at 28 GHz≈ 17.1 cmhalf-wavelength spacing

References for this section3

• E. Basar et al., "Wireless Communications Through Reconfigurable Intelligent Surfaces", IEEE Access, vol. 7, pp. 116753-116773, 2019.
• Q. Wu and R. Zhang, "Towards Smart and Reconfigurable Environment: Intelligent Reflecting Surface Aided Wireless Network", IEEE Communications Magazine, vol. 58, no. 1, pp. 106-112, 2020.
• M. Di Renzo et al., "Smart Radio Environments Empowered by Reconfigurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead", IEEE Journal on Selected Areas in Communications, vol. 38, no. 11, pp. 2450-2525, 2020.

Correct Perspective: RIS is Not Magic

RIS redirects existing energy; a passive surface is not a perfect signal booster. Check these limits:

• Double path loss: BS-to-RIS and RIS-to-user attenuation both matter.
• Channel estimation difficulty: Passive surfaces cannot easily sense the channel themselves.
• Finite phase resolution: Real hardware may have 1-bit, 2-bit, or limited tuning states.
• Control overhead: Large surfaces require many control parameters.
• Hardware nonidealities: Mutual coupling, phase-dependent amplitude, and frequency selectivity can reduce ideal gains.

Link budget calculator: direct path vs RIS path

1. Formula and symbols

$$P_{r,\rm RIS}=\frac{P_tG_tG_r}{(4\pi)^2}\frac{\chi A^2}{d_1^2d_2^2}$$

$$\mathrm{SNR}=P_r/(kTB F)$$

• $P_t$: Transmitted power
• $P_r$: Received power
• $G_t$: Linear transmit antenna gain
• $G_r$: Linear receive antenna gain
• $\chi$: product of the incidence/reflection cosine factors and linear extra-loss factor
• $A$: physical area
• $d_{1}$: Distance from source to RIS
• $d_{2}$: Distance from RIS to user
• $k$: Boltzmann constant
• $T$: Noise temperature in kelvin
• $B$: Receiver bandwidth
• $F$: Linear receiver noise factor

Carrier frequency

3.5 GHz28 GHz100 GHz

RIS elements (N)

2561,0244,09616,384

Bandwidth

20 MHz100 MHz400 MHz

User position (east)

58 m

User position (south)

28 m

RIS position on the facade

36 m

Transmit power

30 dBm

BS antenna gain

20 dBi

User antenna gain

0 dBi

Building penetration loss

30 dB

RIS reflection + quantisation loss

2 dB

Receiver noise figure

7 dB

Interference at the user

-130 dBm

2. Work with the current inputs

Direct

-75.7 dBm

Via RIS

-70.0 dBm

Both (power sum)

-69.0 dBm

Scale -130 to -30 dBm; the orange tick marks the noise floor.

Link summary

Path

Via RISDirect

50.0 dBm leaves the BS → -70.0 dBm reaches the user → SINR 17.0 dB → 567 Mbit/s

1. BS transmits50.0 dBm30 dBm + 20 dBi antenna

   −93.4 dB39.7 m to the RIS

2. RIS reflects+65.0 dB+70.1 dB aperture −3.2 dB angles −2.0 dB loss

   −91.6 dB32.6 m to the user

3. User receives-70.0 dBm

   vs -87.0 dBmnoise + interference at the receiver

4. SINR17.0 dB

   × 100 MHzB · log₂(1 + SINR)

5. Data rate567 Mbit/sShannon limit

Budget item	Direct	Via RIS
Transmit power	30.0 dBm	30.0 dBm
BS antenna gain (beam on target)	+20.0	+20.0
User antenna gain	0.0	0.0
Path length	52.0 m	39.7 + 32.6 m
Free-space loss	−95.7	−93.4 −91.6
Building penetration	−30.0	clear
RIS aperture gain (capture + re-radiate)	—	+70.1
Oblique angles (θᵢ 49°, θᵣ 43°)	—	−3.2
RIS reflection + quantisation loss	—	−2.0
Received power	-75.7 dBm	-70.0 dBm
SINR	11.3 dB	17.0 dB
Shannon rate	385 Mbit/s	567 Mbit/s

Step-by-step calculation (updates with the controls)

1. 1Carrier wavelength

   Every loss and gain below depends on the wavelength, so we start there.

   $$\lambda = \frac{c}{f} = \frac{3\times10^{8}\ \text{m/s}}{28.0\times10^{9}\ \text{Hz}} = 10.71\ \text{mm}$$

2. 2Free-space path loss (Friis)

   A wave spreads over a growing sphere, so its power density falls with distance squared. In decibels:

   $$\mathrm{FSPL}(d) = 20\log_{10}\!\left(\frac{4\pi d}{\lambda}\right)\ \text{dB}$$
   $$\begin{aligned} \mathrm{FSPL}(d_0) &= 20\log_{10}\!\left(\tfrac{4\pi\cdot 52.0}{0.01071}\right) = 95.7\ \text{dB} && \text{(BS} \to \text{user)}\\ \mathrm{FSPL}(d_1) &= 20\log_{10}\!\left(\tfrac{4\pi\cdot 39.7}{0.01071}\right) = 93.4\ \text{dB} && \text{(BS} \to \text{RIS)}\\ \mathrm{FSPL}(d_2) &= 20\log_{10}\!\left(\tfrac{4\pi\cdot 32.6}{0.01071}\right) = 91.6\ \text{dB} && \text{(RIS} \to \text{user)} \end{aligned}$$

3. 3Direct path: received power

   In linear units the link is a product of gains and losses; in decibels the product becomes a sum. Start from the transmit power, add the antenna gains, subtract the losses:

   $$P_r^{\text{dir}} = \frac{P_t\,G_t\,G_r}{\mathrm{FSPL}(d_0)\,L_{\text{pen}}} \;\Longrightarrow\; P_r^{\text{dir}}[\text{dBm}] = P_t + G_t + G_r - \mathrm{FSPL}(d_0) - L_{\text{pen}}$$
   $$P_r^{\text{dir}} = 30.0 + 20.0 + 0.0 - 95.7 - 30.0 = \mathbf{-75.7\ dBm}$$

   The direct line crosses the building, so the penetration loss applies.

4. 4RIS aperture and its gain

   The surface is N elements spaced half a wavelength apart, so its physical area is

   $$A = N\left(\frac{\lambda}{2}\right)^{2} = 1024\times(5.35\ \text{mm})^{2} = 293.5\ \text{cm}^2$$

   An aperture of area A has antenna gain 4πA/λ². The RIS uses it twice: once to capture the incoming wave and once to re-radiate it toward the user. So in decibels:

   $$G_{\text{RIS}} = 20\log_{10}\!\left(\frac{4\pi A}{\lambda^{2}}\right) = 20\log_{10}\!\left(\frac{4\pi\times0.0293}{(0.01071)^{2}}\right) = 70.1\ \text{dB}$$

5. 5RIS path: received power

   Chain the two hops: free space to the surface, the surface gain, free space to the user. Tilted angles shrink the projected area, which adds a cosine factor for each side:

   $$\begin{aligned} P_r^{\text{RIS}} &= P_t G_t G_r \left(\frac{\lambda}{4\pi d_1}\right)^{2}\left(\frac{4\pi A}{\lambda^{2}}\right)^{2}\left(\frac{\lambda}{4\pi d_2}\right)^{2}\frac{\cos\theta_i\cos\theta_r}{L_{\text{RIS}}\,L_{\text{pen}}}\\ &= P_t G_t G_r\left(\frac{A}{4\pi d_1 d_2}\right)^{2}\frac{\cos\theta_i\cos\theta_r}{L_{\text{RIS}}\,L_{\text{pen}}} \end{aligned}$$

   Notice the λ terms cancel: for a fixed physical area, the RIS path does not get weaker at higher frequency, unlike the direct path. In decibels:

   $$\begin{aligned} P_r^{\text{RIS}}[\text{dBm}] = {}& P_t + G_t + G_r - \mathrm{FSPL}(d_1) - \mathrm{FSPL}(d_2)\\ &+ G_{\text{RIS}} + 10\log_{10}(\cos\theta_i\cos\theta_r) - L_{\text{RIS}} - L_{\text{pen}} \end{aligned}$$
   $$10\log_{10}(\cos49^\circ\cdot\cos43^\circ) = -3.2\ \text{dB}$$
   $$P_r^{\text{RIS}} = 30.0 + 20.0 + 0.0 - 93.4 - 91.6 + 70.1 - 3.2 - 2.0 + 0.0 = \mathbf{-70.0\ dBm}$$

6. 6Noise power at the receiver

   Thermal noise is kTB. At room temperature (290 K), kT is −174 dBm per hertz; the receiver adds its noise figure on top:

   $$N = kTB\,F \;\Longrightarrow\; N[\text{dBm}] = -174 + 10\log_{10}B + \mathrm{NF}$$
   $$N = -174 + 10\log_{10}(100\times10^{6}) + 7.0 = -174 + 80.0 + 7.0 = \mathbf{-87.0\ dBm}$$

7. 7SINR

   Signal, interference and noise powers add in linear units (milliwatts), not in dBm, so convert first with P[mW] = 10^(P[dBm]/10):

   $$\mathrm{SINR} = \frac{S}{I + N}, \qquad I = -130.0\ \text{dBm} = 1.00\times10^{-13}\ \text{mW}, \qquad N = 2.00\times10^{-9}\ \text{mW}$$
   $$\mathrm{SINR}_{\text{dir}} = \frac{10^{-75.7/10}}{1.00\times10^{-13} + 2.00\times10^{-9}} = \frac{2.68\times10^{-8}}{2.00\times10^{-9}} = 13.4 = \mathbf{11.3\ dB}$$
   $$\mathrm{SINR}_{\text{RIS}} = \frac{10^{-70.0/10}}{1.00\times10^{-13} + 2.00\times10^{-9}} = \frac{9.94\times10^{-8}}{2.00\times10^{-9}} = 49.8 = \mathbf{17.0\ dB}$$

   With no interference (I ≪ N), SINR equals the SNR.

8. 8Shannon capacity: the data-rate ceiling

   The highest error-free rate over a bandwidth B with this SINR. Real modulation and coding reach a fraction of it.

   $$C = B\log_{2}(1 + \mathrm{SINR})$$
   $$C_{\text{dir}} = 100\ \text{MHz}\times\log_{2}(1 + 13.4) = \mathbf{385\ Mbit/s}$$
   $$C_{\text{RIS}} = 100\ \text{MHz}\times\log_{2}(1 + 49.8) = \mathbf{567\ Mbit/s}$$

9. 9Both paths together

   If the user receives both paths with unrelated phases, their powers add (an RIS that also phase-aligns with the direct path could do up to 3 dB better when the two are equal):

   $$\begin{aligned} P_r^{\text{both}} &= 10\log_{10}\!\left(10^{-75.7/10} + 10^{-70.0/10}\right) = -69.0\ \text{dBm}\\ \mathrm{SINR}_{\text{both}} &= 18.0\ \text{dB}, \qquad C_{\text{both}} = \mathbf{601\ Mbit/s} \end{aligned}$$

3. Result

Direct received power-75.7 dBmSINR 11.3 dB

RIS received power-70.0 dBmSINR 17.0 dB

Noise floor-87.0 dBm

RIS advantage+5.7 dBRIS path is stronger

References for this section3

• E. Basar et al., "Wireless Communications Through Reconfigurable Intelligent Surfaces", IEEE Access, vol. 7, pp. 116753-116773, 2019.
• Q. Wu and R. Zhang, "Towards Smart and Reconfigurable Environment: Intelligent Reflecting Surface Aided Wireless Network", IEEE Communications Magazine, vol. 58, no. 1, pp. 106-112, 2020.
• M. Di Renzo et al., "Smart Radio Environments Empowered by Reconfigurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead", IEEE Journal on Selected Areas in Communications, vol. 38, no. 11, pp. 2450-2525, 2020.

Takeaway

• Core idea: program part of the radio environment rather than relying only on uncontrolled reflections.
• Where useful: blocked coverage, localization, interference management, or sensing when the RIS path geometry is favorable.

References for this section3

• E. Basar et al., "Wireless Communications Through Reconfigurable Intelligent Surfaces", IEEE Access, vol. 7, pp. 116753-116773, 2019.
• Q. Wu and R. Zhang, "Towards Smart and Reconfigurable Environment: Intelligent Reflecting Surface Aided Wireless Network", IEEE Communications Magazine, vol. 58, no. 1, pp. 106-112, 2020.
• M. Di Renzo et al., "Smart Radio Environments Empowered by Reconfigurable Intelligent Surfaces: How It Works, State of Research, and The Road Ahead", IEEE Journal on Selected Areas in Communications, vol. 38, no. 11, pp. 2450-2525, 2020.