5. Research Directions and Hardware Trade-offs

Read Trends as Design Questions

This note uses selected research directions rather than an exhaustive ranking of the latest papers. Ask what capability is added, which cost appears and whether a comparison keeps resources equal. The references span analytical work, architecture surveys and a specific around-the-corner experiment.

References for this section3

• A. Tishchenko et al., "The Emergence of Multi-Functional and Hybrid Reconfigurable Intelligent Surfaces for Integrated Sensing and Communications", IEEE Communications Surveys & Tutorials, early access, 2025.
• M. Rihan et al., "Active RIS-Aided ISAC Systems", IEEE Transactions on Communications, 2024.
• W. Lyu et al., "Hybrid NOMA Assisted Integrated Sensing and Communication via RIS", IEEE Transactions on Vehicular Technology, vol. 73, no. 5, pp. 7368-7373, 2024.

Active RIS: Signal Gain and Noise Gain

An active RIS supplies amplification; a passive RIS supplies reflection without amplifier gain. Amplification can overcome downstream receiver noise but also forwards upstream noise and adds its own noise.

For an illustrative single-stage model, let $S$ be signal power, $N_{\rm in}$ input noise, $N_{\rm amp}$ input-referred added noise, $N_{\rm rx}$ downstream receiver noise and $G$ linear power gain. Refer all powers to the same output reference plane before applying $G$:

$$\gamma=\frac{GS}{G(N_{\rm in}+N_{\rm amp})+N_{\rm rx}},\qquad \lim_{G\to\infty}\gamma=\frac{S}{N_{\rm in}+N_{\rm amp}}.$$

This is a generic amplifier-noise illustration, not a complete active-RIS radar equation. The lab holds signal and noise reference powers fixed to isolate the ceiling. A physical comparison also needs surface output power, saturation, stability and the actual multi-pass noise paths.

Does more active gain always improve SNR?

1. Formula and symbols

$$\gamma=\frac{GS}{G(N_{\rm in}+N_{\rm amp})+N_{\rm rx}}$$

• $G$: active linear power gain
• $S$: input signal power
• $N_{\rm in}$: incident noise
• $N_{\rm amp}$: input-referred added noise
• $N_{\rm rx}$: receiver noise

Convert dBm to linear power before adding noise terms.

Active power gain

10 dB

Input-referred added noise

-95 dBm

2. Work with the current inputs

$$\gamma=\frac{GS}{G(N_{\rm in}+N_{\rm amp})+N_{\rm rx}},\quad G=10^{10/10}$$

$$\begin{aligned}N_{\rm out}/\mathrm{mW}&=10^{10/10}(10^{-105/10}+10^{-95/10})\\&\quad+10^{-100/10}\\N_{\rm out,dBm}&=-84.46\end{aligned}$$

$$\gamma_{\rm dB}=(-100+10)-(-84.46)=-5.54\ \mathrm{dB}$$

3. Result

Active output SNR-5.54 dB

Passive reference SNR-1.19 dB

High-gain ceiling-5.41 dB

References for this section3

• M. Rihan et al., "Active RIS-Aided ISAC Systems", IEEE Transactions on Communications, 2024.
• A. Tishchenko et al., "The Emergence of Multi-Functional and Hybrid Reconfigurable Intelligent Surfaces for Integrated Sensing and Communications", IEEE Communications Surveys & Tutorials, early access, 2025.
• W. Lyu et al., "Hybrid NOMA Assisted Integrated Sensing and Communication via RIS", IEEE Transactions on Vehicular Technology, vol. 73, no. 5, pp. 7368-7373, 2024.

STAR-RIS: Two Sides, One Energy Budget

A simultaneous transmitting and reflecting surface can direct energy to either side. For a passive panel, let $T$ and $R$ be transmitted and reflected power fractions, $\eta$ total usable efficiency and $\rho$ the fraction of usable power assigned to transmission:

$$T=\eta\rho,\qquad R=\eta(1-\rho),\qquad T+R=\eta\leq1.$$

Field-amplitude coefficients have magnitudes $\sqrt T$ and $\sqrt R$. Do not confuse them with power fractions. If an echo traverses the panel twice by transmission, this part of its budget is $T^2$, not $T$. Independent amplitude and phase choices are an idealization; physical implementations can impose coupled transmission/reflection phases.

Can both sides receive all the incident power?

1. Formula and symbols

$$T=\eta\rho,\quad R=\eta(1-\rho),\quad T+R=\eta\leq1$$

$$P_{\rm two\ passes}/P_{\rm incident}=T^2$$

• $\eta$: total passive power efficiency
• $\rho$: split toward transmission
• $T$: Transmitted power fraction
• $R$: Reflected power fraction

T² is the normalized fraction after two independent passages through the transmitting side.

Usable power assigned to transmission

50%

Total usable power efficiency

80%

2. Work with the current inputs

$$T=\eta\rho=0.80\times0.50=0.400,\quad R=\eta(1-\rho)=0.400$$

$$T+R=0.80\leq1,\qquad T^2=0.1600$$

3. Result

Transmitted / reflected40.0% / 40.0%

Absorbed or otherwise lost20%

Two-pass transmission factor16.0%

References for this section3

• M. Rihan et al., "Active RIS-Aided ISAC Systems", IEEE Transactions on Communications, 2024.
• W. Lyu et al., "Hybrid NOMA Assisted Integrated Sensing and Communication via RIS", IEEE Transactions on Vehicular Technology, vol. 73, no. 5, pp. 7368-7373, 2024.
• A. Tishchenko et al., "The Emergence of Multi-Functional and Hybrid Reconfigurable Intelligent Surfaces for Integrated Sensing and Communications", IEEE Communications Surveys & Tutorials, early access, 2025.

Multifunctional and Hybrid Surfaces

Hybrid and multifunctional architectures add receiving, sensing or other functions to a reflecting surface. Compare what each proposed implementation actually measures, how it is calibrated and how much power and control signaling it needs. More functionality is a hardware design choice, not a free channel-estimation solution.

References for this section3

• A. Tishchenko et al., "The Emergence of Multi-Functional and Hybrid Reconfigurable Intelligent Surfaces for Integrated Sensing and Communications", IEEE Communications Surveys & Tutorials, early access, 2025.
• W. Lyu et al., "Hybrid NOMA Assisted Integrated Sensing and Communication via RIS", IEEE Transactions on Vehicular Technology, vol. 73, no. 5, pp. 7368-7373, 2024.
• M. Rihan et al., "Active RIS-Aided ISAC Systems", IEEE Transactions on Communications, 2024.

Learning, NOMA and Near-Field Designs

• Learning-based control: evaluate beam prediction on unseen geometry and movement; include training data, inference latency and scan overhead.
• Non-orthogonal multiple access (NOMA): users share resources with power allocation and successive interference cancellation. Account for decoding errors and user pairing.
• Near-field focusing: use spherical-wave distances when plane-wave assumptions fail. Report the physical aperture and target distance, not only element count.

The phase-focusing lab in the preceding note shows how to inspect a simplified spherical-wave sum before making a system-level claim.

References for this section3

• W. Lyu et al., "Hybrid NOMA Assisted Integrated Sensing and Communication via RIS", IEEE Transactions on Vehicular Technology, vol. 73, no. 5, pp. 7368-7373, 2024.
• A. Tishchenko et al., "The Emergence of Multi-Functional and Hybrid Reconfigurable Intelligent Surfaces for Integrated Sensing and Communications", IEEE Communications Surveys & Tutorials, early access, 2025.
• M. Rihan et al., "Active RIS-Aided ISAC Systems", IEEE Transactions on Communications, 2024.

Experimental Evidence

Yasmeen, Kundu and Ram report an around-the-corner RIS experiment at 5.5 GHz capturing walking micro-Doppler signatures under NLoS conditions. It is a concrete example of a particular geometry and hardware setup, not a universal coverage guarantee.

For a prototype, record surface response, calibration procedure, target trajectory, clutter, bandwidth and an equal-resource baseline. A simulation supports a model; an experiment tests how that model survives a specific environment.

References for this section3

• M. Rihan et al., "Active RIS-Aided ISAC Systems", IEEE Transactions on Communications, 2024.
• W. Lyu et al., "Hybrid NOMA Assisted Integrated Sensing and Communication via RIS", IEEE Transactions on Vehicular Technology, vol. 73, no. 5, pp. 7368-7373, 2024.
• A. Tishchenko et al., "The Emergence of Multi-Functional and Hybrid Reconfigurable Intelligent Surfaces for Integrated Sensing and Communications", IEEE Communications Surveys & Tutorials, early access, 2025.

Takeaway

Choose an architecture by its energy, noise and measurement budget. Next, include control overhead and mobility before calling the design deployable.

References for this section3

• M. Rihan et al., "Active RIS-Aided ISAC Systems", IEEE Transactions on Communications, 2024.
• W. Lyu et al., "Hybrid NOMA Assisted Integrated Sensing and Communication via RIS", IEEE Transactions on Vehicular Technology, vol. 73, no. 5, pp. 7368-7373, 2024.
• A. Tishchenko et al., "The Emergence of Multi-Functional and Hybrid Reconfigurable Intelligent Surfaces for Integrated Sensing and Communications", IEEE Communications Surveys & Tutorials, early access, 2025.