PaperlayerView exact revision on arXiv

Source-linked AI summary

Reconfigurable Intelligent Surface: Power Consumption Modeling and Practical Measurement Validation

Jinghe Wang, Wankai Tang, Jing Cheng Liang, Lei Zhang, Jun Yan Dai, Xiao Li, Shi Jin, Qiang Cheng, Tie Jun Cui

arXiv:2211.00323v3cs.ITeess.SPmath.IT

TL;DR

The paper models RIS power by separating static hardware consumption from unit-cell consumption, then uses measurements to examine how power varies across RIS designs. Its results identify RF switch-based RISs as the most power-efficient of the three types and describe distinct tradeoffs among PIN- and varactor-diode designs.

Abstract

from arXiv · show

The reconfigurable intelligent surface (RIS) has received a lot of interest because of its capacity to reconfigure the wireless communication environment in a cost- and energy-efficient way. However, the realistic power consumption modeling and measurement validation of RIS has received far too little attention. Therefore, in this work, we model the power consumption of RIS and conduct measurement validations using various RISs to fill this vacancy. Firstly, we propose a practical power consumption model of RIS. The RIS hardware is divided into three basic parts: the FPGA control board, the drive circuits, and the RIS unit cells. The power consumption of the first two parts is modeled as $P_{\text {static}}$ and that of the last part is modeled as $P_{\text {units}}$. Expressions of $P_{\text {static}}$ and $P_{\text {units}}$ vary amongst different types of RISs. Secondly, we conduct measurements on various RISs to validate the proposed model. Five different RISs including the PIN diode, varactor diode, and RF switch types are measured, and measurement results validate the generality and applicability of the proposed power consumption model of RIS. Finally, we summarize the measurement results and discuss the approaches to achieve the low-power-consumption design of RIS-assisted wireless communication systems.

I. INTRODUCTION

Practical RIS power modeling and measurement validation have been limited, even though RIS hardware differs from relay hardware and its power budget needs separate consideration. This work proposes a model separating static and unit-cell power and validates it with measurements across RIS types.

I. INTRODUCTION

  • Earlier RIS power studies and measurements left the practical model and validation evidence limited; the proposed model is intended to support system-power characterization and energy-efficiency evaluation.The paper notes that prior models omitted control-circuit power or unit-cell coding effects, or did not examine RIS power in depth.Source: §I · p. 2
  • The model separates RIS power into static draw from the FPGA board and drive circuits and unit-cell draw, with expressions varying by RIS type.The paper proposes this structure as a practical basis for RIS power modeling.Source: §I · pp. 2–3
  • The FPGA board is treated as constant-power, while drive-circuit power depends on control-signal count and circuit characteristics.The model groups the FPGA board and drive circuits into Pstatic.Source: §I · pp. 2–3
  • PIN-diode unit-cell power depends on polarization, cell count, bit resolution, and working state; varactor-cell power is almost negligible, while RF-switch-cell power depends on polarization, cell count, and device characteristics.The paper gives type-specific expressions for Punits.Source: §I · p. 3
  • Measurements of five fabricated PIN-diode, varactor-diode, and RF-switch RISs validate the model's generality and applicability.The authors also discuss low-power design approaches for RIS-assisted systems.Source: §I · p. 3

II. POWER CONSUMPTION MODELING OF RIS

The section models RIS power by separating hardware into an FPGA control board, drive circuits, and unit cells, then assigning power models to these parts. It treats the board as constant-power and drive-circuit power as dependent on control signals and circuit characteristics.

II. POWER CONSUMPTION MODELING OF RIS

A. The Introduction of RIS hardware

  • RIS hardware is divided into the FPGA control board, drive circuits, and unit cells, and its operating power is split between static and unit-cell consumption.The FPGA programs cell coding states, the drive circuits actuate adjustable components, and the cells form beamforming patterns.Source: A. The Introduction of RIS hardware · p. 3

B. The Static Power Consumption Modeling

  • The FPGA control board's power is regarded as constant, while drive-circuit power is modeled as a separate component of static power.The static-power model also accounts for whether power is affected by the number of control signals.Source: B. The Static Power Consumption Modeling · p. 3

lPN

The static power model combines FPGA-board power and drive-circuit power. Drive-circuit power varies with the RIS design and the control signals and circuitry it requires.

lPN

  • Pstatic combines FPGA control-board power with the total power of the drive circuits.The board is modeled as constant power, while drive-circuit power varies with the number of control signals and the circuits’ own consumption.Source: lPN · p. 4
  • Drive-circuit power depends on component type, polarization mode, control degree of freedom, signal count, and each circuit’s power use.Examples include shift registers for PIN diodes and DACs or operational amplifiers for varactors.Source: lPN · p. 4
  • The model estimates the number of drive circuits from adjustable components, control grouping, and the signals each circuit generates.It rounds up to an integer and calculates horizontal and vertical polarization separately when their designs or control degrees of freedom differ.Source: lPN · p. 4
  • For PIN diodes and RF switches, the control-signal count follows unit-cell bit resolution; for varactors, it follows whether a cell is configurable.PIN diodes switch between binary states; a varactor can use different reverse-bias magnitudes for different states.Source: lPN · p. 4
  • For common RISs with uniform bit resolution and consistent polarization control, the general static-power expression can be simplified using polarization indicators.Iv and Ih indicate whether vertical and horizontal polarization, respectively, can be controlled.Source: lPN · p. 5, Table I · p. 5

C. The Power Consumption Modeling of RIS unit cells

Unit-cell power depends on the adjustable component and, for PIN-diode RISs, on the cell’s polarization, bit resolution, and coding state. Varactor-cell power is treated as negligible, while RF-switch-cell power is independent of coding state.

C. The Power Consumption Modeling of RIS unit cells

  • PIN-diode unit-cell power depends on polarization, bit resolution, and coding state because only bits encoded as 1 consume power.For a unit cell, power is proportional to the number of 1-coded bits; a 0-coded bit consumes no power.Source: C. The Power Consumption Modeling of… · p. 5
  • For varactor-diode-based RISs, unit-cell power consumption is regarded as 0 because operating varactor current is negligible, even when unit cells are large in number, supported by practical measurements.Source: C. The Power Consumption Modeling of… · p. 5
  • RF-switch unit-cell power depends on polarization, cell count, and per-cell switch power, but not coding state.The coding-state independence is supported by.Source: C. The Power Consumption Modeling of… · p. 5
  • The general model combines static and unit-cell power, with concise expressions available for common RIS designs.Equations (6) and (9) form the general model; equations (7) and (10) form the concise model.Source: C. The Power Consumption Modeling of… · p. 5

III. PRACTICAL MEASUREMENT AND VALIDATION

The validation uses a concise-model setup with uniform bit resolution and consistent control degree of freedom across polarizations. It measures six RISs spanning PIN-diode, varactor-diode, and RF-switch types.

III. PRACTICAL MEASUREMENT AND VALIDATION

  • The validation assumes all unit cells share bit resolution B and both polarization directions have consistent control degree of freedom Ng.The concise expressions being validated are (7) and (10).Source: §III · p. 5
  • Six fabricated RISs are used for validation: four PIN-diode-based, one varactor-diode-based, and one RF-switch-based.The passage says these measurements validate the concise power-consumption model.Source: §III · p. 5
  • The section presents practical measurement processes and results for RIS power consumption.Source: §III · p. 5

A. The Practical Measurement of the FPGA Control Board

In the measured PIN-diode RIS, the shift-register drive circuits consume 2.24 mW in total, compared with 4.8 W for the FPGA control board. The example uses four 8 × 8 sub-RISs to form a 16 × 16 RIS.

A. The Practical Measurement of the FPGA Control Board

  • The measured XC7K70T FPGA control board draws 4.8 W at 24 V and 0.2 A.The paper also notes that a Xilinx ZYNQ7100 consuming 1.5 W can suit prototypes with fewer cells or lower data-processing-speed requirements.Source: A. The Practical Measurement of the… · p. 5

B. PIN-diode-based RIS

1) Practical Measurement of Drive circuits:

  • 2.24 mW total drive-circuit power is measured for 32 shift registers, while the FPGA control board is measured at 4.8 W.The drive circuits use 0.07 mW each; the paper describes this as an energy-efficient, mW-level solution for hundreds or thousands of unit cells.Source: 1) Practical Measurement of Drive… · p. 6

2) Practical Measurement of RIS unit cells:

Measurements on three PIN-diode RIS designs show power rising with the number of enabled cells, in agreement with the proposed model. Results also distinguish static consumption and per-cell power across designs.

2) Practical Measurement of RIS unit cells:

  • 15.73 W at n = 0 and 12.56 mW per unit cell characterize the 1# RIS; dual-polarization power is about twice single-polarization power.The vertical- and horizontal-polarization curves are nearly coincident, and measurements in both modes match the model.Source: 2) Practical Measurement of RIS unit… · p. 7, Figure 7 · p. 7
  • 6.52 W at n = 0 is the 3# RIS static power, and total power rises linearly with the number of cells encoded as “1”.The authors also conducted a random-encoding measurement for model validation.Source: 2) Practical Measurement of RIS unit… · p. 7
  • 1.25 mW per PIN diode characterizes the 4# 2-bit RIS, whose measured column and cumulative power agree with theoretical results.For coding state “11”, measured column current agrees with the theoretical value obtained by adding the currents for “10” and “01”; Punits is linear in enabled bits.Source: 2) Practical Measurement of RIS unit… · p. 8

C. Varactor-diode-based RIS

The measured 5# varactor-diode RIS uses column control and eight columns of 16 cells each.

C. Varactor-diode-based RIS

  • The 5# varactor-diode RIS operates at 3.2 GHz, with eight columns of 16 cells controlled by reverse bias.The varactor bridges metal patches and serves as an adjustable device.Source: 1) Practical Measurement of Drive… · p. 8

1) Practical Measurement of Drive circuits:

The measured varactor-RIS drive system combines a DAC and an op-amp to provide control signals and amplify bias voltage. Its four drive circuits consume 1.72 W in total.

1) Practical Measurement of Drive circuits:

  • 1.72 W is the total drive-circuit power for four groups of 32 varactor cells, each using a DAC3484 and AD8021 op-amp.The DAC generates analog bias outputs, and the op-amp provides the amplification needed to reach higher bias voltage.Source: 1) Practical Measurement of Drive… · p. 9

2) Practical Measurement of RIS unit cells:

Across the measured RIS types, the authors identify RF switch-based designs as the most power-efficient, while highlighting trade-offs among unit-cell and drive-circuit consumption. Measurements show negligible varactor-cell current and characterize RF-switch control hardware.

2) Practical Measurement of RIS unit cells:

  • Varactor-diode currents were negligible under the tested reverse-bias voltages, so the authors model their unit-cell power as approximately zero.The tested biases were 0 V, −3 V through −8 V, and −20 V.Source: 2) Practical Measurement of RIS unit… · p. 9

D. RF Switch-based RIS

  • The measured RF switch RIS has 64 independently regulated cells and one FPGA driver circuit consuming 240 mW.The FPGA XC3S400AN serves as the RF switch driver.Source: D. RF Switch-based RIS · pp. 9–10

A. Discussion about Measurement Results of three types of RISs

  • Varactor RIS unit-cell power is approximately zero, but its drive circuits can be more complex and energy-hungry; discrete-level bias is proposed as a lower-power direction.The authors note remaining challenges for varactor RIS design and manufacturing at mmWave and Terahertz frequencies.Source: A. Discussion about Measurement… · p. 10
  • PIN-diode RIS drive circuits consume at mW scale, while unit-cell power can be substantial; measured per-cell power is 12.56 mW for 2# and 3# RIS cells and 1.25 mW for 4#.The authors attribute the difference partly to diode type and applied on-voltage; 2# and 3# currents exceed on-state current, reserving operating margin.Source: A. Discussion about Measurement… · p. 10
  • RF switch unit cells are typically at the µW power scale, and their drive circuits could potentially be omitted if the master FPGA supplies control signals directly.The measured FPGA control board consumes watt-level power for data processing.Source: A. Discussion about Measurement… · pp. 10–11
  • RF switch-based RISs are reported as the most power-efficient of the three types, with low consumption in both drive circuits and unit cells.The authors also identify PIN-diode RISs as suitable for high-frequency, discrete-phase scenarios and varactor-diode RISs for high phase resolution or precision regulation.Source: A. Discussion about Measurement… · p. 11

B. Discussion about how is the total power consumption modified by the number of RIS unit cells

RIS power responds differently to unit-cell scaling across PIN-diode, varactor-diode, and RF-switch designs. The simplified models attribute these differences to unit-cell states, drive-circuit power, or nearly constant control-board power.

B. Discussion about how is the total power consumption modified by the number of RIS unit cells

  • For a single-polarization, 1-bit PIN RIS with independent control and all cells ON, power is approximately 4.8 + N × 0.01 W.The model treats drive-circuit power as negligible compared with unit-cell power.Source: B. Discussion about how is the total… · p. 11
  • For a single-polarization varactor RIS with independent continuous-phase control, power is approximately 4.8 + N × 0.43 W.The model assumes one drive circuit per unit cell and negligible unit-cell power.Source: B. Discussion about how is the total… · p. 11
  • For independently controlled, single-polarization RF-switch RISs, power is approximately the constant 4.8 W control-board draw.Drive-circuit and unit-cell power are treated as negligible relative to control-board power.Source: B. Discussion about how is the total… · p. 11
  • As unit-cell count increases, PIN-diode power rises with ON cells, varactor power rises rapidly, and RF-switch power changes little.The varactor increase is attributed to the energy consumed by each drive circuit.Source: B. Discussion about how is the total… · p. 11

C. Discussion about low power consumption design of RIS-assisted system

The discussion links RIS power-saving choices to architecture-specific power drivers and system performance. Measurements support a model separating static and unit-cell power, while the design implications differ across PIN-diode, varactor-diode, and RF-switch RISs.

C. Discussion about low power consumption design of RIS-assisted system

  • PIN-diode drive circuits contribute little, so unit-cell states dominate power and low-power beamforming should use fewer PIN diodes coded as 1.A case study using the model reports that accurately calculating PIN-diode power is important when optimizing RIS energy efficiency.Source: C. Discussion about low power… · pp. 11–12
  • A 10 × 10 varactor RIS changed from unit-cell control (Ng=1) to column control (Ng=10) reduces drive-circuit power to 1/10, but spectral efficiency decreases.The control-degree-of-freedom choice trades spectral efficiency against system energy efficiency.Source: C. Discussion about low power… · p. 12
  • RF-switch RISs keep drive-circuit and unit-cell power low, but mmWave beam alignment may require power-demanding multi-bit PIN configurations; low-power design remains open.The paper identifies the RF-switch type as promising and calls for further investigation.Source: C. Discussion about low power… · p. 12

V. CONCLUSION

  • Measurements support a model separating static RIS power from unit-cell power, with the factors governing each component varying across RIS types.The FPGA control-board draw is constant, drive-circuit draw varies with control signals and circuit characteristics, and unit-cell power depends on RIS type.Source: §V · p. 12

Related papers

Loading 2211.00323v3…