PaperlayerView exact revision on arXiv

Source-linked AI summary

RIS-Aided Wireless Communications: Prototyping, Adaptive Beamforming, and Indoor/Outdoor Field Trials

Xilong Pei, Haifan Yin, Li Tan, Lin Cao, Zhanpeng Li, Kai Wang, Kun Zhang, Emil Björnson

arXiv:2103.00534v2cs.ITeess.SP

TL;DR

RIS-aided wireless communication lacks sufficient prototype and real-world field-trial evidence. This paper develops a 1100-element, 5.8 GHz RIS prototype with feedback-based adaptive beamforming and evaluates it indoors and outdoors, achieving gains up to 27 dB, 32 Mbps over 500 m, and approximately 1 W power consumption.

  • Problem

    Prototyping and real-world field trials of RIS-aided wireless communication remain scarce compared with theoretical studies based on elementary models.

  • Method

    The paper builds a 1100-element, 5.8 GHz RIS prototype and configures its 1-bit reflection coefficients using geometrical properties and real-time receiver-RIS feedback.

  • Results

    The prototype achieved 26 dB indoor gain through a 30 cm concrete wall, 27 dB short-range outdoor gain, and 32 Mbps over 500 m.

  • Takeaways & Limitations

    The field trials demonstrate the effectiveness and practical potential of RIS beamforming for future wireless communication systems.

Abstract

from arXiv · show

The prospects of using a Reconfigurable Intelligent Surface (RIS) to aid wireless communication systems have recently received much attention from academia and industry. Most papers make theoretical studies based on elementary models, while the prototyping of RIS-aided wireless communication and real-world field trials are scarce. In this paper, we describe a new RIS prototype consisting of 1100 controllable elements working at 5.8 GHz band. We propose an efficient algorithm for configuring the RIS over the air by exploiting the geometrical array properties and a practical receiver-RIS feedback link. In our indoor test, where the transmitter and receiver are separated by a 30 cm thick concrete wall, our RIS prototype provides a 26 dB power gain compared to the baseline case where the RIS is replaced by a copper plate. A 27 dB power gain was observed in the short-distance outdoor measurement. We also carried out long-distance measurements and successfully transmitted a 32 Mbps data stream over 500 m. A 1080p video was live-streamed and it only played smoothly when the RIS was utilized. The power consumption of the RIS is around 1 W. Our paper is vivid proof that the RIS is a very promising technology for future wireless communications.

I. INTRODUCTION

The paper addresses propagation limitations that motivate RIS technology by presenting a practical prototype, adaptive feedback-based beamforming, and indoor and outdoor field trials. The system demonstrates substantial power gains, long-range data transmission, and low power consumption.

  • Motivation and RIS concept: RIS technology electronically tunes a two-dimensional metamaterial surface to control electromagnetic-wave reflection toward desired locations.It uses many controllable elements whose reflection coefficients can be varied to synthesize useful electromagnetic responses.
  • System and algorithm: The prototype operates at 5.8 GHz and uses feedback-based real-time beamforming to serve mobile users without modifying existing communication standards.The receiver feeds received-signal strength back to the RIS, which updates reflection coefficients using a self-adaptive algorithm.
  • System and algorithm: The RIS consumed about 1 W of power in the demonstrated communication system.The low-power prototype was evaluated alongside indoor and outdoor communication measurements.
  • Field trials: 27 dB power gain was achieved in short-range outdoor testing, while the 500 m outdoor trial delivered 32 Mbps using the RIS.The long-range trial is reported as the world’s first long-range outdoor test.

II. DESIGN OF RECONFIGURABLE INTELLIGENT SURFACE

The paper develops a 1100-element varactor-tuned RIS at 5.8 GHz, using voltage-controlled reflection responses and a simplified low-bit hardware design. Simulations show at least 180° phase control across 5.5–6.0 GHz, supporting the fabricated board and field-trial configuration.

  • Reflective Element Design: 1100 elements comprise the large-scale RIS hardware prototype.The metasurface is tuned by varactor diodes for C-band operation.
  • Reflective Element Design: Each RIS element uses two varactor-controlled metallic patch pairs over a ground plane, with bottom-layer DC biasing lines.The three-layer structure includes patches, a reflecting ground plane, vias, and biasing circuitry.
  • Reflective Element Design: A series RLC model represents the integrated varactor diode in electromagnetic simulations.The simulations evaluate reflection spectra under different reverse-bias voltages.
  • Reflective Element Design: At 5.5–6.0 GHz, changing bias voltage from 0 to 19 V produces at least 180° phase shift while amplitude variations remain substantially smaller.The largest losses occur when the phase is close to zero; the orange response region is selected for field communications.
  • RIS Board Design: The fabricated board has a 55 × 20 element grid and measures 80.08 cm × 31.30 cm including the control circuit.The RIS portion itself measures 78.65 cm × 20.54 cm.
  • RIS Board Design: Two voltage states provide 1-bit phase control, while groups of five elements in each column share one bias voltage.This reduces implementation complexity by using lower precision vertically while retaining greater azimuthal flexibility.

III. RIS-AIDED WIRELESS COMMUNICATION SYSTEM

The prototype integrates PCs, USRPs, antennas, an FPGA controller, and receiver-RIS feedback to configure the RIS during communication. It optimizes reflection coefficients to increase received signal power while preserving the existing WiFi over-the-air signaling.

  • System Prototype: The communication prototype combines PCs, USRPs, antennas, an RIS, an FPGA master board, and a real-time RIS-UE feedback module.The transmitter and receiver emulate a base station and user equipment.
  • Signal Processing: The system transmits encoded video over UDP using OFDM with 20 MHz bandwidth and 312.5 kHz subcarrier spacing.The receiver synchronizes, compensates carrier-frequency offset, decodes the stream, and plays the video.
  • System Prototype: A serial feedback link connects the receiver and RIS board for real-time reflection-coefficient adjustment.The link supports adaptation to time-varying channels and switching between user equipments.
  • Beamforming Method: The RIS maximizes received signal power by configuring element reflections so their signals arrive at the receiver in phase within hardware granularity.With an RIS-to-receiver LoS path, this optimization forms a reflected beam toward the receiver.
  • Beamforming Method: Existing algorithms were unsuitable because practical operation requires joint CSI acquisition and RIS configuration with low-bit element resolution.This motivated development of a new prototype-oriented algorithm.

A. Codebook-based RIS Beamforming

The RIS beamforming model uses UPA geometry and array-response codebooks to configure reflection phases for dominant propagation paths. Continuous-phase designs align reflected signals at the receiver, while the prototype quantizes phases to two states and searches for a nearby configuration.

  • UPA model: UPA geometry represents RIS elements on a rectangular grid, with array responses factored into y-axis and z-axis components.The model uses column-by-column element ordering and Kronecker products to combine the two dimensions.
  • Channel model: The end-to-end SNR is proportional to |(g_k ⊙ h_k)^Tω|^2 and is tuned by selecting the RIS reflection-coefficient vector ω.The same ω applies across subcarriers, and the single-dominant-path assumption makes the resulting gain independent of subcarrier index.
  • Continuous-phase beamforming: The continuous-phase optimum selects reflection phases so signals from all RIS elements arrive in phase at the receiver.This solution follows from maximizing the inner product under unit-modulus reflection coefficients.
  • Codebook construction: A 2D codebook formed as F(M) ⊗ F(N) approximates RIS configurations that reflect an incident plane wave toward distinct angular directions.For large M and N, its columns approximate the desired angular beamforming configurations.
  • Quantized phases: The prototype restricts each phase shift to −π/2 or +π/2, so the design selects the quantized vector with maximum inner product with the ideal solution.The quantized choice is obtained by selecting the phase state according to the argument of each ideal coefficient.
  • Practical search: A codebook search can maximize received power when angular information is unavailable, with only a minor expected beamforming loss below the Fraunhofer distance.The codebook is designed from far-field array responses, but the paper expects limited loss at shorter distances.

B. Proposed Fast Beamforming Algorithm

The proposed fast beamforming method combines dominant-path geometry with receiver feedback to search RIS rows and columns directly. It progressively accepts phase inversions that increase received power, reducing feedback and computation relative to unstructured channel estimation.

  • Design rationale: The method exploits dominant angular paths and UPA codebook structure to reduce training and feedback requirements.The target configuration is approximately a column of F(M) ⊗ F(N), whose quantized form has repeated adjacent phase states.
  • Greedy search: The greedy algorithm inverts one RIS row or column at a time and keeps the new state when it increases received-signal power.The UE-RIS feedback module supplies the power measurements used for each decision.
  • Search procedure: The reflection matrix is searched horizontally and vertically using configurations R_t and received-power feedback p_t.Columns can be switched by changing their sign because the two available states are −j and j; p_t is averaged over the signal bandwidth.
  • Efficiency: M + N feedback iterations and O(M + N) computational complexity replace the MN iterations required for estimating unstructured channels.The algorithm can support simultaneous data transmission without significant received-power fluctuations.
  • Convergence: The algorithm monotonically increases SNR and is guaranteed to converge under static channel conditions, including circumstances beyond its dominant far-field-path design.The paper notes applicability to near-field propagation as well as the intended single-path setting.

V. EXPERIMENTAL RESULTS

The experimental section evaluates the prototype through electromagnetic, reciprocity, radiation-pattern, indoor, and outdoor measurements. These tests are presented as verification of the proof-of-concept system’s performance.

  • Experimental program: The experiments first test electromagnetic reflection characteristics, then channel reciprocity and radiation patterns, followed by indoor and outdoor system measurements.The radiation-pattern test is conducted in an anechoic chamber to measure the RIS beamforming function.

A. Bias Voltage Response Test

The bias-voltage response tests characterize fabricated RIS elements and show that their reflection phase depends on incidence angle. This angle dependence constrains phase control and motivates measurement-based voltage selection.

  • Calibration: Measured reflection coefficients are used to compensate for fabrication tolerances, dielectric variation, and non-ideal varactor behavior.The paper states that fabricated electromagnetic characteristics can differ from ideal simulations, so control voltages are selected from measurements.
  • Voltage response: Fig. 9 measures the relationship between RIS control voltage and reflection gain and phase at 5.8 GHz for 45° incidence and reflection.The figure directly characterizes voltage-controlled reflection behavior under equal 45° angles.
  • 1-bit calibration: 70° and 250° phase shifts at 3.8 V and 16 V, respectively, produce the desired 180° phase difference for the 1-bit RIS.These voltages identify the two bias states used to realize the target phase separation.
  • Angle dependence: 276° phase range occurs at 15°, 265° at 30°, and 250° at 45°, demonstrating angle-dependent RIS response.The paper attributes the effect to angle-dependent element reflectance and notes that larger incidence angles produce smaller phase variations.

B. Channel Reciprocity Test

The test examines whether RIS uplink and downlink channels are reciprocal, enabling channel knowledge from one direction to inform the other. It also characterizes the RIS wireless channel and its measurement environment.

  • TDD systems exploit reciprocal uplink and downlink channels so channel estimation can use the direction requiring less signaling.The paper motivates testing whether this principle applies to RIS-aided communications.
  • The test environment includes two horn antennas positioned near the RIS, as illustrated in the channel reciprocity setup.
  • The RIS channel reciprocity experiment uses a vector network analyzer to measure uplink and downlink S-parameters.Two horn antennas are placed close to the RIS to minimize the influence of propagation effects.
  • The measured RIS wireless-channel parameters cover the 5.7 GHz to 5.9 GHz frequency range.

C. Radiation Pattern Test

The radiation-pattern test evaluates beamforming by configuring the RIS with a quantized codebook and measuring its directional response in an anechoic chamber.

  • The RIS beamforming experiment is conducted in a microwave anechoic chamber with the RIS and transmit antenna mounted on a rotating platform.The transmit antenna faces the RIS during measurement.
  • A 1-bit quantized 2D-DFT codebook configures the RIS to form a beam at 30 degrees in azimuth.
  • The platform rotation is used to measure the radiation pattern produced by the predefined reflection coefficients.

E. Outdoor Over-the-air Test

The outdoor over-the-air evaluation uses field trials at 50 m and 500 m to assess RIS-aided wireless communication over different propagation distances.

  • The outdoor field trials are conducted at Huazhong University of Science and Technology using two propagation distances: 50 m and 500 m.
  • The outdoor evaluation complements the indoor non-LoS test, which places the transmitter and receiver across a 30 cm concrete wall.
  • The outdoor system settings are summarized in Table IV.

1) 50 meter test:

The 50 m field trial measures RIS performance against a copper-plate reference, while the paper also reports long-distance transmission and discusses why gains can decrease with distance.

  • 50 meter test: 13 dBm transmit power is used in the 50 m rooftop field trial, with receiver spectra measured with and without the RIS.The RIS is configured using the proposed Algorithm 1.
  • 50 meter test: 30.4 dB is the theoretical average power gain for a fully optimized RIS with 1100 elements relative to a random configuration.Accounting for the reported −3.9 dB average loss of 1-bit configurations gives a predicted 26.5 dB gain.
  • 500 meter test: 14 dB power gain at 500 m enabled real-time transmission of 1920 × 1080 video, which played smoothly only with the RIS.
  • 500 meter test: At longer distances, the RIS path can become proportionally weaker than surrounding multipaths, or frequency selectivity can reduce beamforming gain across the band.
  • 50 meter test: RIS gain depends on element patterns, RIS-antenna distances, incidence and reflection angles, and the number of reflection elements.

F. Power Consumption

The prototype’s power consumption is concentrated in its control electronics, while the RIS itself is reported to consume about 1 W. The authors project substantially lower consumption for a minimum-power design.

  • Most RIS-board power is spent on level-regulator chips that enable continuous bias-voltage adjustment.The authors note that this capability is useful for experiments but may not be necessary commercially.
  • 1.5 W is consumed by the high-end FPGA controller in the prototype.The controller is separate from the RIS board.
  • 10 mW is the projected controller consumption when the beamforming algorithm is implemented on an ordinary microcontroller.The authors identify this as a route to reducing controller power.
  • Far below 1 W is the predicted consumption for an equal-sized RIS designed for minimum power consumption.This is a prediction rather than a measured result.
  • 1 W is the reported power consumption of the RIS prototype.This figure refers to the RIS prototype in the field-trial summary.
Loading 2103.00534v2…