PaperlayerView exact revision on arXiv

Source-linked AI summary

Channel Capacity Optimization Using Reconfigurable Intelligent Surfaces in Indoor mmWave Environments

Nemanja Stefan Perović, Marco Di Renzo, Mark F. Flanagan

arXiv:1910.14310v2cs.ITeess.SPmath.IT

TL;DR

Obstacles can block the dominant LOS path in indoor mmWave environments, making communication difficult when few reflective surfaces are available. The paper studies two RIS-based phase-optimization schemes for capacity without an LOS path; simulations report significant gains that increase with the number of RIS elements.

  • Problem

    When obstacles block the dominant LOS path in indoor mmWave environments, communication can be extremely difficult if few surfaces with low reflection losses are available.

    Source: §I · p. 1

  • Method

    The paper proposes a RIS-only phase-optimization scheme with an approximate capacity expression and a joint RIS and transmit-phase-precoder scheme using global co-phasing.

    Source: §I · p. 2

  • Results

    Both proposed schemes produce significant channel-capacity gains, and the gain increases with the number of RIS elements.

    Source: §V · p. 7

  • Takeaways & Limitations

    The simulations indicate that the proposed phase-based schemes are particularly suitable for indoor mmWave systems with many RIS elements.

    Source: §I · p. 2

    The study considers single-stream transmission in indoor mmWave environments without an LOS path.

    Source: §V · p. 6

Abstract

from arXiv · show

Indoor millimeter-wave (mmWave) environment channels are typically sparsely-scattered and dominated by a strong line-of-sight (LOS) path. Therefore, communication over such channels is in general extremely difficult when the LOS path is not present. However, the recent introduction of reconfigurable intelligent surfaces (RISs), which have the potential to influence the propagation environment in a controlled manner, has the potential to change the previous paradigm. Motivated by this, we study the channel capacity optimization utilizing RISs in indoor mmWave environments where no LOS path is present. More precisely, we propose two optimization schemes that exploit the customizing capabilities of the RIS reflection elements in order to maximize the channel capacity. The first optimization scheme exploits only the adjustability of the RIS reflection elements; for this scheme we derive an approximate expression which explains the connection between the channel capacity gains and the system parameters. The second optimization scheme jointly optimizes the RIS reflection elements and the transmit phase precoder; for this scheme, we propose a low-complexity technique called global co-phasing to determine the phase shift values for use at the RIS. Simulation results show that the optimization of the RIS reflection elements produces a significant channel capacity gain, and that this gain increases with the number of RIS elements.

I. INTRODUCTION

Indoor mmWave links can be difficult to establish when obstacles block the dominant LOS path and few reflective surfaces are available. The paper proposes two phase-based RIS capacity-optimization schemes for single-stream MIMO without LOS, and reports gains that increase with the number of RIS elements.

I. INTRODUCTION

  • Indoor 60 GHz communication is suited to short ranges, and obstacles can block the dominant LOS path, making reliable links difficult when few low-loss reflective surfaces are present.RIS elements can adjust reflected-signal phase to modify radio waves and their interactions with surrounding objects.Source: §I · p. 1
  • Prior RIS work in indoor mmWave considered single-antenna transmitter and receiver systems without transmit phase precoding, while other work studied single-antenna-receiver designs.The paper frames its schemes against these restrictions in prior mmWave RIS studies.Source: §I · p. 2
  • The paper proposes two capacity-optimization schemes for single-stream MIMO using RISs in indoor mmWave environments without an LOS path.Both schemes rely on signal phase adjustments and are not generally globally optimal, enabling low-complexity hardware architectures.Source: §I · p. 2
  • The RIS-only scheme derives optimal RIS phase shifts and an approximate capacity expression linking capacity gains with system parameters.The approximation simplifies capacity calculation for this scheme.Source: §I · p. 2
  • The joint scheme adjusts RIS elements and the transmit phase precoder, using low-complexity global co-phasing to set RIS phase shifts.Global co-phasing determines appropriate phase shifts for the RIS reflection elements.Source: §I · p. 2
  • Simulations indicate the proposed schemes are particularly suitable for indoor mmWave systems with many RIS elements.The schemes use phase adjustments and are not generally globally optimal, but require only low-complexity hardware architectures.Source: §I · p. 2

II. SYSTEM MODEL

The system model places transmit and receive antenna arrays on parallel walls and one RIS on a perpendicular surface. It models a single IQ stream and expresses the RIS-assisted channel using approximations under stated assumptions about the RIS elements and geometry.

II. SYSTEM MODEL

  • Transmit and receive ULAs sit on parallel walls separated by distance D, while a one-dimensional RIS is installed on a perpendicular surface.The model specifies array heights and spacings, RIS element spacing, and the distance from the RIS midpoint to the transmit-array plane.Source: §II · p. 2
  • The model assumes ideal RIS elements that independently adjust the phase and reflection angle of each impinging wave.Source: §II · p. 3
  • The system focuses on transmission of a single in-phase and quadrature stream, with a transmit precoding vector and additive complex Gaussian noise.The precoding vector is constrained by the stated norm condition.Source: §II · p. 3
  • The channel-matrix element models propagation from a transmit antenna to a receive antenna through all RIS reflection elements.The distances from the RIS element to each antenna and the element's induced phase shift define the path terms.Source: §II · p. 3
  • Channel coefficients are normalized by the free-space path loss for a center-RIS reflective path between arrays at their mean heights.The approximation uses nearly equal products of transmit-to-RIS and RIS-to-receive distances across the indexed elements and antennas.Source: §II · p. 3
  • The channel matrix is expressed using matrices U, V, and F, representing approximate transmit-side and receive-side responses and the diagonal RIS phase shifts.The geometric approximation is justified by the short antenna arrays relative to the RIS distances and the assumed adjustable reflection angles.Source: §II · p. 3

III. PROPOSED OPTIMIZATION SCHEMES

The schemes optimize single-stream indoor mmWave channel capacity using RIS phase adjustment alone or jointly with transmit phase precoding. Receive SNR is central to capacity, and the assumed receiver combines signals by summing across receive antennas.

III. PROPOSED OPTIMIZATION SCHEMES

  • Receive SNR primarily determines single-stream channel capacity, motivating the assumed combining of signals summed across all receive antennas.Source: §III · p. 3
  • Two schemes optimize single-stream indoor mmWave capacity: one adjusts RIS phases alone, while the other jointly adjusts RIS phases and transmit phase precoding.Source: §III · p. 3

A. RIS-only Optimization Scheme

The RIS-only scheme leaves transmit precoding unadjusted and selects RIS element phases according to an optimization condition. Its received-signal expression incorporates the resulting noise model and is used to calculate capacity.

A. RIS-only Optimization Scheme

1) Approximate Expression:

The section derives an approximate expression to simplify RIS-only capacity computation using approximate channel responses and an auxiliary function. The function's lobe behavior implies that the RIS-only scheme is primarily suited to systems with limited transmit and receive antennas.

1) Approximate Expression:

  • The RIS-only capacity approximation simplifies computation of (9), evaluates the contribution via an arbitrary RIS element, and is reported to produce very accurate simulation results.Source: 1) Approximate Expression · p. 4
  • The approximation models transmit-to-RIS and RIS-to-receive channel responses using array midpoints, element distances, and departure or arrival angles.Its validity assumes adjacent-antenna separations are significantly smaller than the RIS-to-array distances; the channel response expressions neglect free-space path loss.Source: 1) Approximate Expression · p. 4
  • The expression is a product of two instances of g(N, x) = sin(Nx)/sin(x), whose main lobes narrow and sidelobes multiply as N increases.Source: 1) Approximate Expression · p. 4
  • For large transmit or receive antenna counts, the expression is more likely to be small, making RIS-only optimization primarily suitable for systems with limited antenna counts.Source: 1) Approximate Expression · p. 4

B. Joint Optimization Scheme

The joint scheme adjusts both RIS-element phases and the transmit phase precoder, using phase-only precoding for practical, low-complexity hardware. It replaces an analytically intractable optimization with global co-phasing, whose gains can be limited when there are many receive antennas.

B. Joint Optimization Scheme

  • The joint scheme uses phase-only transmit precoding alongside RIS phase adjustments, favoring low-complexity hardware over the best possible performance.The authors note that phase precoding may not achieve the best possible performance.Source: B. Joint Optimization Scheme · p. 4
  • Because the joint phase optimization is analytically intractable and exhaustive search is too large, global co-phasing minimizes the absolute sum of phase deviations for each RIS element.The target phase is set to 0; the phase choices cannot co-phase every transmit-term expression simultaneously.Source: B. Joint Optimization Scheme · p. 5
  • Global co-phasing can yield only limited capacity gains with many receive antennas, so the scheme is most suitable for systems with few receive antennas.The authors attribute the limitation to the amplitude of the RIS-phase-influenced term becoming more likely to be small as the receive-antenna count grows.Source: B. Joint Optimization Scheme · p. 5

IV. SIMULATION RESULTS

Simulations compare the RIS-only and joint schemes with co-phasing MIMO and basic MIMO in an indoor environment without a LOS path. The joint scheme has the highest capacity in all cases, while RIS-only approaches it as the RIS grows.

IV. SIMULATION RESULTS

  • The simulations assume an indoor mmWave channel with no LOS path, with the RIS-provided paths as the only non-negligible NLOS paths.The two benchmarks are co-phasing MIMO, which uses transmit and receive phase precoding, and basic MIMO, which uses neither precoding nor RIS phase adjustment.Source: §IV · p. 5
  • To obtain geometry-independent channel-capacity results, antenna-array heights ht and hr are sampled uniformly over 2–3 m and 0.8–1.8 m, respectively, at 2 cm resolution, and capacity is averaged by Monte Carlo.The simulations use a 60 GHz carrier, a 5 m transmitter–receiver distance, and a RIS 2.5 m from the transmitter.Source: §IV · p. 5
  • The joint optimization scheme achieves higher channel capacity than the other schemes in all cases; RIS-only approaches it as the number of RIS elements increases.When RIS elements greatly outnumber transmit antennas, transmit-precoding gains are negligible compared with RIS phase-adjustment gains.Source: Figure 3 · p. 6, §IV · p. 6

V. CONCLUSION

The paper studies single-stream capacity optimization with RISs in indoor mmWave environments without a LOS path. It proposes RIS-only and joint RIS/transmit-precoder schemes, reporting significant capacity gains that increase with RIS size.

V. CONCLUSION

  • The paper proposes two channel-capacity optimization schemes for single-stream RIS-assisted transmission in indoor mmWave environments without a LOS path.The first adjusts RIS phases alone; the second also adjusts the transmit phase precoder.Source: §V · pp. 6–7
  • For RIS-only optimization, an approximate expression simplifies capacity calculation and relates achievable capacity gains to system parameters.The joint scheme uses a low-complexity technique to determine RIS phase shifts while jointly adjusting RIS elements and the transmit phase precoder.Source: §V · pp. 6–7
  • Simulations show significant channel-capacity gains from both schemes, with gains increasing as the number of RIS elements grows.Source: §V · p. 7

APPENDIX A PHASE ADJUSTMENTS FOR CO-PHASING MIMO

The co-phasing MIMO benchmark uses transmit and receive phase precoders before combining the receive signals. Its phase choices are designed to maximize the receive-signal term that primarily determines channel capacity.

APPENDIX A PHASE ADJUSTMENTS FOR CO-PHASING MIMO

  • Co-phasing MIMO uses phase-only transmit and receive precoders, with their phase values defining the precoding vectors.Source: Appendix A · p. 7
  • The phase precoders are designed to maximize the receive-signal term because single-stream channel capacity is primarily determined by receive SNR.Source: Appendix A · p. 7
  • The procedure first co-phases receive-antenna signals before combining them, then applies global co-phasing to the transmit-precoder elements.Source: Appendix A · p. 7
Loading 1910.14310v2…