PaperlayerView exact revision on arXiv

Source-linked AI summary

Fluid Antenna Multiple Access

Kai-Kit Wong, Kin-Fai Tong

arXiv:2006.05508v1cs.IT

TL;DR

The paper asks whether fluid antennas can simplify multiuser access by exploiting interference deep fades. It analyzes FAMA through SIR outage and outage-capacity bounds, then characterizes multiplexing gain and port requirements. The results indicate that one fluid antenna per user can support hundreds of users in a few wavelengths while increasing network outage capacity.

  • Problem

    Conventional multiuser systems place substantial processing and resource-management burdens on base stations as the number of users grows.

  • Method

    FAMA scans N fluid-antenna ports, selects the position with the strongest SIR, and analyzes outage probability, outage capacity, multiplexing gain, and required N.

  • Results

    The results show that sufficiently large N can yield arbitrarily small SIR outage probability, while network multiplexing gain grows linearly with N and outage capacity can support hundreds of users.

  • Takeaways & Limitations

    A single small fluid antenna at each user can enable FAMA for hundreds of users sharing the same radio resource and produce significant capacity gain.

Abstract

from arXiv · show

Fluid antenna is a novel technology that can make an antenna appear instantly at one of N preset locations in a predefined space. An important application is to adopt fluid antenna in a small space of mobile device for obtaining the tremendous diversity hidden in the small space. Previous results have revealed that a single-antenna fluid antenna system, even with a very small space, can outperform a multiple antenna maximum ratio combining (MRC) system if $N$ is large enough. This paper explores the potential of using fluid antenna for multiple access through performance analysis. Fluid antenna multiple access (FAMA) exploits moments of deep fade experienced by the interference to achieve a favourable channel condition for the desired signal, without requiring sophisticated signal processing. We analyze the FAMA system by first deriving the outage probability of the signal-to-interference ratio (SIR) in a double integral form. We then obtain an outage probability upper bound in closed form and an average outage capacity lower bound for the FAMA system, with an arbitrary number of interferers, from which the multiplexing gain of FAMA is characterized. We also estimate how large N is required to achieve a given multiplexing gain using fluid antennas with a given size. Results illustrate that it is possible for FAMA to support hundreds of users using only one fluid antenna at each user in a few wavelengths of space, giving rise to significant enhancement in the network outage capacity.

I. Introduction

FAMA uses fluid antennas to exploit interference deep fades for multiple access, avoiding sophisticated signal processing. The paper analyzes outage probability, outage capacity, and multiplexing gain as functions of ports, antenna size, and users.

  • Motivation: Massive-MIMO approaches can strain base stations because supporting many users requires many antennas and additional interference-management tasks.The introduction cites channel inversion, pilot decontamination, power control, and user-cell association as associated processing burdens.
  • Motivation: 50dB interference fades are desirable because FAMA exploits them to improve the desired signal’s channel condition.Rather than avoiding deep fades, the approach selects spatial positions where interference is deeply faded.
  • Fluid Antennas: A fluid antenna switches among N preset ports in a small space, using a common RF chain to scan fading envelopes.Prior work reported that a single fluid antenna can match multi-antenna MRC capacity in half-wavelength or less when N is large enough.
  • FAMA: FAMA users select the position with the strongest SIR, naturally avoiding inter-user interference without coordination or sophisticated processing.Each user independently switches its antenna position to suppress interference from coexisting transmitters.
  • Analysis: The paper derives exact and bounded SIR outage expressions, an average outage-capacity lower bound, a multiplexing-gain definition, and a sufficient port-count condition.The analysis characterizes how performance scales with the number of ports and fluid-antenna size.
  • Findings: FAMA’s network outage capacity and multiplexing gain scale linearly with N, while antenna-size benefits diminish beyond λ and gains are bounded by coexisting users.A single fluid antenna can achieve arbitrarily small SIR outage probability when N is sufficiently large.

II. Network Model

The network model places a fluid antenna with N switchable ports along a linear space of length Wλ. It models correlated Rayleigh fading and idealizes ports as point antennas sharing one RF chain.

  • A. Single-User Fluid Antenna System: Each mobile station has a fluid antenna whose location can switch among N preset ports distributed over a linear length Wλ.Here λ is the communication wavelength, and a switchable location is called a port.
  • A. Single-User Fluid Antenna System: All ports share one RF chain, and switching delay is assumed negligible for the analysis.The paper uses an ideal point-antenna abstraction for each port.
  • A. Single-User Fluid Antenna System: The received channel coefficient at each port is modeled as circularly symmetric complex Gaussian, giving Rayleigh-distributed channel amplitude.The model assigns zero mean and variance σ^2 to the complex channel coefficient.
  • A. Single-User Fluid Antenna System: Port channels are correlated because ports may be arbitrarily close, with correlation determined through an isotropic-scattering autocorrelation model.The autocorrelation uses the zero-order Bessel function of the first kind.
  • A. Single-User Fluid Antenna System: The model parameterizes the correlated channels using Gaussian random variables and autocorrelation parameters while preserving E[|g_k|^2] = σ^2.The first port serves as the reference location for port displacements.

B. FAMA

FAMA selects the port maximizing the signal-to-interference ratio in a multiuser interference channel. The model assumes independent desired and interference channels across links but spatial correlation across ports.

  • B. FAMA: The FAMA model considers one user with N ports and N_I interferers, corresponding to N_I + 1 users.The same statistical model applies to every user when their statistics are identical.
  • B. FAMA: Interferer signals are combined into a total interference channel at each port, modeled using parameters analogous to those of the desired signal.The interference model includes the transmitted data from each interferer.
  • B. FAMA: Each FAMA user switches to the port with the largest desired-to-interference ratio to obtain the best reception condition.The selected random variable corresponds to the square root of the received-signal SIR.
  • B. FAMA: Desired and interference channels are independent across links but correlated across spatial ports, making their ratios spatially correlated.The correlation arises because channels at nearby ports share spatial fading structure.
  • B. FAMA: The analysis uses SIR rather than SINR because it simplifies the derivation and is a reasonable approximation in interference-limited environments.The approximation is motivated when the number or power of interferers is large.

III. Main Results

The paper derives outage and capacity characterizations for FAMA, then uses them to quantify multiplexing gain and the roles of ports, antenna size, correlation, and users.

  • Outage capacity: The average network outage capacity is lower bounded using the per-user outage capacity and the total number of independent FAMA users.For a target SIR γ, per-user outage capacity is (1−θ) log2(1+γ), and the network capacity scales with the number of users.
  • Outage probability: FAMA’s exact SIR outage probability is derived in double-integral form, followed by closed-form upper bounds for analysis.The upper bound addresses the high complexity and limited insight of direct numerical integration when N is large.
  • Outage probability: As N →∞, the SIR outage probability approaches zero when |µ_k| ≠ 1 for every interferer-related port correlation.This establishes interference avoidance through port selection when sufficiently many ports are available.
  • System parameters: Increasing the number of interferers can relax the correlation requirement when total interference power is fixed, because additional multipath creates exploitable envelope fluctuations.The benefit depends on the fixed-total-power setting and does not remove the user-count limit on multiplexing gain.
  • Multiplexing gain: FAMA’s multiplexing gain grows with the number of ports N but is discounted by the target SIR γ and ultimately limited by the number of users.The analysis provides sufficient conditions on N and correlation µ for achieving a target multiplexing gain m.

IV. Numerical Results

Numerical results examine how ports, antenna size, users, interferers, and SIR targets affect FAMA capacity, showing strong gains with sufficiently many ports but conservative bound-based estimates.

  • Outage and multiplexing gain: With W = 2, N = 20, and NI = 5, FAMA supports 6 users at about 30% SIR outage and achieves multiplexing gain 4.2; N = 30 raises it to 4.8.These results use the upper-bound-based numerical study.
  • Capacity trends: Average outage capacity increases with N and W, but the W difference disappears for extremely large N, while capacity saturates at (NI + 1) log2(1 + γ).The saturation value is achieved as N →∞ when outage probability tends to zero.
  • Capacity trends: For large N, an optimal target SIR γ maximizes average outage capacity because overly small γ limits per-user rates while overly large γ sharply increases outage.The favorable effect of more interferers can reverse when γ is too large for a fixed N.
  • Outage and multiplexing gain: For γ = 0dB and N = 20, 101 users achieve multiplexing gain 13, which doubles to 26 when N increases to 50.The reported increase is attributed to reduced SIR outage probability and illustrates substantial gains at moderate N.
  • Antenna size: N is more influential than W: increasing W helps multiplexing gain, especially at larger N, but its benefit diminishes beyond roughly λ/2.The curves show correlation-induced ripples, while port increases continue to provide larger gains.
  • Antenna size: If N is too small, no feasible W achieves a specified multiplexing gain; once N is sufficiently large, required W decreases sharply and then gradually.The number of interferers has little effect on required W, while reducing W below λ/2 requires exponential increases in N.

V. Conclusion

The conclusion presents FAMA as a multiple-access approach that exploits interference fading through fluid-antenna port selection. It reports arbitrarily small SIR outage probability with sufficiently many ports, linear multiplexing-gain scaling, diminishing size returns beyond 2λ, and support for hundreds of users.

  • Conclusion: FAMA handles inter-user interference by scanning fading envelopes and selecting the best among closely located ports.The approach uses software-controlled, position-flexible antennas to exploit spatial moments when interference experiences deep fades.
  • Conclusion: Sufficiently many ports allow a fluid antenna to achieve arbitrarily small SIR outage probability.The conclusion identifies this result as demonstrating feasibility for interference elimination.
  • Conclusion: FAMA network multiplexing gain grows linearly with the number of ports at each user.The paper analyzes both average outage capacity and multiplexing gain for the FAMA network.
  • Conclusion: FAMA can accommodate hundreds of users in the same radio resource using one small fluid antenna at each user.The conclusion associates this configuration with a significant capacity gain.

A. Proof of Theorem 1

The proof of Theorem 1 builds the required probability expression by conditioning on interference-related variables and evaluating the resulting integrations. It uses a conditional cdf, a pdf, an identity, and variable substitutions.

  • A. Proof of Theorem 1: The derivation expresses the probability through an integral over conditional distributions of the relevant channel magnitudes.The integrand includes a conditional probability involving |g_1| through |g_N| and interference variables.
  • A. Proof of Theorem 1: The proof uses a conditional cdf and pdf for the interference-channel magnitudes to evaluate the integral.These distributional components are introduced from cited prior results.
  • A. Proof of Theorem 1: The remaining integration is simplified by noting the total probability property of a Rician random variable.This observation motivates the auxiliary lemma used in the derivation.
  • A. Proof of Theorem 1: Lemma 1 supplies an identity used after changing integration variables.The proof applies substitutions including x = t_k r and z = t_2 while evaluating the integrations.

B. Proof of Theorem 2

The proof of Theorem 2 derives an upper bound by applying Lemma 2 inside the outage-probability integral and then simplifying the resulting expression with additional bounds and substitutions.

  • B. Proof of Theorem 2: Lemma 2 provides a lower bound for Q1(α, β) used in the proof.The lemma is obtained by retaining only the first term of the cited definition.
  • B. Proof of Theorem 2: Applying Lemma 2 to the Q1(·, ·) term transforms the outage-probability integral.The proof then evaluates a difference and applies the resulting bound over the integration variable.
  • B. Proof of Theorem 2: The derivation further uses the lower bound I0(x) ≥ e^x/(1+2x) and substitutes t = 0.These steps produce the intermediate expression used for the final bound.
  • B. Proof of Theorem 2: A simple integration over t yields the final outage-probability upper bound.The bound follows after substituting the preceding result into the outage-probability expression.

C. Proof of Theorem 3

The proof of Theorem 3 imposes equal nonzero-mean magnitudes across ports, applies a binomial expansion, and combines the result with interference and SIR conditions.

  • C. Proof of Theorem 3: The proof assumes |µ_2| = ··· = |µ_N| = µ before simplifying the relevant expression.This equal-magnitude condition is used to obtain the subsequent closed-form manipulation.
  • C. Proof of Theorem 3: A binomial expansion is then applied to derive the theorem’s target expression.The expansion follows directly from the equal-magnitude assumption.
  • C. Proof of Theorem 3: The final result follows under interference and an ambitious SIR target.The proof states that these conditions yield the desired result.

D. Proof of Theorem 5

The proof constructs a lower-bounding fluid-antenna model using the distance-dependent behavior of the autocorrelation function, then derives a sufficient condition for the required antenna size and number of ports.

  • Proof construction: The proof models the fluid antenna as a linear space of Wλ with N ports and uses port autocorrelation parameters ordered by distance.The model retains the spatial structure needed to bound performance across antenna positions.
  • Proof construction: Although J0(·) oscillates with distance, the general trend of |J0(·)| decreases as the reference-position distance increases.This trend motivates enforcing a bound |J0(ρ)| ≤ µ* beyond a threshold ρ*.
  • Lower bound: The constructed model assigns greater diversity contribution to farther ports and lower-bounds the N-port system using a reduced set of ports.This agrees with the stated intuition that more distant ports contribute more diversity.
  • Sufficient condition: A sufficient condition based on J0^-1(µ*) determines the fluid-antenna size needed to achieve the SIR target and multiplexing gain.The condition is obtained by setting autocorrelation parameters using a port separation tied to Wλ.
Loading 2006.05508v1…