Fluid Antenna Multiple Access (FAMA)

• Fluid Antenna Multiple Access (FAMA) is a wireless access paradigm that uses reconfigurable antennas to dynamically switch ports and exploit micro-scale spatial diversity.
• FAMA employs user-centric port selection to maximize the signal-to-interference ratio, reducing outage probability even in compact devices.
• It offers a cost-effective alternative to traditional MIMO systems by using a single RF chain to achieve high scalability and effective interference mitigation.

Fluid Antenna Multiple Access (FAMA) is a multiuser wireless access paradigm leveraging reconfigurable antenna systems that can instantaneously switch the active radiating port among a large set of physically proximate but spatially distinct positions. By exploiting the fine-grained spatial diversity inherent in multipath channels—even within physically constrained mobile devices—FAMA enables opportunistic interference mitigation and user-centric diversity gains. Its operational core is user-side spatial port selection to maximize the signal-to-interference ratio (SIR) or signal-to-interference-plus-noise ratio (SINR), yielding a system capacity that can surpass that of traditional multi-antenna maximum ratio combining (MRC) systems at large port counts, even when the physical antenna aperture is as small as a few wavelengths (Wong et al., 2020).

1. Physical Principle and Fluid Antenna Configuration

A fluid antenna consists of a “position-reconfigurable” structure whereby a single RF chain is associated with N preset antenna locations (“ports”) distributed over a compact spatial region. Typical implementations employ flexible elements—liquid metal, conductive fluids, or electronically controlled pixels—that can be dynamically activated to emulate the antenna being present at one spatial location at a time. The port positions are parameterized as

$$d_k = \frac{k-1}{N-1}\cdot W\lambda, \quad k=1,2,\ldots,N,$$

where $W$ is the normalized aperture size (in wavelengths) and $\lambda$ the carrier wavelength.

Due to the compactness, the underlying spatial channels across ports are highly correlated—often modeled using the zeroth-order Bessel function as

$$\phi_{g_k,g_\ell}(d_k-d_\ell) = (\sigma^2/2) J_0\left(\frac{2\pi}{N-1}W|k-\ell|\right),$$

where $J_0(\cdot)$ is the Bessel function. Critically, despite strong correlation, substantial micro-scale spatial power variations remain—especially in multipath environments—causing both the desired signal and the aggregate interference to experience rapid spatial “fading” across the ports.

2. User-Centric Interference Avoidance via Port Selection

The FAMA access protocol operates without centralized coordination. Each user terminal independently selects, on a per-coherence-interval (slow FAMA) or per-symbol (fast FAMA) basis, the port $k^*$ that maximizes a user-centric metric, typically the SIR: $$k^* = \arg\max_{k} \left\{\frac{|g_k|^2}{|g_k^I|^2}\right\},$$ where $g_k$ is the channel coefficient at port $k$, and $g_k^I$ is the corresponding interference projection (often modeled as complex Gaussian due to aggregated multiuser interference). The receiver then uses only the selected port’s output for further baseband processing.

This approach “rides the deep fades” of the interference process—even leveraging those channel spatial positions where the desired signal is not maximized, but where the interference is exceptionally weak. Thus, the port switching mechanism opportunistically exploits the statistical independence between signal and interference spatial fades, circumventing the need for sophisticated multiuser detection or transmitter-side precoding.

3. Outage Probability and Diversity–Multiplexing Scaling

The central performance metric is the outage probability of the SIR per user: $$\mathrm{Prob}(\mathrm{SIR} \leq \gamma) = \mathrm{Prob}\left(\max_k \frac{|g_k|^2}{|g_k^I|^2} \leq \gamma\right).$$ The joint distribution is determined via a double-integral (see Theorem 1, (Wong et al., 2020)) involving the spatial correlation structure, the envelope distributions (Rayleigh or more generally Rician), and special functions (modified Bessel $I_0$, Marcum-Q $Q_1$). This double integral captures the interplay between the desired and interfering channel envelopes across spatial positions, reflecting nontrivial spatial correlation.

A closed-form upper bound on the outage probability, under simplifying correlation assumptions ($|\mu_2| = \cdots = |\mu_N| = \mu$), is derived using bounds on the Marcum-Q function and Bessel inequality: $$\mathrm{Prob}(\mathrm{SIR} \leq \gamma) \leq \epsilon^{(\mathrm{II})}_{UB}(\gamma)$$ (with the closed-form expression depending on $\mu$, $\gamma$, and port count $N$).

This enables the derivation of a lower bound on the average outage capacity: $$C_{\mathrm{FAMA}}(\gamma) \geq (N_I + 1)\left[1 - \epsilon_{UB}(\gamma)\right]\log_2(1+\gamma),$$ where $N_I$ is the number of interferers.

The achievable “multiplexing gain” is thus quantified as

$m \equiv (N_I+1)(1-\epsilon_{UB}(\gamma)),$

which reveals that, for large $N$ and moderate spatial correlation ($\mu^2 < 1$), $m$ can approach $N_I+1$, i.e., all users can be supported without outage. For practical regimes, the following asymptotic holds: $$m \approx \min\left\{ \frac{(N-1)(1-\mu^2)}{\gamma},\, N_I + 1\right\},$$ explicitly showing that multiplexing gain scales linearly in $N$, inversely with the SIR target $\gamma$, and is strongly limited by spatial correlation ($\mu$). The required number of ports to guarantee a target $m$ is

$N \gtrsim \frac{m\gamma}{1-\mu^2} + 1.$

4. Spatial Correlation and Port Design Trade-offs

A fundamental design constraint in FAMA is imposed by port spatial correlation (parameterized by $\mu$ or the autocorrelation function). For a fixed aperture, increasing $N$—and thus densifying ports—does not indefinitely improve performance: as $\mu^2 \to 1$, ports become nearly redundant, and the effective diversity saturates.

Increasing aperture size $W$ (up to $\sim \lambda/2$) decreases $\mu^2$ and enhances spatial independence, yielding higher attainable diversity and multiplexing. In practice, a few wavelengths’ span and $N \sim 100$–$1000$ achieve most of the available performance.

Careful design balances the aperture, number of ports, and spatial arrangement to optimize the trade-off between attainable diversity (and thus outage reduction) and practical antenna realization constraints.

5. Hardware, Complexity, and Network Implications

FAMA’s architecture requires only a single RF chain per terminal, regardless of the number of ports, as only one port is active at any instant. This contrasts with massive MIMO, which needs as many RF chains as antennas, thus incurring much higher hardware and power costs. The instantaneous port activation is typically achieved by fast electronic or microfluidic switching, allowing selection on timescales commensurate with channel coherence intervals.

The lack of transmitter-side CSI or coordinated scheduling simplifies the network stack. Each user independently exploits its own spatial diversity, and FAMA scales to hundreds of users in the same resource block, with network-wide outage capacity increasing almost linearly in the number of users until physically limited by correlation or SIR target.

6. Comparison and Relative Merits

Compared to traditional MRC with multiple fixed antennas, FAMA can outperform even for a much smaller occupied spatial region, provided $N$ is large and spatial correlation is managed. The multiuser nature is preserved without multiuser detection or transmitter-side spatial processing. FAMA is especially attractive for compact devices, Internet-of-Things (IoT) nodes, or densely populated indoor/urban scenarios where antenna size is at a premium.

Moreover, the opportunistic port selection principle is robust to the statistics of both the desired signal and the interference, allowing deployment in diverse propagation and multiuser environments.

7. Prospects and Design Guidelines

Strong analytical results illustrate FAMA's feasibility as a scalable multiuser access scheme. For network deployment:

Parameter	Influence on FAMA Performance	Design Approach
N (number of ports)	Higher $N$ improves diversity/multiplexing until saturation by $\mu$	Set $N$ just large enough for target gain
$W$ (antenna size / λ)	Larger $W$ reduces correlation, increases effective channels	Select $W\sim$ a few λ where feasible
SIR threshold ($\gamma$)	Higher $\gamma$ reduces multiplexing gain for fixed $N$	Adjust $N$ and $W$ to meet QoS/SIR constraints
Interference environment	More interferers increase benefit from smart port selection	FAMA scales well with user density

Designers should ensure that $N$ and $W$ are chosen so that

$N \gtrsim \frac{m\gamma}{1-\mu^2} + 1,$

for target multiplexing gain $m$, SIR threshold $\gamma$, and spatial autocorrelation $\mu^2 \approx J_0(\pi W)$ (Wong et al., 2020). By this principle, even a few wavelengths’ aperture with rapid port switching can suffice for high-reliability, massive-access applications.

Summary

Fluid Antenna Multiple Access provides a user-driven, spatially opportunistic method for massive multiuser access in dense networks using space-constrained hardware. Fundamental probabilistic analysis reveals that, with suitable port count and spatial design, single-RF-chain systems can achieve diversity, interference mitigation, and linear user scaling comparable to or exceeding that of more complex multi-antenna systems. The essential mechanisms are user-side port selection to maximize instantaneous SIR and the leveraging of micro-scale spatial channel variations arising from environmental scattering. FAMA thus enables scalable, practical deployments for future wireless networks, especially where massive connectivity and interference management are paramount (Wong et al., 2020).