How Many Independent Modes Does a Fluid Antenna Have? A Closed-Form Outage Analysis via Equivalent Degrees of Freedom

Technical Summary: How Many Independent Modes Does a Fluid Antenna Have? A Closed-Form Outage Analysis via Equivalent Degrees of Freedom

Problem Statement
Fluid Antenna Systems (FAS) offer a promising solution for achieving spatial diversity in compact devices by dynamically selecting the best port among $N$ correlated locations within a physical aperture. However, accurate performance analysis, specifically outage probability, remains a significant challenge. Exact evaluation requires computing the joint cumulative distribution function (CDF) of $N$ correlated channel gains, which involves an intractable $N$-dimensional multivariate integral. Existing closed-form approximations, such as the Block-Correlation Model (BCM) and its variable extension (VBCM), simplify the problem by partitioning ports into independent blocks. While analytically convenient, these models systematically discard inter-block correlations, leading to an optimistic underestimation of the true outage probability. Furthermore, no existing closed-form expression simultaneously satisfies three critical criteria: (i) trivial evaluation without numerical integration, (ii) provable conservativeness (an upper bound on true outage), and (iii) consistency with the correct diversity order.

Methodology
The paper bridges FAS spatial correlation with the classical theory of bandlimited signals, specifically the Slepian-Landau-Pollak concentration theorem. The authors establish that for a FAS with a normalized linear aperture length $W$ (in wavelengths), the spatial correlation matrix possesses a specific spectral structure. Regardless of the total number of deployed ports $N$, the eigenvalues of the correlation matrix exhibit a sharp transition: the first $K^*$ eigenvalues form a dominant cluster, while the remaining eigenvalues decay exponentially to zero.

The number of significant eigenmodes is derived as:
$$K^* = 2\lceil W \rceil + 1$$
This quantity, termed the Equivalent Degrees of Freedom (EDoF), depends solely on the aperture size $W$, not on the port count $N$.

Based on this spectral concentration, the paper proposes two approximation frameworks:

1. EDoF Approximation: The correlated $N$-port system is modeled as selection combining over $K^*$ independent and identically distributed (i.i.d.) Rayleigh branches. This yields a simple closed-form outage expression: $P_{out}^{EDoF} = (1 - e^{-x})^{K^*}$, where $x$ is the normalized outage threshold.
2. Refined Weighted Independent Modes (WIM): To improve accuracy at moderate Signal-to-Noise Ratios (SNR), this variant incorporates the specific magnitudes of the dominant eigenvalues. It models the system as selection combining over $K^*$ independent branches with unequal weights $\{\beta_k\}$ derived from the normalized eigenvalues, resulting in a product-form expression: $P_{out}^{Ref} = \prod_{k=1}^{K^*} (1 - e^{-x/\beta_k})$.

Theoretical proofs utilize the Karhunen-Loève (KL) expansion and Anderson's inequality to demonstrate that these approximations never underestimate the true outage probability and achieve the exact diversity order $K^*$.

Key Contributions

• EDoF Approximation: Derivation of the first conservative, closed-form outage bound for FAS that requires no eigendecomposition, numerical integration, or knowledge of the port count $N$. It depends only on the normalized aperture $W$.
• Refined WIM: Development of a product-form refinement that accounts for eigenvalue spread, offering improved accuracy at moderate SNR while retaining a closed-form structure and the exact diversity order.
• Diversity Order and Asymptotic Tightness: Proof that both approximations achieve the exact diversity order $K^* = 2\lceil W \rceil + 1$. The EDoF formula is proven to be asymptotically exact at high SNR, with a bounded multiplicative gap to the true outage.
• Ergodic Capacity: Derivation of a closed-form series expression for ergodic capacity under the EDoF model using order statistics, enabling rapid system-level assessment without Monte Carlo simulation.
• Multi-User Extension (FAMA): Extension of the framework to Fluid Antenna Multiple Access (FAMA) systems, providing a closed-form expression for interference-limited outage floors. This reveals how user count, aperture size, and diversity order jointly determine the outage floor.
• 2D Planar FAS: Generalization of the EDoF concept to two-dimensional planar apertures. The diversity order scales multiplicatively with the aperture dimensions ($K^*_{2D} = K^*_x \cdot K^*_y$), demonstrating that modest 2D apertures can achieve high diversity orders.
• Comparative Analysis: A systematic comparison establishing the ordering of approximations: $P_{iid} \le P_{BCM} \le P_{exact} \le P_{EDoF} \le P_{Ref}$. This confirms that BCM/VBCM models are optimistic (underestimating outage), while the proposed EDoF/WIM models are conservative (overestimating outage).

Results
Numerical results via Monte Carlo simulations validate the analytical findings:

• Accuracy: The EDoF approximation tracks the exact outage probability closely, acting as a tight conservative upper bound. At high SNR, the ratio between the EDoF estimate and the exact outage converges to a finite constant (typically between 1.5 and 4 for practical apertures), confirming asymptotic tightness.
• Port Count Independence: Simulations confirm that increasing the number of ports $N$ beyond $N \approx 4K^*$ yields negligible diversity gains, as the additional ports merely oversample the existing $K^*$ spatial modes.
• Aperture Scaling: Outage probability decreases monotonically with aperture size $W$. The "staircase" behavior of the EDoF model reflects the discrete nature of $K^*$, where increasing $W$ slightly beyond an integer value activates additional spatial modes.
• FAMA Performance: In multi-user scenarios, the outage probability exhibits a floor at high SNR due to interference. The results show that increasing the aperture size (and thus $K^*$) is more effective at lowering this floor than increasing transmit power.
• 2D vs. 1D: A 2D planar FAS achieves significantly higher diversity orders than a 1D linear FAS of equivalent total aperture, offering substantial SNR gains for compact terminals.

Significance
The paper claims that the proposed framework resolves a critical gap in the FAS analytical toolkit. By providing a closed-form expression that is simultaneously easy to evaluate, provably conservative, and accurate in diversity order, the work enables reliable system dimensioning and performance certification. Unlike previous models that risk under-designing systems by underestimating outage, the EDoF and WIM approximations provide engineers with a safe, quantifiable margin for link budget analysis and resource allocation. The identification of $K^*$ as the fundamental limit of spatial diversity, determined by aperture size rather than port count, offers a clear design guideline: adding ports beyond the resolution of the aperture's eigenmodes yields diminishing returns, whereas increasing the physical aperture is the primary lever for improving reliability.