Performance Limits of FRIS Systems in Nakagami-$m$ Fading

Technical Summary: Performance Limits of FRIS Systems in Nakagami-m Fading

Problem Statement
Fluid Reconfigurable Intelligent Surfaces (FRIS) have emerged as a promising technology to enhance wireless link reliability by dynamically selecting subsets of weakly correlated elements, thereby overcoming the spatial correlation limitations of conventional fixed-structure RIS. However, the performance analysis of FRIS-assisted systems remains challenging due to the complex sum-product structure of the cascaded channel. Existing literature largely relies on asymptotic arguments, Central Limit Theorem (CLT) approximations, or simplified fading assumptions (e.g., Rayleigh fading). These approaches often fail to provide rigorous performance guarantees, such as strict upper or lower bounds, and are limited in their applicability to general propagation environments characterized by Nakagami-m fading with arbitrary spatial correlation.

Methodology
The paper develops a rigorous analytical framework for FRIS-assisted wireless systems operating over arbitrarily correlated Nakagami-m fading channels. The methodology proceeds through the following steps:

1. System Modeling: The authors model a single-antenna base station communicating with a user terminal via an FRIS composed of $L_{tot}$ elements, where a subset of $L_{ON}$ elements is activated. The FRIS is modeled as a uniform planar array (UPA) with spatial correlation governed by an isotropic Jakes model. A correlation-aware sub-lattice selection strategy is employed to activate elements while adhering to a prescribed correlation threshold.
2. Weighted Cauchy–Schwarz (WCS) Framework: To handle the intractable sum-product structure of the effective cascaded channel gain ($S = \sum X_\ell Y_\ell$), the authors apply a weighted Cauchy–Schwarz inequality. This transforms the problem into deriving bounds for the product of two weighted sums of squared fading envelopes ($A_w$ and $B_w$).
3. Physically Consistent Correlation Model: A physically consistent model for correlated Nakagami-m fading is introduced. Instead of imposing correlation directly on the envelopes, the model represents Nakagami-m variables as sums of squared magnitudes of correlated complex Gaussian random variables. This allows the correlation structure to emerge naturally from the underlying scattering clusters.
4. Statistical Characterization: Using the Gaussian representation, the weighted sums $A_w$ and $B_w$ are expressed as sums of quadratic forms in correlated Gaussian vectors. Through eigenvalue decomposition, these are reformulated as sums of independent Gamma random variables. The probability density functions (PDFs) of these sums are derived using the Moschopoulos Gamma-sum representation.
5. Derivation of Bounds: The PDF and cumulative distribution function (CDF) of the product of the two Gamma-mixture variables are derived. By truncating the infinite series representations of these functions, the authors establish a rigorous finite-truncation lower bound for the CDF of the squared channel gain. This directly leads to a closed-form lower bound for the Outage Probability (OP).

Key Contributions

• Rigorous OP Lower Bounds: The paper derives the first strict lower bounds for the outage probability of an FRIS-aided wireless system under arbitrarily correlated Nakagami-m fading.
• Generalized Fading Model: Unlike prior works restricted to Rayleigh fading, this framework accommodates general Nakagami-m fading with arbitrary positive real parameters ($m_X, m_Y$), naturally including Rayleigh fading as a special case ($m=1$).
• Physically Consistent Correlation: The proposed correlation model relates the Gaussian correlation matrix to the induced Nakagami-envelope correlation in a physically consistent manner, avoiding heuristic assumptions.
• Tractable Closed-Form Expressions: The derived bounds are expressed in closed form involving Meijer G-functions and finite sums, making them computationally tractable for performance evaluation.
• Simplified i.i.d. Case: A simplified closed-form expression is provided for the independent and identically distributed (i.i.d.) scenario.

Numerical Results
Monte Carlo simulations were conducted for a $10 \times 10$ FRIS surface with varying numbers of active elements ($L_{ON} = 25, 36$). The results demonstrate:

• Accuracy: The proposed analytical lower bound closely matches the simulated outage probability, confirming the validity of the derivation.
• Correlation Awareness: A correlation-aware element selection strategy (Configuration 2) significantly outperforms a compact central selection (Configuration 1) by selecting weakly correlated elements, thereby improving spatial diversity.
• Comparison with Approximations: While CLT and Gamma-based approximations may align with simulation results over limited Signal-to-Noise Ratio (SNR) ranges, they fail to provide rigorous bounds (sometimes appearing above or below the true curve). In contrast, the proposed bound consistently remains below the true OP, providing a guaranteed performance floor. Notably, the proposed bound yields a tighter characterization than the CLT approximation in the high-SNR regime.

Significance and Claims
The authors claim that this work establishes fundamental benchmarks for the achievable system performance of FRIS systems. By avoiding heuristic and asymptotic distributional approximations, the proposed framework offers provable performance guarantees for arbitrary numbers of FRIS elements. The paper positions these results as a significant step toward a more rigorous analytical understanding of FRIS systems, moving beyond the limitations of existing Rayleigh-fading-only studies and approximation-based analyses. The derived bounds serve as a reliable reference for assessing the minimum attainable outage probability in realistic, correlated propagation environments.