Performance Analysis of Movable Antenna Arrays

Technical Summary: Performance Analysis of Movable Antenna Arrays

Problem Statement
This paper addresses the performance analysis of continuous movable antenna (MA) arrays within the broader context of Fluid Antenna Systems (FAS). While existing literature has explored single fluid antennas and discrete multi-antenna configurations, there is a lack of thorough mathematical analysis regarding continuous multi-antenna arrays where elements move together along a line. Specifically, the authors aim to characterize the system performance of a linear MA array with fixed inter-element spacing that moves continuously along an axis perpendicular to its orientation. The primary focus is on the upper tail of the cumulative distribution function (CDF) of the received Signal-to-Noise Ratio (SNR), as this metric quantifies the system's ability to achieve high instantaneous data rates by optimizing antenna position.

Methodology
The study models a system with a single-antenna transmitter and a receiver equipped with a linear MA array consisting of $M$ fixed elements separated by a distance $\Delta$. The entire array moves within a continuous range $[0, T]$. The channel is modeled using a full, spatially coherent correlation model based on Jakes' model, where the correlation between antennas depends on their relative positions and the movement of the array.

To analyze the performance, the authors derive an asymptotically exact approximation for the upper tail of the SNR CDF. The methodology relies on a lower bound for the CDF of the supremum of the SNR process, $S^* = \sup_{0 \le t \le T} S(t)$. This bound is expressed as:
$$F_{S^*}(s_{th}) \ge P(S(0) \le s_{th}) - T \times \text{LCR}_{S(t)}(s_{th})$$
where $\text{LCR}_{S(t)}$ is the Level Crossing Rate (LCR) of the SNR process across a threshold $s_{th}$.

The analysis is conducted for two scenarios:

1. Uncorrelated Antennas: Where the separation $\Delta$ is large enough to treat elements as independent.
2. Correlated Antennas: Where $\Delta$ is small, requiring a non-separable two-dimensional correlation model. For this case, the authors derive a novel closed-form expression for the LCR, which is mathematically challenging due to the continuous nature of the movement and the full correlation modeling.

Key Contributions

1. System Model: The paper analyzes a multi-antenna FAS where a linear array of $M$ elements moves continuously along a perpendicular axis, distinct from single-fluid or discrete-port systems.
2. Asymptotically Exact Bounds: The authors provide asymptotically exact approximations for the upper tail of the SNR CDF for both correlated and uncorrelated array elements.
3. Novel LCR Expression: A significant contribution is the derivation of a new closed-form expression for the Level Crossing Rate (LCR) of the SNR under correlated array elements. This expression involves a complex integral formulation derived from the eigenvalues of specific matrices related to the correlation structure.
4. Performance Insights: The analysis yields insights into how antenna spacing and movable length affect system performance, specifically regarding the trade-off between spatial diversity and correlation gain.

Results
Numerical results, validated through Monte Carlo simulations, confirm the accuracy of the derived LCR expressions and the tightness of the CDF bounds in the upper tail. Key findings include:

• Array vs. Single Antenna: The proposed MA array outperforms both single fluid antenna (SFA) and fixed-array systems.
• Impact of Spacing ($\Delta$): Contrary to fixed arrays where larger spacing is often preferred to reduce correlation and avoid deep fades, the MA array benefits from reduced inter-element spacing. Smaller $\Delta$ leads to higher spatial correlation, allowing all antennas to simultaneously encounter favorable channel conditions when the array is positioned optimally. This "correlation gain" improves overall system performance.
• Impact of Movable Length ($T$): Increasing the movable length $T$ consistently improves performance by expanding the search space for optimal channel conditions.
• Trade-offs: For nearly static arrays (small $T$), there is a trade-off where increased spacing provides spatial diversity (beneficial for lower thresholds), while reduced spacing provides correlation gain (beneficial for high SNR regions). For large $T$, mobility becomes the dominant factor, and compact arrays (small $\Delta$) are superior as they ensure all elements align with high-power channel regions simultaneously.

Significance
The paper claims that its primary significance lies in filling the gap of a thorough mathematical analysis for continuous multi-antenna FAS configurations. By deriving the LCR for correlated elements and establishing tight bounds on the SNR CDF, the work provides a theoretical foundation for understanding the limiting performance of such systems. The results demonstrate that exploiting mobility in conjunction with compact antenna spacing can yield significant performance gains over traditional fixed arrays and single fluid antennas, particularly in maximizing instantaneous SNR for high data rate applications. The analysis is noted to be applicable to any correlation model satisfying specific smoothness conditions, though the Jakes' model is used for the specific numerical validation presented.