Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AI-Limited Fluid Antenna-Aided Integrated Sensing and Communication Systems

Farshad Rostami Ghadi, Kai-Kit Wong, F. Javier Lopez-Martinez, Zhentian Zhang, Hyundong Shin, Christos Masouros

TL;DR

This paper addresses ISAC when the transmitter is constrained by an AI representation bottleneck and the receiver employs a fluid antenna system (FAS). It models the AI encoder as a Gaussian bottleneck Z = X + W_z with $N_z^* = \frac{P}{2^{C_{AI}}-1}$ and derives a capacity-distortion region where the AI noise degrades both communication and sensing at the selected port via effective SNRs $\Gamma_c^*$ and $\Gamma_s^*$. The analysis shows that the port-selection gain is limited by the physical length $W$ of the FAS, with the effective DoF given by the numerical rank $L'(W)$ of the Jakes correlation, and that increasing $W$ allows the AI-limited performance bounds to be approached as $W\to\infty$. A practical VIB encoder is proposed to realize the Gaussian bottleneck, together with training procedures, and numerical results validate the theory and illustrate the trade-offs between AI capacity, spatial DoF, and ISAC performance. Overall, the work demonstrates that FAS can compensate AI-induced limitations and recover near AI-limited joint performance in ISAC systems.

Abstract

This paper characterizes the fundamental limits of integrated sensing and communication (ISAC) when the transmitter is subject to an artificial intelligence (AI) representation bottleneck and the receiver employs a fluid antenna system (FAS). Specifically, the message is first encoded into an ideal Gaussian waveform and mapped by an AI encoder into a finite-capacity latent representation that constitutes the physical channel input, while the FAS receiver selects the port experiencing the most favorable channel conditions. We reveal that the AI bottleneck is equivalent to an additive representation noise, which reduces both the communication and sensing signal-to-noise ratios (SNRs) at the selected port. We then derive the resulting ISAC capacitydistortion region and establish tight converse and achievability bounds under general fading models, including Jakes-correlated channels. Leveraging the spatial degrees of freedom (DoF) characterization of the Jakes' model, we furthermore prove that the port-selection gain is fundamentally constrained by the physical length of the FAS region: the effective diversity order equals the numerical rank of the Jakes' correlation matrix and increases only with the FAS length. Consequently, enlarging the FAS length allows the selected-port SNR to approach the AI-imposed ceiling, driving the achievable communication rate and sensing mean square error (MSE) toward their AI-limited fundamental bounds. Numerical results corroborate the analysis and scaling laws.

AI-Limited Fluid Antenna-Aided Integrated Sensing and Communication Systems

TL;DR

This paper addresses ISAC when the transmitter is constrained by an AI representation bottleneck and the receiver employs a fluid antenna system (FAS). It models the AI encoder as a Gaussian bottleneck Z = X + W_z with and derives a capacity-distortion region where the AI noise degrades both communication and sensing at the selected port via effective SNRs and . The analysis shows that the port-selection gain is limited by the physical length of the FAS, with the effective DoF given by the numerical rank of the Jakes correlation, and that increasing allows the AI-limited performance bounds to be approached as . A practical VIB encoder is proposed to realize the Gaussian bottleneck, together with training procedures, and numerical results validate the theory and illustrate the trade-offs between AI capacity, spatial DoF, and ISAC performance. Overall, the work demonstrates that FAS can compensate AI-induced limitations and recover near AI-limited joint performance in ISAC systems.

Abstract

This paper characterizes the fundamental limits of integrated sensing and communication (ISAC) when the transmitter is subject to an artificial intelligence (AI) representation bottleneck and the receiver employs a fluid antenna system (FAS). Specifically, the message is first encoded into an ideal Gaussian waveform and mapped by an AI encoder into a finite-capacity latent representation that constitutes the physical channel input, while the FAS receiver selects the port experiencing the most favorable channel conditions. We reveal that the AI bottleneck is equivalent to an additive representation noise, which reduces both the communication and sensing signal-to-noise ratios (SNRs) at the selected port. We then derive the resulting ISAC capacitydistortion region and establish tight converse and achievability bounds under general fading models, including Jakes-correlated channels. Leveraging the spatial degrees of freedom (DoF) characterization of the Jakes' model, we furthermore prove that the port-selection gain is fundamentally constrained by the physical length of the FAS region: the effective diversity order equals the numerical rank of the Jakes' correlation matrix and increases only with the FAS length. Consequently, enlarging the FAS length allows the selected-port SNR to approach the AI-imposed ceiling, driving the achievable communication rate and sensing mean square error (MSE) toward their AI-limited fundamental bounds. Numerical results corroborate the analysis and scaling laws.
Paper Structure (20 sections, 2 theorems, 36 equations, 5 figures, 2 tables, 1 algorithm)

This paper contains 20 sections, 2 theorems, 36 equations, 5 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Context
  3. State-of-the-Art
  4. Motivation and Contributions
  5. Organization and Mathematical Notations
  6. System Model
  7. Pre-Channel AI Representation Bottleneck
  8. FAS-Equipped Receiver Model
  9. Performance Metrics
  10. Capacity-Distortion Region
  11. FAS Statistics and Finite-$L$ Performance
  12. Asymptotic Behavior: FAS length, Spatial DoF, and AI Compensation
  13. Spatial DoF and Selected Port Gain Saturation
  14. Consequences for AI-Limited ISAC
  15. VIB Encoder Design
  16. ...and 5 more sections

Key Result

Theorem 1

Consider an ISAC system with the AI representation constraint $I(X;Z)\le C_{\mathrm{AI}}$, employing the Gaussian representation model eq:test_channel, and having an FAS receiver using the selection rule in eq:selection_rule. The achievable communication rate $R$ and sensing distortion $D_s$ satisfy and where $N_z^\star$ denotes the minimum representation noise variance permitted by the AI bottle

Figures (5)

  • Figure 1: Information-theoretic abstraction of the AI-constrained FAS-aided ISAC system model.
  • Figure 2: Achievable communication rate $R$ versus AI-capacity $C_\mathrm{AI}$.
  • Figure 3: Sensing distortion $D_s$ versus AI-capacity $C_\mathrm{AI}$.
  • Figure 4: Joint rate-sensing trade-off for the proposed model when $W=8\lambda$.
  • Figure 5: Validation of theoretical and achieved performance.

Theorems & Definitions (2)

  • Theorem 1: Capacity-Distortion Region
  • Theorem 2: Length-Dependent Compensation of the AI Bottleneck