Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep-Learning-Based Channel Estimation for IRS-Assisted ISAC System

Yu Liu, Ibrahim Al-Nahhal, Octavia A. Dobre, Fanggang Wang

TL;DR

This work tackles channel estimation in IRS-assisted ISAC networks by introducing a DL-based framework with two specialized networks: SE-DNN at the BS for sensing-channel estimation ($\mathbf{A}$) and CE-DNNs at the UEs for BS–IRS–UE channels ($\mathbf{B}_k$). It designs input-output pairs and augments training data to robustly map pilot- and reflection-informed observations to accurate channel estimates, with offline training and online testing phases. Empirical results show substantial NMSE improvements over LS benchmarks across varied SNRs and system dimensions, demonstrating strong generalization (e.g., up to ~15 dB NMSE gain for $\mathbf{A}$ at $\text{NMSE}=10^{-1}$ and ~5 dB for $\mathbf{B}_k$). The approach enables more reliable CSI for IRS-assisted ISAC, facilitating enhanced beamforming and sensing performance in future networks.

Abstract

Integrated sensing and communication (ISAC) and intelligent reflecting surface (IRS) are viewed as promising technologies for future generations of wireless networks. This paper investigates the channel estimation problem in an IRS-assisted ISAC system. A deep-learning framework is proposed to estimate the sensing and communication (S&C) channels in such a system. Considering different propagation environments of the S&C channels, two deep neural network (DNN) architectures are designed to realize this framework. The first DNN is devised at the ISAC base station for estimating the sensing channel, while the second DNN architecture is assigned to each downlink user equipment to estimate its communication channel. Moreover, the input-output pairs to train the DNNs are carefully designed. Simulation results show the superiority of the proposed estimation approach compared to the benchmark scheme under various signal-to-noise ratio conditions and system parameters.

Deep-Learning-Based Channel Estimation for IRS-Assisted ISAC System

TL;DR

This work tackles channel estimation in IRS-assisted ISAC networks by introducing a DL-based framework with two specialized networks: SE-DNN at the BS for sensing-channel estimation () and CE-DNNs at the UEs for BS–IRS–UE channels (). It designs input-output pairs and augments training data to robustly map pilot- and reflection-informed observations to accurate channel estimates, with offline training and online testing phases. Empirical results show substantial NMSE improvements over LS benchmarks across varied SNRs and system dimensions, demonstrating strong generalization (e.g., up to ~15 dB NMSE gain for at and ~5 dB for ). The approach enables more reliable CSI for IRS-assisted ISAC, facilitating enhanced beamforming and sensing performance in future networks.

Abstract

Integrated sensing and communication (ISAC) and intelligent reflecting surface (IRS) are viewed as promising technologies for future generations of wireless networks. This paper investigates the channel estimation problem in an IRS-assisted ISAC system. A deep-learning framework is proposed to estimate the sensing and communication (S&C) channels in such a system. Considering different propagation environments of the S&C channels, two deep neural network (DNN) architectures are designed to realize this framework. The first DNN is devised at the ISAC base station for estimating the sensing channel, while the second DNN architecture is assigned to each downlink user equipment to estimate its communication channel. Moreover, the input-output pairs to train the DNNs are carefully designed. Simulation results show the superiority of the proposed estimation approach compared to the benchmark scheme under various signal-to-noise ratio conditions and system parameters.
Paper Structure (14 sections, 20 equations, 6 figures, 1 table, 1 algorithm)

This paper contains 14 sections, 20 equations, 6 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model
  3. Sensing Model
  4. Communication Model
  5. Proposed DL-based Estimation Approach
  6. Input-Output Pairs Design
  7. Design for Sensing Channel
  8. Design for Communication Channel
  9. Training Dataset Generation
  10. Proposed DL Estimation Framework
  11. Offline Training
  12. Online Testing
  13. Simulation Results
  14. Conclusion

Figures (6)

  • Figure 1: IRS-assisted ISAC system model.$\qquad$$\qquad$$\qquad$$\qquad$$\qquad$$\qquad$$\qquad$$\quad$
  • Figure 2: Pilot transmission protocol.$\qquad$$\qquad$$\qquad$$\qquad$$\qquad$$\qquad$$\qquad$$\qquad$$\qquad$
  • Figure 3: The proposed DL estimation framework: (a) Offline training and online testing phases, (b) Architectures of the SE-DNN and CE-DNN.
  • Figure 4: NMSE of S&C channel estimation versus SNR for$M=4$, $L=30$, and $K=3$.
  • Figure 5: NMSE of communication channel estimation versus $L$at$M=4$ and $K=3$for$\text{SNR}=5\,\rm dB$and$15\,\rm dB$.
  • ...and 1 more figures