Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Enhancing Battlefield Awareness: An Aerial RIS-assisted ISAC System with Deep Reinforcement Learning

Hyunsang Cho, Seonghoon Yoo, Bang Chul Jung, Joonhyuk Kang

TL;DR

This work tackles battlefield ISAC under challenging propagation with blocked LoS, SI, and clutter. It deploys an aerial RIS (ARIS) to reconfigure the environment and uses a DRL framework (DDPG) to jointly optimize AP beamforming, ARIS trajectory, and RIS phase shifts under SINR constraints, while employing a NSP-based receive beamformer to suppress self-interference and clutter. The sensing objective is formulated as minimizing the CRB for target coordinates, with CRB derived from the FIM and measurement noise linked to the received ISAC echo SNR. Numerical results show the proposed DRL-based ARIS-ISAC system outperforms fixed-RIS and NSP-less baselines, achieving faster convergence and improved target localization accuracy, which enhances practical battlefield situational awareness.

Abstract

This paper considers a joint communication and sensing technique for enhancing situational awareness in practical battlefield scenarios. In particular, we propose an aerial reconfigurable intelligent surface (ARIS)-assisted integrated sensing and communication (ISAC) system consisting of a single access point (AP), an ARIS, multiple users, and a sensing target. With deep reinforcement learning (DRL), we jointly optimize the transmit beamforming of the AP, the RIS phase shifts, and the trajectory of the ARIS under signal-to-interference-noise ratio (SINR) constraints. Numerical results demonstrate that the proposed technique outperforms the conventional benchmark schemes by suppressing the self-interference and clutter echo signals or optimizing the RIS phase shifts.

Enhancing Battlefield Awareness: An Aerial RIS-assisted ISAC System with Deep Reinforcement Learning

TL;DR

This work tackles battlefield ISAC under challenging propagation with blocked LoS, SI, and clutter. It deploys an aerial RIS (ARIS) to reconfigure the environment and uses a DRL framework (DDPG) to jointly optimize AP beamforming, ARIS trajectory, and RIS phase shifts under SINR constraints, while employing a NSP-based receive beamformer to suppress self-interference and clutter. The sensing objective is formulated as minimizing the CRB for target coordinates, with CRB derived from the FIM and measurement noise linked to the received ISAC echo SNR. Numerical results show the proposed DRL-based ARIS-ISAC system outperforms fixed-RIS and NSP-less baselines, achieving faster convergence and improved target localization accuracy, which enhances practical battlefield situational awareness.

Abstract

This paper considers a joint communication and sensing technique for enhancing situational awareness in practical battlefield scenarios. In particular, we propose an aerial reconfigurable intelligent surface (ARIS)-assisted integrated sensing and communication (ISAC) system consisting of a single access point (AP), an ARIS, multiple users, and a sensing target. With deep reinforcement learning (DRL), we jointly optimize the transmit beamforming of the AP, the RIS phase shifts, and the trajectory of the ARIS under signal-to-interference-noise ratio (SINR) constraints. Numerical results demonstrate that the proposed technique outperforms the conventional benchmark schemes by suppressing the self-interference and clutter echo signals or optimizing the RIS phase shifts.
Paper Structure (12 sections, 16 equations, 4 figures, 1 table, 1 algorithm)

This paper contains 12 sections, 16 equations, 4 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. System model
  3. Set-up
  4. Channel model
  5. Communication Model
  6. Radar Sensing Model
  7. Receive beamforming design
  8. CRB Derivation of Target Coordinates
  9. Proposed ARIS-Assisted ISAC System with DRL
  10. Simulation result
  11. Conclusion
  12. Derivation of measurement covariance $\mathbf{C}(\mathbf{d}_s)$ and FIM ${\mathbf{J}}\left({\mathbf{d}}_{s}\right)$

Figures (4)

  • Figure 1: Illustration of the proposed ARIS-assisted ISAC systems.
  • Figure 2: Reward versus number of episodes.
  • Figure 3: The squared estimation error versus time slot. ($\Gamma_{th}=10$ dB).
  • Figure 4: Optimal ARIS trajectories of the proposed method according to the different SINR threshold $\Gamma_{th}$. (The coordinates of the estimated target $\tilde{x}_{s,l}, \tilde{y}_{s,l}$ is depicted when $\Gamma_{th}=10$ dB.