Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Learned Intelligent Recognizer with Adaptively Customized RIS Phases in Communication Systems

Yixuan Huang, Jie Yang, Chao-Kai Wen, Shuqiang Xia, Xiao Li, Shi Jin

TL;DR

This work addresses embedding target recognition in RIS-aided wireless systems by jointly optimizing RIS phase shifts and neural network parameters. It introduces an LSTM-based recognizer that fuses successive measurements to adapt RIS configurations for the next moment, leveraging a physical channel model to tailor sensing patterns. The approach jointly optimizes hardware and software to achieve high recognition accuracy with limited sensing overhead, while maintaining near-full spectral efficiency for communication. Results show superior recognition performance over state-of-the-art methods with only negligible SE loss, and achievable gains with modest RIS sizes, enabling practical deployment.

Abstract

This study presents an advanced wireless system that embeds target recognition within reconfigurable intelligent surface (RIS)-aided communication systems, powered by cuttingedge deep learning innovations. Such a system faces the challenge of fine-tuning both the RIS phase shifts and neural network (NN) parameters, since they intricately interdepend on each other to accomplish the recognition task. To address these challenges, we propose an intelligent recognizer that strategically harnesses every piece of prior action responses, thereby ingeniously multiplexing downlink signals to facilitate environment sensing. Specifically, we design a novel NN based on the long short-term memory (LSTM) architecture and the physical channel model. The NN iteratively captures and fuses information from previous measurements and adaptively customizes RIS configurations to acquire the most relevant information for the recognition task in subsequent moments. Tailored dynamically, these configurations adapt to the scene, task, and target specifics. Simulation results reveal that our proposed method significantly outperforms the state-of-the-art method, while resulting in minimal impacts on communication performance, even as sensing is performed simultaneously.

Learned Intelligent Recognizer with Adaptively Customized RIS Phases in Communication Systems

TL;DR

This work addresses embedding target recognition in RIS-aided wireless systems by jointly optimizing RIS phase shifts and neural network parameters. It introduces an LSTM-based recognizer that fuses successive measurements to adapt RIS configurations for the next moment, leveraging a physical channel model to tailor sensing patterns. The approach jointly optimizes hardware and software to achieve high recognition accuracy with limited sensing overhead, while maintaining near-full spectral efficiency for communication. Results show superior recognition performance over state-of-the-art methods with only negligible SE loss, and achievable gains with modest RIS sizes, enabling practical deployment.

Abstract

This study presents an advanced wireless system that embeds target recognition within reconfigurable intelligent surface (RIS)-aided communication systems, powered by cuttingedge deep learning innovations. Such a system faces the challenge of fine-tuning both the RIS phase shifts and neural network (NN) parameters, since they intricately interdepend on each other to accomplish the recognition task. To address these challenges, we propose an intelligent recognizer that strategically harnesses every piece of prior action responses, thereby ingeniously multiplexing downlink signals to facilitate environment sensing. Specifically, we design a novel NN based on the long short-term memory (LSTM) architecture and the physical channel model. The NN iteratively captures and fuses information from previous measurements and adaptively customizes RIS configurations to acquire the most relevant information for the recognition task in subsequent moments. Tailored dynamically, these configurations adapt to the scene, task, and target specifics. Simulation results reveal that our proposed method significantly outperforms the state-of-the-art method, while resulting in minimal impacts on communication performance, even as sensing is performed simultaneously.
Paper Structure (19 sections, 14 equations, 6 figures, 1 table)

This paper contains 19 sections, 14 equations, 6 figures, 1 table.

Table of Contents

  1. Introduction
  2. System Model
  3. Signal and Channel Models for Communication
  4. Signal and Channel Models for Sensing
  5. Protocol Design and Spectral Efficiency Analysis
  6. Learned Recognizer with Adaptive RIS Phase Customization
  7. Overall Design of the Network
  8. Key Modules of the Network
  9. Network Training
  10. Comparison with Prior Studies
  11. Numerical Results
  12. Simulation Settings
  13. Results and Discussions
  14. Performance Comparison of Various RIS Phase Designs
  15. Influence of Distance, Transmit Power, and Training Data Size
  16. ...and 4 more sections

Figures (6)

  • Figure 1: Illustration of the proposed joint communication and sensing system with the aid of the RIS.
  • Figure 2: The proposed protocol with time-division RIS configurations.
  • Figure 3: The proposed NN architecture.
  • Figure 4: Comparison of $\eta$ with different RIS array sizes and configurations.
  • Figure 5: $\eta$ with respect to $D$, $P_{\text{t}}$, and $\rho$.
  • ...and 1 more figures