Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Integrated Communication and Learned Recognizer with Customized RIS Phases and Sensing Durations

Yixuan Huang, Jie Yang, Chao-Kai Wen, Shi Jin

TL;DR

This work presents a RIS-aided wireless system that multiplexes downlink signals for environmental sensing and target recognition. An LSTM-based recognizer jointly optimizes RIS phases and NN parameters to adapt to prior scene and target information, while a learned decisioner dynamically terminates sensing to allocate a varying number of measurements per target. The approach yields significant improvements in recognition accuracy over state-of-the-art methods with only minor SE degradation, and it supports both continuous and discrete RIS phase implementations via differentiable training and temperature-based discretization. Extensive simulations on MNIST, Fashion-MNIST, and synthesized datasets demonstrate robustness to noise, distance, and SI, and reveal practical insights into RIS design and sensing-duration tradeoffs. The proposed framework offers a scalable, real-time pathway for integrated communication and sensing in future wireless networks, with open-source code available at the provided repository.

Abstract

Future wireless communication networks are expected to be smarter and more aware of their surroundings, enabling a wide range of context-aware applications. Reconfigurable intelligent surfaces (RISs) are set to play a critical role in supporting various sensing tasks, such as target recognition. However, current methods typically use RIS configurations optimized once and applied over fixed sensing durations, limiting their ability to adapt to different targets and reducing sensing accuracy. To overcome these limitations, this study proposes an advanced wireless communication system that multiplexes downlink signals for environmental sensing and introduces an intelligent recognizer powered by deep learning techniques. Specifically, we design a novel neural network based on the long short-term memory architecture and the physical channel model. This network iteratively captures and fuses information from previous measurements, adaptively customizing RIS phases to gather the most relevant information for the recognition task at subsequent moments. These configurations are dynamically adjusted according to scene, task, target, and quantization priors. Furthermore, the recognizer includes a decision-making module that dynamically allocates different sensing durations, determining whether to continue or terminate the sensing process based on the collected measurements. This approach maximizes resource utilization efficiency. Simulation results demonstrate that the proposed method significantly outperforms state-of-the-art techniques while minimizing the impact on communication performance, even when sensing and communication occur simultaneously. Part of the source code for this paper can be accessed at https://github.com/kiwi1944/CRISense.

Integrated Communication and Learned Recognizer with Customized RIS Phases and Sensing Durations

TL;DR

This work presents a RIS-aided wireless system that multiplexes downlink signals for environmental sensing and target recognition. An LSTM-based recognizer jointly optimizes RIS phases and NN parameters to adapt to prior scene and target information, while a learned decisioner dynamically terminates sensing to allocate a varying number of measurements per target. The approach yields significant improvements in recognition accuracy over state-of-the-art methods with only minor SE degradation, and it supports both continuous and discrete RIS phase implementations via differentiable training and temperature-based discretization. Extensive simulations on MNIST, Fashion-MNIST, and synthesized datasets demonstrate robustness to noise, distance, and SI, and reveal practical insights into RIS design and sensing-duration tradeoffs. The proposed framework offers a scalable, real-time pathway for integrated communication and sensing in future wireless networks, with open-source code available at the provided repository.

Abstract

Future wireless communication networks are expected to be smarter and more aware of their surroundings, enabling a wide range of context-aware applications. Reconfigurable intelligent surfaces (RISs) are set to play a critical role in supporting various sensing tasks, such as target recognition. However, current methods typically use RIS configurations optimized once and applied over fixed sensing durations, limiting their ability to adapt to different targets and reducing sensing accuracy. To overcome these limitations, this study proposes an advanced wireless communication system that multiplexes downlink signals for environmental sensing and introduces an intelligent recognizer powered by deep learning techniques. Specifically, we design a novel neural network based on the long short-term memory architecture and the physical channel model. This network iteratively captures and fuses information from previous measurements, adaptively customizing RIS phases to gather the most relevant information for the recognition task at subsequent moments. These configurations are dynamically adjusted according to scene, task, target, and quantization priors. Furthermore, the recognizer includes a decision-making module that dynamically allocates different sensing durations, determining whether to continue or terminate the sensing process based on the collected measurements. This approach maximizes resource utilization efficiency. Simulation results demonstrate that the proposed method significantly outperforms state-of-the-art techniques while minimizing the impact on communication performance, even when sensing and communication occur simultaneously. Part of the source code for this paper can be accessed at https://github.com/kiwi1944/CRISense.

Paper Structure

This paper contains 32 sections, 30 equations, 14 figures, 7 tables.

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. Protocol Design
  7. SE Analysis
  8. Recognizer with Adaptive RIS Phase Customization
  9. Overall Design of the NN
  10. Key Modules of the NN
  11. Training and Implementation
  12. Implementation of Discrete RIS Phase Shifts
  13. Enhanced Recognizer with Dynamic Sensing Durations
  14. NN Enhancement with a Learned Decisioner
  15. Training
  16. ...and 17 more sections

Figures (14)

  • Figure 1: Illustration of the proposed RIS-aided joint communication and recognition system.
  • Figure 2: The proposed protocol with time-division RIS configurations.
  • Figure 3: (a) The proposed NN based on the LSTM architecture; (b) Structure of the physical model and feature extraction module; (c) Details of the information processing module.
  • Figure 4: The proposed RIS phase generator under quantization constraints.
  • Figure 5: The modified information processing module with a decisioner.
  • ...and 9 more figures

Theorems & Definitions (2)

  • Remark 1
  • Remark 2