Related Experiment Video
Updated: Aug 15, 2025
07:56
A Photonic System for Generating Unconditional Polarization-Entangled Photons Based on Multiple Quantum Interference
Published on: September 5, 2019
8.6K
Intelligent Reflecting Surface-Assisted Physical Layer Key Generation with Deep Learning in MIMO Systems
Shengjie Liu1, Guo Wei2, Haoyu He1
1School of Computer Science, Sichuan University, Chengdu 610065, China.
Sensors (Basel, Switzerland)
|January 8, 2023
Summary
This study introduces a low-overhead physical layer secret key generation (PLKG) method using reconfigurable intelligent surfaces (IRS) in multi-input multi-output (MIMO) systems. It leverages deep learning to efficiently extract channel reciprocity, improving key generation rates and security.
Area of Science:
- Wireless Communications
- Information Security
- Artificial Intelligence
Background:
- Physical layer secret key generation (PLKG) is crucial for secure communications but is often limited by ideal environments and interference.
- Existing reconfigurable intelligent surface (IRS)-assisted PLKG schemes primarily focus on single-input-single-output (SISO) systems, which are insufficient for future multi-input multi-output (MIMO) networks.
- MIMO systems introduce significant overhead for channel reciprocity extraction, a challenge that needs addressing for effective PLKG.
Purpose of the Study:
- To propose a novel, low-overhead IRS-assisted PLKG scheme tailored for MIMO communication environments.
- To address the challenge of high overhead in channel reciprocity extraction within MIMO systems.
- To enhance the performance of PLKG in terms of key generation rate, error rate, and randomness.
Main Methods:
- Combining direct and IRS-reflecting channels to construct the channel response function.
- Developing a theoretically optimal interaction matrix to approximate the optimal achievable rate.
- Designing a deep learning model, IRS-CRNet (IRS-assisted Channel Reciprocity-learning Neural Network), for efficient channel reciprocity extraction in Time Division Duplexing (TDD) systems.
- Implementing a PLKG scheme based on the IRS-CRNet.
Main Results:
- The proposed IRS-CRNet effectively extracts channel reciprocity in MIMO systems.
- The IRS-assisted PLKG scheme demonstrates superior performance compared to existing methods.
- Simulation results validate the scheme's effectiveness in key generation rate, key error rate, and randomness.
Conclusions:
- The novel low-overhead IRS-assisted PLKG scheme using deep learning is effective for MIMO environments.
- IRS-CRNet significantly reduces the overhead associated with channel reciprocity extraction in TDD-MIMO systems.
- This approach enhances the security and efficiency of secret key generation in advanced wireless networks.
