Reconfigurable intelligent surface-aided secret key generation using an autoencoder and K-means quantization

Abstract

In quasi-static wireless channel scenarios, the generation of physical layer keys faces the challenge of invariant spatial and temporal channel characteristics, resulting in a high key disagreement rate (KDR) and low key generation rate (KGR). To address these issues, we propose a novel reconfigurable intelligent surface (RIS)-aided secret key generation approach using an autoencoder and K-means quantization algorithm. The proposed method uses channel state information (CSI) for channel estimation and dynamically adjusts the reflection coefficients of the RIS to create a rapidly fluctuating channel. This strategy enables the extraction of dynamic channel parameters, thereby enhancing channel randomness. Additionally, by integrating the autoencoder with the K-means clustering quantization algorithm, the method efficiently extracts random bits from complex, ambiguous, and high-dimensional channel parameters, significantly reducing KDR. Simulations demonstrate that, under various signal-to-noise ratios (SNRs), the proposed method performs excellently in terms of KGR and KDR. Furthermore, the randomness of the generated keys is validated through the National Institute of Standards and Technology (NIST) test suite.

1. Introduction

With the rapid advancement of wireless communication technologies, ensuring secure data transmission has become an increasingly critical challenge. Unlike wired networks, wireless communication relies on electromagnetic waves, which are inherently exposed to potential eavesdropping and malicious interference. The openness of wireless transmission, coupled with the broadcast characteristics of radio frequency signals, significantly increases the risk of unauthorized interception, tampering, and data leakage. This vulnerability has spurred the demand for more robust encryption methods to safeguard communications in dynamic and hostile environments.

Traditional cryptographic approaches, while effective, often rely on computational complexity to secure information. However, they may be insufficient to counter evolving threats, particularly in resource-constrained environments. In response, physical layer key generation has emerged as a promising solution, leveraging the unique characteristics of the wireless channel to generate symmetric keys for encryption. One notable method, proposed by Mathur et al. (2008), capitalizes on the inherent properties of the wireless channel reciprocity, time-variability, and spatial decorrelation to generate highly secure and dynamic keys for data encryption. Reciprocity ensures that both communicating parties, typically referred to as Alice and Bob, experience correlated channel responses, allowing them to independently generate identical cryptographic keys without transmitting sensitive information over the air. Time-variability exploits the time-varying nature of the wireless environment to continuously update keys, enhancing their unpredictability and resilience against eavesdropping, even in the presence of fading and interference. Finally, spatial decorrelation ensures that an eavesdropper, located beyond a half-wavelength distance from the legitimate users, cannot reconstruct the shared key due to the uncorrelated nature of the channel at their positions.

However, the variability required for robust key generation can be limited in real-world applications, particularly in time-domain quasi-static environments where the wireless channel changes gradually. These scenarios often suffer from reduced time-domain randomness and smaller key update rates, making it difficult to maintain a high level of security and efficiency in key generation. Fig. 1 illustrates the challenges presented by such quasi-static environments, where maintaining sufficient channel randomness becomes a crucial concern.