PHY5 – AI/ML in the PHY/PHY layer security
Wednesday, 7 June 2023, 11:00-12:30, Room J1
Session Chair: Christoph Lipps (German Research Center for Artificial Intelligence, DE)
Multi-Feature Physical Layer Authentication for URLLC Based on Linear Supervised Learning
Andreas Weinand (Rheinland-Pfälzische Technische Universität Kaiserslautern-Landau, Germany); Christoph Lipps (German Research Center for Artificial Intelligence, Germany); Michael Karrenbauer and Hans D. Schotten (University of Kaiserslautern, Germany)
Physical Layer Authentication (PLA) can be a lightweight alternative to conventional security schemes such as certificates or Message Authentication Codes (MACs) for secure message transmission within Ultra Reliable Low Latency Communication (URLLC) scenarios. Single features such as Received Signal Strength Indicator (RSSI) are however not providing sufficient authentication accuracy. Therefore, multifeature techniques for PLA are introduced within this work and evaluated using a Universal Software Radio Peripheral (USRP) based testbed in a mobile URLLC campus network scenario. Linear supervised classification is proposed for PLA and evaluated under different attacker scenarios. The results show promising authentication performances in most of the evaluated senarions and can be increased by the application of multi-feature authentication.
The Role of Physical Layer Security in Satellite-Based Networks
Rupender Singh (VTT Technical Research Centre of Finland, Finland); Ijaz Ahmad (VTT Technical Research Centre of Finland & VTT Technical Research Center of Finland, Finland); Jyrki M Huusko (VTT Technical Research Centre of Finland, Finland)
6G will revolutionize the world with a large amount of bandwidth, high data rates, and extensive coverage in remote and rural areas. These goals can only be achieved by integrating terrestrial networks with non-terrestrial networks. However, the integration of terrestrial and non- terrestrial networks are raising security concerns about malicious attacks on satellite-terrestrial links due to their openness. Physical layer security (PLS) has emerged as a good candidate to deal with security threats by exploring the randomness of wireless channels. In this direction, this paper reviews how PLS methods are implemented in satellite communications. Firstly, we discuss the ongoing research on satellite-based networks by highlighting the key points in the literature. Then, we revisit the research activities on PLS in satellite-based networks by categorizing the different system architectures. Finally, we highlight research directions and opportunities to leverage the PLS in future satellite-based networks.
Secrecy Energy Efficiency in PAPR-Aware Artificial Noise Scheme for Secure Massive MIMO
Idowu Iseoluwa Ajayi (Institut Supérieur d’Electronique de Paris, France); Yahia Medjahdi (IMT Nord Europe, France); Lina Mroueh (Institut Supérieur d’Electronique de Paris, France); Rafik Zayani (CEA-LETI, France); Fatima Kaddour (Agence Nationale des Frequences, France)
In this paper, we study the secrecy energy efficiency (SEE) in an artificial noise (AN)-aided secure massive multipleinput multiple-output (MIMO) scheme. The scheme uses instantaneous information to design a peak-to-average power (PAPR)- aware AN that simultaneously improves secrecy and reduces PAPR. High PAPR leads to non-linear in-band signal distortion and out-of-band radiation causing adjacent channel interference. To ensure optimal secrecy performance, high power amplifiers (HPAs) at the base station (BS) are backed off to operate in the linear region only. The amount of back-off needed to ensure linearity of the HPA has a direct impact on the energy efficiency of the system and by extension the SEE. For our scheme, the magnitude of this back-off is determined by the power allocation ratio between the data and AN. Hence, we propose an optimal power allocation ratio for the scheme. This is to ensure a good trade-off between the energy efficiency, security, and reliability of the system. Simulation results show a better SEE performance for our scheme compared to legacy massive MIMO schemes with or without random AN injection. Finally, we study the impact of spatially correlated Rayleigh fading on the proposed scheme.
Low-Complexity Neural Networks for Denoising Imperfect CSI in Physical Layer Security
Idowu Iseoluwa Ajayi (Institut Supérieur d’Electronique de Paris, France); Yahia Medjahdi (IMT Nord Europe, France); Lina Mroueh (Institut Supérieur d’Electronique de Paris, France); Olumide Okubadejo (ESIEE, France); Fatima Kaddour (Agence Nationale des Frequences, France)
Channel adaptation physical layer security (PLS) schemes are degraded when the channel state information (CSI) is imperfect. Imperfect CSI is due to factors such as noisy feedback, outdated CSI, etc. In this paper, we propose a lowcomplexity noisy CSI denoising scheme based on the autoencoder architecture of deep neural networks referred to as DenoiseSecNet. To further reduce complexity, we then propose a hybrid version (HybDenoiseSecNet) that combines a legacy denoising scheme and a shallow neural network to achieve a similar performance as DenoiseSecNet. Simulation results, in terms of bit error rate (BER), secrecy capacity, and normalized mean squared error (NMSE), show the performance improvement of our proposed scheme compared to conventional denoising schemes. Finally, we study the significant reduction in computational complexity of the proposed scheme compared to another neural network scheme.
Deep Learning-Supported Kriging for Interpolation of High-Resolution Indoor REMs
Friedrich Burmeister (Technische Universität Dresden, Germany); Alexandros Palaios and Philipp Geuer (Ericsson Research, Germany); Anton Krause (Technische Universität Dresden, Germany); Richard Jacob (TU Dresden, Germany); Philipp Schulz and Gerhard P. Fettweis (Technische Universität Dresden, Germany)
In future communications systems, precise locationinf ormation of users is a declared target. To improve the radio systems, spatial channel knowledge with the same local accuracy in form of precise Radio Environment Maps (REMs) is beneficial. Constructing REMs with channel measurements is not only costly but often not feasible for specific regions of interest. Consequently, it is necessary to construct REMs based on a limited number of observations. Kriging is typically used for interpolation in the literature. The solely distance-dependent semi-variogram inherently assumes an isotropic environment. However, radio environments, especially indoor, are not isotropic and modeling the directionality of the spatial correlation is not possible by means of a simple variogram function. That is why we propose to enhance the Kriging spatial interpolation by exchanging the semi-variogram model by a Deep Neural Network (DNN) to better describe the anisotropic channel correlations in real-world environments. Ordinary Kriging and our proposed approach are compared for different sampling resolutions and sampling methodologies, namely random and regular. Our proposed method improves the average accuracy and more importantly further increases the confidence in the provided predictions. Higher confidence in the prediction is a way to unlock the usage of such techniques for future communication networks.