Generative AI has received significant attention among a spectrum of diverse industrial and academic domains, thanks to the magnificent results achieved from deep generative models such as generative pre-trained transformers (GPT) and diffusion models. In this paper, we explore the applications of denoising diffusion probabilistic models (DDPMs) in wireless communication systems under practical assumptions such as hardware impairments (HWI), low-SNR regime, and quantization error. Diffusion models are a new class of state-of-the-art generative models that have already showcased notable success with some of the popular examples by OpenAI and Google Brain. The intuition behind DDPM is to decompose the data generation process over small “denoising” steps. Inspired by this, we propose using denoising diffusion model-based receiver for a practical wireless communication scheme, while providing network resilience in low-SNR regimes, non-Gaussian noise, different HWI levels, and quantization error. We evaluate the reconstruction performance of our scheme in terms of mean-squared error (MSE) metric. Our results show that more than 25 dB improvement in MSE is achieved compared to deep neural network (DNN)-based receivers. We also highlight robust out-of-distribution performance under non-Gaussian noise.

In this paper, a generalization of deep learning-aided joint source channel coding (Deep-JSCC) approach to secure communications is studied. We propose an end-to-end (E2E) learning-based approach for secure communication against multiple eavesdroppers over complex-valued fading channels. Both scenarios of colluding and non-colluding eavesdroppers are studied. For the colluding strategy, eavesdroppers share their logits to collaboratively infer private attributes based on ensemble learning method, while for the non-colluding setup they act alone. The goal is to prevent eavesdroppers from inferring private (sensitive) information about the transmitted images, while delivering the images to a legitimate receiver with minimum distortion. By generalizing the ideas of privacy funnel and wiretap channel coding, the trade-off between the image recovery at the legitimate node and the information leakage to the eavesdroppers is characterized. To solve this secrecy funnel framework, we implement deep neural networks (DNNs) to realize a data-driven secure communication scheme, without relying on a specific data distribution. Simulations over CIFAR-10 dataset verifies the secrecy-utility trade-off. Adversarial accuracy of eavesdroppers are also studied over Rayleigh fading, Nakagami-m, and AWGN channels to verify the generalization of the proposed scheme. Our experiments show that employing the proposed secure neural encoding can decrease the adversarial accuracy by 28%.

Digital healthcare services are rapidly evolving for new methodologies, including hospital-to-home (H2H) services and intelligent Internet-of-Medical-Things (IIoMT). The sixth generation (6G) technology is considered as the fabric that facilitates the realization of these technologies, creating a paradigm shift towards personalized e-health services. To deal with the high security requirements and the energy constraints of 6G-enabled e-health services, we propose a lightweight learning-based key generation scheme for a pair of wireless-powered nodes in a cooperative communication system, where the legitimate nodes and the intermediate node have low-cost hardware-impaired transceivers. We utilize an echo state network (ESN) to enhance the “randomness distillation” phase, in which the legitimate parties try to obtain a common source of randomness as the raw data for key agreement. The PHY-based observed data is passed to the ESN, containing a reservoir of sparsely connected neurons to compensate for observation mismatches caused by the unbalanced hardware impairments. The output of the ESN can then be utilized to extract the secret key between e-health endpoints. Numerical experiments verify the performance gain of our proposed echo-based approach, resulting in 50% less required inference time compared with a fully-connected neural network (FCNN). Moreover, a performance gain of about 32% in terms of mean-square error (MSE) is achieved compared with a conventional PHY-only scheme.

In this paper, a learning-aided content-based image transmission scheme is proposed, where a multi-antenna source wishes to securely deliver an image to a legitimate destination in the presence of randomly-distributed passive eavesdroppers (Eves). We take into account the fact that not all regions of an image have the same importance from the security perspective. Hence, we employ a hybrid method to realize both the error-free data delivery of public regions—containing less-important pixels; and an artificial noise (AN)-aided transmission scheme for securing the confidential packets. To reinforce system’s security, fountain-based packet delivery is also adopted, where the source node encodes images into fountain-like packets prior to sending them over the air. The secrecy is achieved when the legitimate destination correctly receives the entire source packets before Eves obtain the important regions, while conforming to the latency limits of the system. Accordingly, the secrecy performance of our scheme is characterized by deriving a closed-form expression for the quality-of-security (QoSec) violation probability. Moreover, our proposed image delivery scheme leverages a deep neural network (DNN) and learns to maintain optimized transmission parameters, while achieving a low QoSec violation probability. Simulation results are provided to illustrate that our proposed learning-assisted scheme outperforms the state-of-the-arts by achieving considerable gains in terms of security and delay requirement.

Wireless secret key generation (WSKG) allows efficient key agreement protocols for securing the sixth generation (6G) wireless networks. Nevertheless, due to external adversaries or internal impairments, WSKG schemes might become vulner-able during the randomness distillation, where the legitimate nodes try to observe their source of common randomness. In this paper, we investigate the WSKG scheme with legitimate parties suffering from hardware impairments (HIs), while an active adversary acts as a man-in-the-middle (MiM) via injecting fake pilot signals. We first utilize randomized pilots to overcome the MiM. We also leverage the concept of recurrent neural networks (RNNs) to further enhance the randomness distillation. More specifically, the long short-term memory networks (LSTMs)-as a well-established type of RNNs-are implemented to learn the long-term dependencies between the observations of legitimate parties. The achievable secret key rate (SKR) and the impact of MiM on system's performance are analyzed. Our numerical results verify the performance gain of our proposed learning-based approach compared with the state-of-the-art methods and provide useful insights on system design. We show that our RNN-based approach achieves 30% and 15% improvement in terms of observation mismatches compared with the naïve scheme and the fully-connected benchmarks, respectively.

In this article, we propose a lightweight security scheme for ensuring both information confidentiality and transmission resiliency in the Internet-of-Things (IoT) communication. A single-antenna transmitter communicates with a half-duplex single-antenna receiver in the presence of a sophisticated multiple-antenna-aided passive eavesdropper and a multiple-antenna-assisted hostile jammer (HJ). A low-complexity artificial noise (AN) injection scheme is proposed for drowning out the eavesdropper. Furthermore, for enhancing the resilience against HJ attacks, the legitimate nodes exploit their own local observations of the wireless channel as the source of randomness to agree on shared secret keys. The secret key is utilized for the frequency hopping (FH) sequence of the proposed communication system. We then proceed to derive a new closed-form expression for the achievable secret key rate (SKR) and the ergodic secrecy rate (ESR) for characterizing the secrecy benefits of our proposed scheme, in terms of both information secrecy and transmission resiliency. Moreover, the optimal power sharing between the AN and the message signal is investigated with the objective of enhancing the secrecy rate. Finally, through extensive simulations, we demonstrate that our proposed system model outperforms the state-of-the-art transmission schemes in terms of secrecy and resiliency. Several numerical examples and discussions are also provided to offer further engineering insights.

Unmanned Aerial Vehicles (UAVs) have found compelling applications in intelligent logistics, search and rescue as well as in air-borne Base Station (BS). However, their communications are prone to both channel errors and eavesdropping. Hence, we investigate the max-min secrecy fairness of UAV-aided cellular networks, in which Cooperative Rate-Splitting (CRS) aided downlink transmissions are employed by each multi-antenna UAV Base Station (UAV-BS) to safeguard the downlink of a two-user Multi-Input Single-Output (MISO) system against an external multi-antenna Eavesdropper (Eve). Realistically, only Imperfect Channel State Information (ICSI) is assumed to be available at the transmitter. Additionally, we consider a realistic total power constraint and guarantee the specific Quality of Service (QoS) requirements of the legitimate users. To handle the worst-case channel uncertainty of the legitimate users and an external Eve, we conceive a robust secure resource allocation algorithm, which maximizes the minimum worst-case secrecy rate of the legitimate users. Based on the CRS principle, the transmitter splits and encodes the messages of legitimate users into common as well as private streams and the user having stronger CSI is asked to help the cell-edge user by opportunistically forwarding its decoded common message. In contrast to the existing schemes adopted in the literature for ensuring secure transmission of the first cooperative phase only, in our proposed solution the common message has a twin-fold mission. Explicitly, apart from serving as the desired message, it also acts as Artificial Noise (AN) for drowning out Eve without consuming extra power. This is in stark contrast to the conventional AN designs. In the second phase, the pure AN is directed towards the Eve, deploying a robust Maximum Ratio Transmitter (MRT) beamformer at the UAV-BS. To solve the resultant non-convex optimization problem we resort to the Sequential Parametric Convex Approximation (SPCA) method together with a bespoke initialization algorithm to avoid any failure due to infeasibility. Our simulation results confirm that the proposed secure transmission scheme outperforms the existing cooperative benchmarkers.

In this paper, by taking into account the requirements of information secrecy and transmission resiliency, we present a comprehensive scheme enabling secure device-to-device (D2D) networks, where a single-antenna transmitter communicates with a half-duplex single-antenna receiver in the presence of a passive eavesdropper and an adversary jammer. Motivated by physical layer security techniques, artificial noise injection scheme is proposed to ensure communication secrecy. To improve the resiliency against jamming attack, the D2D nodes utilize the frequency hopping technique. Under this system model, we examine the achievable ergodic secrecy rate (ESR) by deriving a new closed-form expression. Furthermore, the optimal power allocation between the artificial noise and data signal is studied for maximizing the ESR. Numerical examples and discussions are provided to depict the efficiency of our proposed scheme compared with the state-of-the-arts.

Cooperative relaying can be introduced as a promising approach for data communication in the Internet of Things (IoT), where the source and the destination may be placed far away. In this paper, by taking a variety of realistic hardware imperfections (HWIs) and channels estimation errors (CEEs) into account, the secrecy performance of a three-hop cooperative network with a source, a destination and two consecutive amplify-and-forward (AF) relays is investigated. The relays are considered to be untrusted, i.e., while they are mandatory helpers for data transmission, they may overhear the received signals. We adopt the artificial noise injection scheme, to keep the source message secret from being captured by the untrusted relays. Given this system model, a novel closed-form expression is obtained in the high signal-to-noise ratio (SNR) regime for the ergodic secrecy rate (ESR) performance over Nakagami-m fading. Our simulation results highlight that the secrecy performance of the system is improved when the tolerable HWIs are distributed beneficially across the transmission and reception radio-frequency (RF) front-ends of each node. Our work reveals that unlike the ideal case, the realistic scenario of non-ideal hardware with CEEs faces with the secrecy rate ceiling. Finally, under a constraint on the total energy consumption which is applicable for battery-limited IoT equipment, we maximize the achievable secrecy rate. Our results highlight the importance of the destination's jamming cooperation and the first relay's role on the secrecy performance.

In this paper, we propose a light-weight jamming-resistant scheme for the Internet-of-Things (IoT) in 5G networks to ensure high-quality communication in a two-hop cooperative network. In the considered system model, a source communicates with a destination in the presence of an untrusted relay and a powerful multi-antenna adversary jammer. The untrusted relay is an authorized necessary helper who may wiretap the confidential information. Meanwhile, the jammer is an external attacker who tries to damage both the training and transmission phases. Different from traditional frequency hopping spread spectrum (FHSS) techniques that require a pre-determined pattern between communicating nodes, in our scheme, the source and destination enjoy the local observations of the two-hop channels. Then they exploit the measured channel as the source of common randomness to generate shared secret keys. By collecting multiple time slots into a frame, the sequence of channels observed in each frame is utilized to specify the adopted FHSS sequence in the next frame. Based on the derived FHSS sequence from the key generation phase, the source starts to transmit its message supporting by the the destination-assisted cooperative jamming (DACJ) technique which prevents the untrusted relay from discovering the secret message. For the mentioned system model, we present new closed-form expressions for characterizing the achievable secret key rate (SKR) and ergodic secrecy rate (ESR) to highlight the efficiency of our proposed scheme compared to the state-of-the-art. We next determine the optimal power allocation (OPA) between the pilot and data transmission phases that maximizes the ESR performance while escaping from jamming attack. Finally, several numerical examples and discussions are presented to gain engineering insights behind the studied communication scenario.