Biblio

Machine learning offers performance improvements and novel functionality, but its life cycle performance is understudied. In areas like cognitive communications, where systems are long-lived, life cycle trade-offs are key to system design. Herein, we consider the use of deep learning to classify spectrograms. We vary the label-space over which the network makes classifications, as may emerge with changes in use over a system’s life cycle, and compare heterogeneous transfer learning performance across label-spaces between model architectures. Our results offer an empirical example of life cycle challenges to using machine learning for cognitive communications. They evidence important trade-offs among performance, training time, and sensitivity to the order in which the label-space is changed. And they show that fine-tuning can be used in the heterogeneous transfer of spectrogram classifiers.

We investigate cognitive radio networks where the unlicensed sender operates in the overlay mode to relay the information of the licensed transmitter as well as send its individual information. To secure information broadcasted by the unlicensed sender against the wire-tapper, we invoke jammers to limit eavesdropping. Also, to exploit efficiently radio frequency energy in licensed signals, we propose the unlicensed sender and all jammers to scavenge this energy source. To assess the security measures of both licensed and unlicensed networks, we first derive rigorous closed-form formulas of licensed/unlicensed secrecy outage probabilities. Next, we validate these formulas with Monte-Carlo simulations before using them to achieve insights into the security capability of the proposed jamming-aided energy harvesting cognitive radio networks in crucial system parameters.

Recent work has shown the impact of adversarial machine learning on deep neural networks (DNNs) developed for Radio Frequency Machine Learning (RFML) applications. While these attacks have been shown to be successful in disrupting the performance of an eavesdropper, they fail to fully support the primary goal of successful intended communication. To remedy this, a communications-aware attack framework was recently developed that allows for a more effective balance between the opposing goals of evasion and intended communication through the novel use of a DNN to intelligently create the adversarial communication signal. Given the near ubiquitous usage of for-ward error correction (FEC) coding in the majority of deployed systems to correct errors that arise, incorporating FEC in this framework is a natural extension of this prior work and will allow for improved performance in more adverse environments. This work therefore provides contributions to the framework through improved loss functions and design considerations to incorporate inherent knowledge of the usage of FEC codes within the transmitted signal. Performance analysis shows that FEC coding improves the communications aware adversarial attack even if no explicit knowledge of the coding scheme is assumed and allows for improved performance over the prior art in balancing the opposing goals of evasion and intended communications.

Millimeter Wave (mmWave) band can be a solution to serve the vast number of Internet of Things (IoT) and Vehicle to Everything (V2X) devices. In this context, Cognitive Radio (CR) is capable of managing the mmWave spectrum sharing efficiently. However, Cognitive mmWave Radios are vulnerable to malicious users due to the complex dynamic radio environment and the shared access medium. This indicates the necessity to implement techniques able to detect precisely any anomalous behaviour in the spectrum to build secure and efficient radios. In this work, we propose a comparison framework between deep generative models: Conditional Generative Adversarial Network (C-GAN), Auxiliary Classifier Generative Adversarial Network (AC-GAN), and Variational Auto Encoder (VAE) used to detect anomalies inside the dynamic radio spectrum. For the sake of the evaluation, a real mmWave dataset is used, and results show that all of the models achieve high probability in detecting spectrum anomalies. Especially, AC-GAN that outperforms C-GAN and VAE in terms of accuracy and probability of detection.

Cognitive Radio Networks (CRN) are opportunistic networks which aim to harness the white space in the television frequency spectrum, on a need-to-need basis, without interfering the incumbent, called the Primary User (PU). Cognitive radios (CR) that sense the spectrum periodically for sensing the PU activity, are called Secondary Users (SU). CRNs are susceptible to two major attacks, Byzantine attacks and Primary User Emulation Attack (PUEA). Both the attacks are capable of rendering a CRN useless, by either interfering with the PU itself or capturing the entire channel for themselves. Byzantine attacks detection and mitigation is an important security issue in CRN. Hence, the current work proposes using a multi-Hidden Markov Model system with an aim to detect different types of random-Byzantine attacks. Simulation results show good detection rate across all the attacks.

Cognitive radio technology addresses the spectrum scarcity challenges by allowing unlicensed cognitive devices to opportunistically utilize spectrum band allocated to licensed devices. However, the openness of the technology has introduced several attacks to cognitive radios, one which is the spectrum sensing data falsification attack. In spectrum sensing data falsification attack, malicious devices share incorrect spectrum observations to other cognitive radios. This paper investigates the spectrum sensing data falsification attack in cognitive radio networks. We use the modified Z-test to isolate extreme outliers in the network. The q-out-of-m rule scheme is implemented to mitigate the spectrum sensing data falsification attack, where a random number m is selected from the sensing results and q is the final decision from m. The scheme does not require the services of a fusion Centre for decision making. This paper presents the theoretical analysis of the proposed scheme.

Software Defined Radio (SDR) can move the complicated signal processing and handling procedures involved in communications from radio equipment into computer software. Consequently, SDR equipment could consist of only a few chips connected to an antenna. In this paper, we present an implemented SDR testbed, which consists of four complete SDR nodes. Using the designed testbed, we have conducted two case studies. The first is designed to facilitate video transmission via adaptive LTE links. Our experimental results demonstrate that adaptive LTE link video transmission could reduce the bandwidth usage for data transmission. In the second case study, we perform UE location estimation by leveraging the signal strength from nearby cell towers, pertinent to various applications, such as public safety and disaster rescue scenarios where GPS (Global Position System) is not available (e.g., indoor environment). Our experimental results show that it is feasible to accurately derive the location of a UE (User Equipment) by signal strength. In addition, we design a Hardware In the Loop (HIL) simulation environment using the Vienna LTE simulator, srsLTE library, and our SDR testbed. We develop a software wrapper to connect the Vienna LTE simulator to our SDR testbed via the srsLTE library. Our experimental results demonstrate the comparative performance of simulated UEs and eNodeBs against real SDR UEs and eNodeBs, as well as how a simulated environment can interact with a real-world implementation.