Secure Millimeter Wave Communications for Next Generation Networks

Next-generation (NextG) wireless networks will realize a new communication paradigm characterized by high spectral, energy, and cost efficiency, ultra-high data rates, ultra-low latency, dense connectivity, and advanced physical sensing. These advancements will foster a wide range of emerging applications, including Industry 5.0, connected autonomous vehicles, collaborative robotics, and augmented reality. To meet these ambitious targets, new advancements in all layers of wireless networks are necessary. Millimeter-wave (mmWave) communications and integrated and sensing (ISAC) systems stand out as keyenabling technologies at the physical layer. Despite the performance promises, the new capabilities pose novel security concerns which are the main focus of this dissertation. The first security challenge stems from the fundamental requirement for beam management in mmWave communications due to the directional nature of mmWave transmissions. Compared with omnidirectional systems, directional transmissions require continuous beam alignment and tracking. We investigate security vulnerabilities in beam management functions and demonstrate the feasibility of various beam stealing and denial-of-service attacks. We counter such attacks using non-malleable properties of the physical propagation environment, which allow detecting adversarial interference in beam management functions. We further investigate attacks against Integrated Sensing and Communication (ISAC) systems, where the dual use of waveforms for communication and sensing creates an expanded attack surface. We demonstrate that ISAC is susceptible to sophisticated sensing attacks, including target spoofing and target elimination, which can deceive the system into perceiving ghost targets or hiding real ones. As a countermeasure, we develop a moving target defense that randomizes key waveform parameters, thus making it difficult to launch effective online attacks. The introduced waveform randomization does not disrupt the communication function of ISAC, and remains compliant with relevant standards. As an alternative way to defend against active attacks, we advance radio frequency fingerprinting (RFF) identification and device distinction methods for the mmWave frequencies. The main idea is to detect unauthorized signal injections based on the intrinsic hardware imperfections imprinted on radio signals. We propose and train new deep learning networks that generalize the mmWave channel and isolate the the hardware variations for the purposes of device classification and device distinction. Finally, We utilize the secure primitives developed at the physical layer to bridge the gap between physical and digital trust in cyber-physical systems. We propose novel ISAC-enabled cyber-physical trust binding protocols for connected autonomous vehicles.