Mobility-Aware Robust Design for Secure and Energy-Efficient Integrated Sensing and Communication Systems

Abstract

Integrated sensing and communication (ISAC) is regarded as a promising paradigm for next-generation wireless networks, due to its capability to simultaneously provide sensing and communication functionalities within a unified hardware and spectral resource framework. ISAC brings many new opportunities, one of which is sensing-enhanced physical-layer security (PLS) design. Specifically, the sensing signal can be utilized to localize the eavesdropper and serve as the artificial noise (AN) for information jamming toward the eavesdropper [1]. The echoes reflected from the eavesdropper can also be utilized to estimate its channel state information (CSI). However, the existing literature mainly focuses on the PLS design with a static eavesdropper. The mobile eavesdropper poses a greater security risk and requires a dynamic system design. This thesis develops a unified, mobility-aware, robust design framework to systematically study the sensing-assisted PLS design against the mobile eavesdropper. To achieve this, we need to detect and continuously track the eavesdropper. The sensing-induced error should be accounted for to meet the security requirement, and the system design should be tailored to counter the eavesdropper’s movement.

First, we start from the most basic building block: resolution-aware beam scanning for target detection (potential eavesdroppers). Distinguishing the closely located targets imposes restrictive constraint on beam pattern with beamwidth satisfying the resolution requirement. Traditional methods control beamwidth by varying the number of activated antennas, which is less flexible and sacrifices spatial degrees of freedom (DoFs). To overcome this obstacle, we design an ideal beam pattern satisfying the resolution requirement and restrict the beam pattern via beam pattern matching error to achieve the desired resolution requirement.

Second, to ensure a broad monitoring coverage, we explore the networked ISAC system. The existing works assume the designated sensing transceivers, which limits fully exploiting the macro-diversity provided by the distributed sensing nodes. To this end, we propose the BS mode selection and user association design, which reveals the benefits of collaborative sensing in exploiting the macro-diversity for sensing purposes.

Third, to provide a continuous secure communication service, we investigate the near-field PLS design against a mobile eavesdropper, where the sensing is used to track the eavesdropper. Due to the CSI of the passive eavesdropper is challenging to obtain, we propose to estimate the CSI via sensing results and evaluate the sensing-induced uncertainty for CSI estimation. Further, we employ sensing to track the eavesdropper and propose an integrated tracking and beamforming design. A Pareto optimization framework is developed to characterize the key system performance indicators.

Fourth, to adapt sensing performance to a dynamic environment, we investigate the implicit sensing performance requirement design for far-field PLS in networked ISAC systems. It is challenging to determine the sensing performance due to the changing environment with the mobile eavesdropper. Existing studies normally impose explicit sensing performance requirements without considering the varying communication conditions, which hinders the system from fully exploiting the synergy between sensing and communication. To address this issue, we implicitly formulate sensing performance into the information leakage rate, enabling the system to adaptively adjust sensing accuracy according to varying configurations. Further, to thoroughly investigate the impact of network structure on system performance and complexity, both centralized and decentralized system designs are considered.

Finally, to reduce location uncertainty for achieving energy-efficient design, we propose a two-phase mobility-aware design for wireless power transfer (WPT) with a mobile energy-harvesting receiver (EHR). Due to the movement of EHR and sensing estimation errors, directly performing WPT results in significant power transfer. To achieve an energy-efficient design, we propose a variable-length two-phase robust design to reduce the location uncertainty and improve the power transfer efficiency. In the first phase (sensing phase), the system performs collaborative sensing to obtain accurate location estimation. The sensing results are then utilized in the second phase (WPT phase) to transfer energy to EHRs. By jointly optimizing the time-splitting strategy and allocated resources in both phases, the system achieves a considerable power consumption reduction. We also show the existence of an optimal time-splitting ratio for given location uncertainty.

Overall, this thesis provides a unified framework for mobility-aware, robust, and secure ISAC system design, addressing several tightly coupled challenges: fundamental sensing–communication trade-offs, physical-layer security under user and eavesdropper mobility, and the complexity of networked ISAC architectures. We show that the sensing-refinement is necessary and the sensing accuracy should be carefully determined to maximize the system performance. Comprehensive simulations validate the effectiveness of the proposed mobility-aware design. The proposed methodology can be widely applied to various wireless services for mobile entities.

Date of Award	2026
Original language	English
Awarding Institution	The Hong Kong University of Science and Technology
Supervisor	Shenghui SONG (Supervisor)