Physical-layer security (PLS) is expected to play a crucial role in next-generation wireless

networks, where ultra-reliable and adaptive communication is required. Emerging

technologies, particularly reconfigurable intelligent surfaces (RIS), offer new opportunities

to enhance PLS by enabling programmable control of the wireless propagation environment.

Designing effective RIS-assisted PLS frameworks under dynamic and imperfect

channel conditions therefore becomes a critical research challenge.

In the first part of this thesis, we investigate an RIS-assisted secure uncrewed aerial

vehicle (UAV) communication scheme with multiple colluding eavesdroppers, where the

average secrecy rate (ASR) is enhanced through joint optimization of the UAV trajectory,

artificial noise (AN) and information beamforming, as well as RIS phase shifts. To tackle

the high complexity of the resulting non-convex multi-variable problem, a block coordinate

descent (BCD) framework combined with successive convex approximation (SCA) and

majorization–minimization (MM) techniques is developed. Simulation results demonstrate

significant ASR improvements over benchmark schemes.

In the second part, we propose a simultaneously transmitting and reflecting (STAR)-

RIS-enabled full-duplex (FD) secure communication scheme. The proposed design not

only suppresses self-interference (SI) to a level comparable to conventional self-interference

cancellation (SIC) techniques, but also intelligently controls the jamming power received

by the eavesdropper, thereby enhancing secrecy capacity. The transmission and reflection

phase shifts of the STAR-RIS are jointly optimized with beamforming vectors to simultaneously

facilitate SI suppression and jamming enhancement. Numerical results confirm

the superiority of the proposed scheme under different STAR-RIS operation modes.

In the third part, we develop a STAR-RIS-aided integrated sensing and communication

(ISAC) framework to address secure transmission in scenarios where the eavesdropper’s

channel state information (CSI) is unavailable. Artificial noise signals are leveraged

to simultaneously perform sensing and prevent information leakage, while a self-refining

mechanism is introduced to iteratively improve eavesdropper localization accuracy. The

transmit beamforming and STAR-RIS coefficients are jointly optimized to enhance sensing

performance, mitigate SI in FD operation, and strengthen jamming effectiveness. Simulation

results in both static and mobile scenarios demonstrate that the proposed design

significantly improves secrecy capacity through sensing-assisted channel acquisition.