Abstract
As wireless communication moves toward 6G, millimeter-wave (mmWave) and Terahertz (THz) bands offer massive bandwidth but suffer from severe path loss and susceptibility to physical blockage. This paper addresses the security challenges inherent in these high-frequency bands by leveraging Intelligent Reflecting Surfaces (IRS). We propose a joint optimization framework for base station (BS) beamforming and IRS phase shifts to maximize the secrecy rate—the rate of secure data transmission. Our findings demonstrate that IRS can effectively bypass eavesdroppers, even in "beam blocking" scenarios, and that efficient algorithms like Block Coordinate Descent (BCD) can achieve near-optimal performance under realistic hardware constraints.
Introduction
The transition to mmWave and THz frequencies is a cornerstone of future 6G networks. However, the propagation characteristics of these bands—high directional gain but extreme sensitivity to obstacles—create unique security vulnerabilities. If an eavesdropper (Eve) physically blocks the direct line-of-sight between a base station (BS) and a legitimate user (Bob), she can intercept the signal while simultaneously degrading Bob's reception. This paper explores how an IRS—a "smart mirror" that can steer radio waves—can be used to create a secure, alternative path for data that bypasses such eavesdroppers.
System Model and Problem Formulation
The proposed system utilizes an IRS to create a controllable "spatial detour" for the signal. By adjusting the phase shifts of the IRS elements, the BS can steer the reflected beam toward Bob while minimizing the leakage toward Eve. The core challenge lies in the joint optimization of the BS's hybrid beamforming and the IRS's reflecting matrix, which is traditionally a complex non-convex problem.
Methodology
The authors simplify the optimization by proving that under rank-one channel conditions (typical for the line-of-sight nature of mmWave/THz), the BS beamforming design can be decoupled from the IRS configuration. This allows for a two-step approach:
- BS Beamforming: A closed-form solution is derived and implemented via an Orthogonal Matching Pursuit (OMP) algorithm to handle the hybrid (analog + digital) beamforming structure.
- IRS Optimization: Two methods are proposed for finding the best discrete phase shifts: a Semidefinite Programming (SDP) approach and an Element-wise Block Coordinate Descent (BCD) method. The BCD method is particularly notable for its low computational complexity, making it practical for real-time systems.
Key Results
Simulations conducted at 0.3 THz reveal that the proposed IRS-assisted strategy significantly outperforms "secure oblivious" baselines that do not account for the eavesdropper's presence.
As shown in Figure 3, the secrecy rate increases monotonically with the number of IRS elements ($N$). A larger IRS can generate narrower, more precise beams, effectively focusing energy on Bob and away from Eve. Furthermore, the study demonstrates that using just 3-bit discrete phase shifts (8 possible values) achieves performance nearly identical to "perfect" continuous phase shifts, validating the feasibility of the approach with realistic hardware. The BCD algorithm achieved nearly the same performance as an exhaustive search but with a fraction of the computational complexity.
Conclusion and Future Work
IRS technology provides a robust mechanism for securing high-frequency wireless links against physical blockages and eavesdropping. While this work assumes a rank-one channel and perfect channel state information (CSI), future research will need to address more complex scattering environments and the difficulties of estimating channels for passive IRS elements in multi-user scenarios.
Summary updated