The system model assumes perfect channel state information (CSI) for all links, including BS-to-STAR-RIS, STAR-RIS-to-user, and STAR-RIS-to-eavesdropper channels. In practice, however, the eavesdroppers (sensing targets) are passive and their CSI may only be partially or imperfectly known. How would the proposed joint beamforming and STAR-RIS coefficient optimization framework be extended to account for bounded or statistical CSI uncertainty, and what impact would such robust design have on the achievable secrecy-sensing trade-off demonstrated in Fig. 4(c)?
We agree that assuming perfect CSI for passive eavesdroppers (sensing targets) is idealized. In practice, their channel information is often imperfect or only partially known.
To address this, the proposed framework can be extended using a robust design approach, where the joint beamforming and STAR-RIS coefficients are optimized to remain secure under bounded or statistical CSI uncertainty. Specifically, the secrecy objective can be reformulated in a worst-case or outage-based manner, ensuring that the system maintains secrecy performance even under unfavorable channel realizations within a predefined uncertainty region.
Such a robust formulation would inevitably lead to a more conservative design. As a result, the achievable secrecy–sensing trade-off demonstrated in Fig. 4(c) would shift downward compared to the perfect-CSI case. In other words, for the same sensing requirement, the achievable secrecy rate would be reduced. However, the resulting performance would be significantly more reliable and practically meaningful, as it reflects guaranteed or high-confidence secrecy under realistic channel uncertainty.
