Robust Beamforming for Near-Field STAR-RIS-Enabled ISCPT

Claude D'Amours; Francois Chan; Zahra Rostamikafaki

arxiv: 2605.15954 · v1 · pith:XDQN2EFNnew · submitted 2026-05-15 · 📡 eess.SP

Robust Beamforming for Near-Field STAR-RIS-Enabled ISCPT

Zahra Rostamikafaki , Francois Chan , Claude D'Amours This is my paper

Pith reviewed 2026-05-20 16:04 UTC · model grok-4.3

classification 📡 eess.SP

keywords STAR-RISnear-fieldISCPTrobust beamformingharvested powerimperfect CSIS-proceduresuccessive convex approximation

0 comments

The pith

Robust beamforming maximizes harvested power for near-field STAR-RIS ISCPT under imperfect CSI while meeting secrecy and sensing targets.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a STAR-RIS aided framework for near-field integrated sensing, communication, and power transfer that maximizes the energy harvested by users. It sets up a non-convex optimization problem with constraints on user communication rates, eavesdropper rates, and sensing beampattern gains, all under imperfect cascaded channel knowledge. Alternating optimization solves it by applying the S-procedure to robustify the semi-infinite constraints and using successive convex approximation with a penalty term to handle the passive coefficients of the surface. Simulations confirm higher harvested power than standard baselines while still satisfying the secrecy and beampattern requirements. Readers would care because future wireless networks must deliver power, data, and sensing simultaneously in realistic near-field settings where channel knowledge is always limited.

Core claim

The authors formulate a robust harvested power maximization problem for a near-field STAR-RIS-enabled ISCPT system subject to imperfect cascaded CSI, required user rate, eavesdropper tolerable rate, and minimum sensing beampattern gain. They solve it with alternating optimization: the S-procedure approximates the semi-infinite robust constraints and sequential rank-one constraint relaxation yields the active beamformer, while a penalty-based successive convex approximation updates the passive STAR-RIS coefficients. Simulations in the near field show notable gains in harvested power that meet all targets and outperform conventional baselines.

What carries the argument

Alternating optimization that uses the S-procedure to approximate semi-infinite robust constraints, sequential rank-one constraint relaxation for active beamforming, and a penalty-based successive convex approximation scheme for the passive STAR-RIS coefficients.

If this is right

The design achieves higher harvested power in near-field conditions while enforcing the secrecy and sensing requirements.
It remains feasible and effective despite errors in cascaded channel estimates.
It outperforms conventional far-field or non-robust beamforming methods in the evaluated scenarios.
It enables joint optimization of sensing, communication, and wireless power transfer using the same surface.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same alternating-optimization structure with robust approximations could apply to other reconfigurable-surface systems that combine sensing and power delivery.
Deployment in real environments might reduce the frequency of channel re-estimation needed for acceptable performance.
Testing the approach with user mobility or multiple simultaneous users would reveal how well the near-field gains hold up.

Load-bearing premise

The S-procedure and successive convex approximation provide accurate enough surrogates for the original non-convex robust problem so that the solution still works well in realistic near-field channels.

What would settle it

A simulation or measurement in which the obtained beamforming violates the secrecy rate or beampattern constraint, or produces lower harvested power than a non-robust baseline, under realistic near-field channel estimation errors.

Figures

Figures reproduced from arXiv: 2605.15954 by Claude D'Amours, Francois Chan, Zahra Rostamikafaki.

**Figure 1.** Figure 1: The Proposed ISCPT System Model. information transmission. The sets of IRs, ERs, and sensing targets are denoted by K , {1, . . . , K}, E , {1, . . . , E}, and T , {1, . . . , T }, respectively. As depicted in [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** Figure 2: Total harvested power vs. iteration index, maximum tr [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

read the original abstract

A simultaneously transmitting and reflecting reconfigurable intelligent surface (STAR-RIS)-aided near-field integrated sensing, communication, and power transfer (ISCPT) framework is proposed. We formulate a robust harvested power maximization problem under imperfect cascaded channel state information (CSI), with constraints on required user rate, eavesdropper tolerable rate, and minimum sensing beampattern gain. To address this non-convex problem, we adopt alternating optimization (AO). First, we approximate the semi-infinite inequality constraints using the S-procedure and obtain rank-one active beamforming via sequential rank-one constraint relaxation (SROCR); then we update the passive STAR-RIS coefficients with a penalty-based scheme refined by successive convex approximation (SCA). Simulations in the near field demonstrate notable gains in harvested power while meeting secrecy and beampattern targets, outperforming conventional baselines.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper applies AO with S-procedure and SCA to a new near-field STAR-RIS ISCPT setup under imperfect CSI, but the simulation claims rest on unexamined approximation gaps.

read the letter

The paper formulates a robust harvested-power maximization for near-field STAR-RIS aided ISCPT with constraints on user rate, secrecy rate, and sensing beampattern under imperfect cascaded CSI. It solves this with alternating optimization: S-procedure to convert the semi-infinite robust constraints to LMIs, SROCR to get rank-one active beamforming, and penalty-SCA to update the passive coefficients. That specific combination for this system model looks new relative to the baselines they cite, and the approach is a direct, workable extension of tools already common in RIS beamforming papers. The simulations reportedly show higher harvested power while still meeting the targets, which is the practical payoff they emphasize. The main limitation is that both the S-procedure and the SCA are conservative or locally optimal by construction. In the near-field regime the cascaded channels are quadratic in distance and angle, so the uncertainty set is not a plain Euclidean ball; the linearization can therefore leave noticeable sub-optimality or hidden infeasibility that the reported gains might not survive. The abstract gives no convergence guarantees, no error bars on the approximations, and no comparison against tighter or exact methods, so it is difficult to judge how much performance is left on the table. This is the kind of incremental optimization paper that wireless-systems researchers working on RIS and ISAC variants would look at for the formulation and the simulation setup. It is not foundational, but the problem is cleanly stated and the method is reproducible enough that a referee could check the derivations and ask for tighter validation of the approximations. I would send it to peer review rather than desk-reject it.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes a robust beamforming design for a near-field STAR-RIS-enabled integrated sensing, communication, and power transfer (ISCPT) system. It formulates an optimization problem to maximize harvested power under imperfect cascaded CSI, subject to constraints on user rate, eavesdropper rate, and sensing beampattern gain. The non-convex problem is addressed via alternating optimization, employing the S-procedure for semi-infinite constraints, sequential rank-one constraint relaxation for active beamforming, and penalty-based successive convex approximation for passive STAR-RIS coefficients. Simulations indicate notable improvements in harvested power over conventional baselines while satisfying the constraints.

Significance. If the approximations are sufficiently accurate, the work offers a practical method for robust design in emerging near-field ISCPT systems with STAR-RIS, which could have implications for secure and energy-efficient wireless networks. The combination of standard techniques like S-procedure and SCA provides a feasible computational approach, and the simulation results suggest potential performance gains in harvested power.

major comments (2)

[AO algorithm description (S-procedure step)] The S-procedure linearization of the robust secrecy and beampattern constraints (described in the abstract and the AO procedure) replaces the semi-infinite worst-case constraints over the uncertainty ball with finite LMIs. In the near-field cascaded channel model, where channel vectors depend quadratically on distance and angle, the uncertainty set is not a Euclidean ball; this can make the resulting LMI overly conservative or allow solutions that are infeasible for the original problem, directly affecting the validity of the reported harvested-power gains.
[Passive coefficient update (penalty-SCA step)] The penalty-SCA update for the passive STAR-RIS coefficients (abstract and corresponding algorithm subsection) replaces the non-concave objective with a first-order surrogate. No convergence proof or bound on the gap to the original non-convex problem is provided, and the central claim that the AO solution simultaneously meets rate, secrecy, and beampattern targets while outperforming baselines rests on this local approximation; without quantification of sub-optimality, the simulation gains cannot be confidently attributed to the true robust optimum.

minor comments (2)

[Abstract] The abstract mentions 'notable gains' but does not quantify them (e.g., in dB) or list the specific simulation parameters such as number of elements, SNR values, or uncertainty radii; adding these would improve readability.
[Notation and system model] Notation for cascaded channels and uncertainty sets should be introduced once and used consistently; minor inconsistencies appear in the transition between active and passive beamforming variables.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We address each major comment point by point below, providing clarifications and indicating revisions made to strengthen the presentation and address the concerns.

read point-by-point responses

Referee: [AO algorithm description (S-procedure step)] The S-procedure linearization of the robust secrecy and beampattern constraints (described in the abstract and the AO procedure) replaces the semi-infinite worst-case constraints over the uncertainty ball with finite LMIs. In the near-field cascaded channel model, where channel vectors depend quadratically on distance and angle, the uncertainty set is not a Euclidean ball; this can make the resulting LMI overly conservative or allow solutions that are infeasible for the original problem, directly affecting the validity of the reported harvested-power gains.

Authors: We appreciate the referee highlighting the modeling of the uncertainty set in the near-field context. In our formulation, the imperfect cascaded CSI is modeled as an additive error bounded within a Euclidean ball, which is a standard assumption in robust optimization literature for handling CSI uncertainty. Although the underlying near-field channel expressions involve quadratic dependence on distance and angle, the error is defined directly on the effective cascaded channel vector. The S-procedure is applied to this modeled set to obtain the LMIs. We acknowledge that this approximation may introduce conservatism. In the revised manuscript, we have added a clarification paragraph in Section III-B discussing the uncertainty modeling choice, its applicability to near-field scenarios, and potential conservatism. We have also included additional numerical verification showing that the obtained solutions satisfy the original rate, secrecy, and beampattern constraints for randomly sampled error realizations within the assumed ball. revision: partial
Referee: [Passive coefficient update (penalty-SCA step)] The penalty-SCA update for the passive STAR-RIS coefficients (abstract and corresponding algorithm subsection) replaces the non-concave objective with a first-order surrogate. No convergence proof or bound on the gap to the original non-convex problem is provided, and the central claim that the AO solution simultaneously meets rate, secrecy, and beampattern targets while outperforming baselines rests on this local approximation; without quantification of sub-optimality, the simulation gains cannot be confidently attributed to the true robust optimum.

Authors: We thank the referee for noting the lack of theoretical guarantees on the penalty-SCA step. The penalty-based successive convex approximation is employed to handle the non-convex unit-modulus constraints on the STAR-RIS coefficients by augmenting the objective with a penalty term and replacing the non-concave parts with first-order Taylor surrogates. While a rigorous proof of convergence to the global optimum or an explicit bound on the optimality gap is not provided (as is common for such non-convex problems involving AO and SCA), the algorithm exhibits monotonic improvement in the surrogate objective at each iteration and rapid convergence in practice. In the revised manuscript, we have expanded the discussion in the algorithm subsection to elaborate on the surrogate properties, the penalty update rule ensuring feasibility, and added a new figure illustrating the convergence behavior of the objective and constraint satisfaction across iterations. These additions support attributing the observed performance gains to the proposed robust design while meeting all constraints. revision: yes

Circularity Check

0 steps flagged

No significant circularity in the robust optimization framework

full rationale

The paper formulates a non-convex robust harvested-power maximization problem under imperfect CSI and solves it via alternating optimization that applies the S-procedure to convert semi-infinite constraints into LMIs, SROCR for rank-one beamforming, and penalty-SCA for the passive STAR-RIS coefficients. These are standard approximation techniques whose outputs are then evaluated in Monte-Carlo simulations against conventional baselines; the reported performance gains are empirical outcomes of the numerical experiments rather than quantities that reduce by construction to the problem statement or to any fitted parameter. No self-citations are invoked as load-bearing uniqueness theorems, no ansatz is smuggled through prior work, and no known empirical pattern is merely renamed. The derivation chain therefore remains self-contained and externally falsifiable through the simulation setup.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard assumptions about channel models and the validity of convex relaxations; no new entities are introduced.

axioms (2)

domain assumption Imperfect cascaded CSI can be modeled with bounded uncertainty sets suitable for S-procedure application.
Invoked when formulating the robust optimization problem in the abstract.
domain assumption Successive convex approximation and penalty-based schemes converge to feasible solutions for the passive beamforming subproblem.
Used in the update step for STAR-RIS coefficients.

pith-pipeline@v0.9.0 · 5676 in / 1304 out tokens · 46284 ms · 2026-05-20T16:04:26.503130+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[1]

Integrating sensing, communication, and po wer transfer: From theory to practice,

X. Li, Z. Han, G. Zhu, Y . Shi, J. Xu, Y . Gong, Q. Zhang, K. Hua ng, and K. B. Letaief, “Integrating sensing, communication, and po wer transfer: From theory to practice,” IEEE Communications Magazine , vol. 62, no. 9, pp. 122–127, 2024

work page 2024
[2]

Outage-con strained secrecy rate maximization for star-ris with energy-harves ting eavesdrop- pers and imperfect csi,

Z. Rostamikafaki, F. Chan, and C. D’Amours, “Outage-con strained secrecy rate maximization for star-ris with energy-harves ting eavesdrop- pers and imperfect csi,” IEEE Access , vol. 13, pp. 47 927–47 937, 2025

work page 2025
[3]

Terahertz band: The last piece of rf spectrum puzzle for com munication systems,

H. Elayan, O. Amin, B. Shihada, R. M. Shubair, and M.-S. Al ouini, “Terahertz band: The last piece of rf spectrum puzzle for com munication systems,” IEEE Open Journal of the Communications Society , vol. 1, pp. 1–32, 2020

work page 2020
[4]

Near-ﬁeld communications: A comprehensive survey,

Y . Liu, C. Ouyang, Z. Wang, J. Xu, X. Mu, and A. L. Swindlehu rst, “Near-ﬁeld communications: A comprehensive survey,” IEEE Commu- nications Surveys and Tutorials , vol. 27, no. 3, pp. 1687–1728, 2025

work page 2025
[5]

Model ing and analysis of near-ﬁeld isac,

B. Zhao, C. Ouyang, Y . Liu, X. Zhang, and H. V . Poor, “Model ing and analysis of near-ﬁeld isac,” IEEE Journal of Selected Topics in Signal Processing, vol. 18, no. 4, pp. 678–693, 2024

work page 2024
[6]

Simultaneou s wireless information and power transfer in near-ﬁeld communication s,

Z. Zhang, Y . Liu, Z. Wang, X. Mu, and J. Chen, “Simultaneou s wireless information and power transfer in near-ﬁeld communication s,” IEEE Internet of Things Journal , vol. 11, no. 16, pp. 27 760–27 774, 2024

work page 2024
[7]

Robust beamforming design for secure near-ﬁeld isac systems,

Z. Chen, F. Wang, G. Han, X. Wang, and V . K. N. Lau, “Robust beamforming design for secure near-ﬁeld isac systems,” IEEE Wireless Communications Letters , pp. 1–1, 2025

work page 2025
[8]

An-aided secure beamforming for elaa-swipt in mixed near-and far-ﬁeld,

Y . Yi, G. Zhang, M. Cui, C. Y ou, and Q. Wu, “An-aided secure beamforming for elaa-swipt in mixed near-and far-ﬁeld,” IEEE Wireless Communications Letters , pp. 1–1, 2025

work page 2025
[9]

Matrix-calibration -based cas- caded channel estimation for reconﬁgurable intelligent su rface assisted multiuser mimo,

H. Liu, X. Y uan, and Y .-J. A. Zhang, “Matrix-calibration -based cas- caded channel estimation for reconﬁgurable intelligent su rface assisted multiuser mimo,” IEEE Journal on Selected Areas in Communications , vol. 38, no. 11, pp. 2621–2636, 2020

work page 2020
[10]

Robust resourc e allocation for star-ris assisted swipt systems,

G. Zhu, X. Mu, L. Guo, A. Huang, and S. Xu, “Robust resourc e allocation for star-ris assisted swipt systems,” IEEE Transactions on Wireless Communications, vol. 23, no. 6, pp. 5616–5631, 2024

work page 2024
[11]

Joint beamfo rming for star-ris in near-ﬁeld communications,

H. Li, Y . Liu, X. Mu, Y . Chen, and P . Zhiwen, “Joint beamfo rming for star-ris in near-ﬁeld communications,” in Proc. GLOBECOM 2023 - 2023 IEEE Global Communications Conference , 2023, pp. 625–630

work page 2023
[12]

Boyd and L

S. Boyd and L. V andenberghe, Convex optimization . Cambridge university press, 2004

work page 2004
[13]

A sequential constr aint relaxation algorithm for rank-one constrained problems,

P . Cao, J. Thompson, and H. V . Poor, “A sequential constr aint relaxation algorithm for rank-one constrained problems,” in Proc. Eur . Signal Process. Conf. (EUSIPCO) , 2017, pp. 1060–1064

work page 2017
[14]

Sem ideﬁnite relaxation of quadratic optimization problems,

Z.-q. Luo, W.-k. Ma, A. M.-c. So, Y . Y e, and S. Zhang, “Sem ideﬁnite relaxation of quadratic optimization problems,” IEEE Signal Processing Magazine, vol. 27, no. 3, pp. 20–34, 2010

work page 2010

[1] [1]

Integrating sensing, communication, and po wer transfer: From theory to practice,

X. Li, Z. Han, G. Zhu, Y . Shi, J. Xu, Y . Gong, Q. Zhang, K. Hua ng, and K. B. Letaief, “Integrating sensing, communication, and po wer transfer: From theory to practice,” IEEE Communications Magazine , vol. 62, no. 9, pp. 122–127, 2024

work page 2024

[2] [2]

Outage-con strained secrecy rate maximization for star-ris with energy-harves ting eavesdrop- pers and imperfect csi,

Z. Rostamikafaki, F. Chan, and C. D’Amours, “Outage-con strained secrecy rate maximization for star-ris with energy-harves ting eavesdrop- pers and imperfect csi,” IEEE Access , vol. 13, pp. 47 927–47 937, 2025

work page 2025

[3] [3]

Terahertz band: The last piece of rf spectrum puzzle for com munication systems,

H. Elayan, O. Amin, B. Shihada, R. M. Shubair, and M.-S. Al ouini, “Terahertz band: The last piece of rf spectrum puzzle for com munication systems,” IEEE Open Journal of the Communications Society , vol. 1, pp. 1–32, 2020

work page 2020

[4] [4]

Near-ﬁeld communications: A comprehensive survey,

Y . Liu, C. Ouyang, Z. Wang, J. Xu, X. Mu, and A. L. Swindlehu rst, “Near-ﬁeld communications: A comprehensive survey,” IEEE Commu- nications Surveys and Tutorials , vol. 27, no. 3, pp. 1687–1728, 2025

work page 2025

[5] [5]

Model ing and analysis of near-ﬁeld isac,

B. Zhao, C. Ouyang, Y . Liu, X. Zhang, and H. V . Poor, “Model ing and analysis of near-ﬁeld isac,” IEEE Journal of Selected Topics in Signal Processing, vol. 18, no. 4, pp. 678–693, 2024

work page 2024

[6] [6]

Simultaneou s wireless information and power transfer in near-ﬁeld communication s,

Z. Zhang, Y . Liu, Z. Wang, X. Mu, and J. Chen, “Simultaneou s wireless information and power transfer in near-ﬁeld communication s,” IEEE Internet of Things Journal , vol. 11, no. 16, pp. 27 760–27 774, 2024

work page 2024

[7] [7]

Robust beamforming design for secure near-ﬁeld isac systems,

Z. Chen, F. Wang, G. Han, X. Wang, and V . K. N. Lau, “Robust beamforming design for secure near-ﬁeld isac systems,” IEEE Wireless Communications Letters , pp. 1–1, 2025

work page 2025

[8] [8]

An-aided secure beamforming for elaa-swipt in mixed near-and far-ﬁeld,

Y . Yi, G. Zhang, M. Cui, C. Y ou, and Q. Wu, “An-aided secure beamforming for elaa-swipt in mixed near-and far-ﬁeld,” IEEE Wireless Communications Letters , pp. 1–1, 2025

work page 2025

[9] [9]

Matrix-calibration -based cas- caded channel estimation for reconﬁgurable intelligent su rface assisted multiuser mimo,

H. Liu, X. Y uan, and Y .-J. A. Zhang, “Matrix-calibration -based cas- caded channel estimation for reconﬁgurable intelligent su rface assisted multiuser mimo,” IEEE Journal on Selected Areas in Communications , vol. 38, no. 11, pp. 2621–2636, 2020

work page 2020

[10] [10]

Robust resourc e allocation for star-ris assisted swipt systems,

G. Zhu, X. Mu, L. Guo, A. Huang, and S. Xu, “Robust resourc e allocation for star-ris assisted swipt systems,” IEEE Transactions on Wireless Communications, vol. 23, no. 6, pp. 5616–5631, 2024

work page 2024

[11] [11]

Joint beamfo rming for star-ris in near-ﬁeld communications,

H. Li, Y . Liu, X. Mu, Y . Chen, and P . Zhiwen, “Joint beamfo rming for star-ris in near-ﬁeld communications,” in Proc. GLOBECOM 2023 - 2023 IEEE Global Communications Conference , 2023, pp. 625–630

work page 2023

[12] [12]

Boyd and L

S. Boyd and L. V andenberghe, Convex optimization . Cambridge university press, 2004

work page 2004

[13] [13]

A sequential constr aint relaxation algorithm for rank-one constrained problems,

P . Cao, J. Thompson, and H. V . Poor, “A sequential constr aint relaxation algorithm for rank-one constrained problems,” in Proc. Eur . Signal Process. Conf. (EUSIPCO) , 2017, pp. 1060–1064

work page 2017

[14] [14]

Sem ideﬁnite relaxation of quadratic optimization problems,

Z.-q. Luo, W.-k. Ma, A. M.-c. So, Y . Y e, and S. Zhang, “Sem ideﬁnite relaxation of quadratic optimization problems,” IEEE Signal Processing Magazine, vol. 27, no. 3, pp. 20–34, 2010

work page 2010