Digital Twin Assisted Beamforming Design for Integrated Sensing and Communication Systems

Ahmed Alkhateeb; Shuaifeng Jiang

REVIEW 3 major objections 6 minor 12 references

Digital Twin Assisted Beamforming Design for Integrated Sensing and Communication Systems

T0 review · 3 major / 6 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A digital twin can guide ISAC beamforming to near-optimal sensing performance.

desk verdict A useful, clearly written DT-aided ISAC beamforming idea, but the near-optimal claim rests on an untested dominant-path heuristic and a self-referential simulation. read the letter →

arxiv 2412.07180 v1 pith:PIYWSMIE submitted 2024-12-10 eess.SP cs.ITmath.IT

Shuaifeng Jiang , Ahmed Alkhateeb This is my paper · ORCID

classification eess.SPcs.ITmath.IT

keywords digitaltwinintegratedsensingandcommunicationISACbeamformingMIMOraytracingNLoSchannel

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper sets out to show that a static digital twin—a 3D electromagnetic model of the environment plus ray tracing—can provide enough information about the sensing channel for a base station to design near-optimal joint communication and sensing beams, without ever estimating the full sensing channel. The proposed design traces paths from the base station to the candidate target position, predicts the directions and pre-target gains of line-of-sight and non-line-of-sight paths, and points the sensing beam along the strongest predicted path while keeping the communication user's SINR above a threshold. The paper reports that this digital-twin-assisted approach achieves sensing SNR close to an upper bound that knows the entire sensing channel, in both LoS-dominant and NLoS-dominant areas. A sympathetic reader would care because non-cooperative targets and unknown materials normally make sensing channel acquisition a chicken-and-egg problem, and this result suggests the environment model alone can stand in for that channel.

What carries the argument

The load-bearing mechanism is the digital twin's ray-traced prediction of partial path parameters: for each BS-to-target path it gives the departure angles and the pre-target complex gain $\beta_{l,1}$, and the algorithm takes the largest $\beta_{l,1}$ as a proxy for the dominant total sensing path. That single predicted direction converts the unknown sensing channel matrix $H_t$ into a known rank-one array-response term, which makes the non-convex beamforming problem solvable through semidefinite relaxation. A second mechanism is the reuse of the communication signal for sensing: the optimization objective includes both $\|H_t^H f_u\|^2$ and $\|H_t^H f_t\|^2$, so the communication beam contributes to sensing, while the sensing beam is shaped to place a null toward the user to control interference.

What would settle it

Take a target whose scattering is strongly anisotropic—for example, a flat plate oriented so it reflects best along a path with smaller $\beta_{l,1}$—and compare the sensing SNR of the digital-twin design with the full-channel upper bound; if the gap is much larger than the few dB shown for the spherical target, the dominant-partial-path selection is wrong.

Watch

Extended reading notes

Core claim

The central claim is that the joint beamforming problem for a MIMO ISAC base station—one communication user, one point-like sensing target, unknown sensing channel—can be nearly solved from the target position and a static digital twin. The digital twin ray-traces from the base station to the target position, returning for each propagation path the departure angles $\phi_l^{\mathrm{AoD}}, \theta_l^{\mathrm{AoD}}$ and the partial path gain $\beta_{l,1}$ accumulated before the signal hits the target. Because the target's own scattering gain and the return-path gain are unknown, the design selects the path with the largest $\beta_{l,1}$ as the dominant sensing direction and optimizes the sensing SNR along that direction, subject to a minimum communication SINR, using semidefinite relaxation. The paper argues that this is sufficient: in simulation with a spherical target behind a concrete wall, the digital-twin design matches the full-channel upper bound in both the LoS-dominant and the NLoS-dominant area, while the communication user's SINR constraint is met.

Load-bearing premise

The load-bearing assumption is that the sensing path with the largest gain before it reaches the target is also the path that dominates the total sensing channel; the paper ignores the target's unknown scattering pattern and the return-path gain, and the spherical target in the simulation reflects equally in all directions, so this assumption is not tested adversarially.

Editorial extensions

If this is right

If the claim holds, ISAC systems can sense targets that are not in line of sight using only the target position and a static environment map, removing the need for pilot-based sensing channel estimation.
The digital-twin beamforming design can be implemented with standard convex optimization, so it is a practical drop-in for the full-channel upper-bound formulation when the channel is unknown.
The same partial ray tracing can identify which environment object, such as a glass wall, dominates target illumination, pointing to where richer digital-twin updates would help most.
The result implies that the bottleneck for NLoS ISAC sensing shifts from channel acquisition to the accuracy of the digital twin's 3D geometry and ray-tracing model.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Editorial inference: the single-dominant-path selection is only as good as the correlation between $\beta_{l,1}$ and total path gain; a natural extension is to combine the top-$K$ predicted paths with weights, which should be more robust to anisotropic target scattering.
Editorial inference: the same digital-twin partial channel information could be used for beam tracking or for choosing between sensing and communication waveforms, not only for a fixed joint beam.
Editorial inference: testing the design with non-spherical targets, such as flat plates or corner reflectors, in simulation or measurement would directly probe the weakest assumption; the spherical target in the paper is a favorable isotropic scatterer.
Editorial inference: because the twin is static, moving targets or changing environments require periodic map updates, and the cost of keeping the EM 3D model fresh is an open system-level question the paper does not address.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

Read Jiang and Alkhateeb. The new bit is real: using a digital twin's ray-traced partial path gains to pick the dominant NLoS sensing direction and then steering the joint sensing/communication beamformer there, with no target cooperation and no full sensing channel knowledge. That is a sensible way around a known ISAC bottleneck, and the formulation (max sensing SNR subject to comm SINR, SDR baseline) is competent. The paper is easy to read and mostly honest about what the twin can and cannot provide.

The soft spots are in proportion. First, the central selection heuristic in Sec V-B: choose the path with the largest pre-target partial gain β_l,1, relying on the intuition that higher β_l,1 means higher total path gain. Total gain is α_l = α_l,jl β_l,1 β_l,2, so the ranking ignores the target scattering amplitude and the post-target return. The paper says this is reasonable when target EM information is unavailable, which is fair, but the simulation then validates it only with a 1-meter sphere, an isotropic scatterer. For aspect-dependent targets the ranking can flip. That is a load-bearing weakness, because even with a perfect digital twin the beam could point at the wrong path. Second, the evaluation is self-referential: the same ray tracer produces both the twin's prediction and the ground-truth sensing channel. So the numbers measure performance under a perfect model, and external twin error is untested. For a digital twin paper, that is a significant limitation. Third, the "near-optimal" claim is relative to the full-channel SDR heuristic, which is itself only a relaxed upper bound; the paper should say "near the full-channel SDR baseline," not "near-optimal." I would not call these fatal; the idea survives, but the advertised claim needs narrowing.

The stress-test concern about the sphere holds up on reading. The paper's own limitation sentence is not enough; they need at least one adversarial target model (plate, vehicle, dihedral) or an analytic counterexample to test the ranking heuristic.

Verdict: this deserves a serious referee. A referee should ask for a robustness study and a reworded claim. I would not cite it today except as an example of DT-aided sensing; after the robustness fix it would be citable. Bring to reading group? Maybe, mostly as a case study in self-referential validation.

Recommendation: send to review, expect major revision, require the adversarial target and noisy-twin experiments before acceptance.

Referee Report

3 major / 6 minor

Summary. The paper studies digital-twin-assisted beamforming for a MIMO ISAC system in which a base station serves a communication user and senses a target with the same transmitted signal. The authors formulate a joint optimization that maximizes sensing SNR subject to a minimum communication SINR, propose an SDR-based full-channel baseline and a LoS-direction baseline, and then introduce a digital-twin-based design. The digital twin runs ray tracing on an EM 3D model to predict the directions and partial gains of all sensing paths, selects the path with the largest pre-target partial gain, and optimizes the beams to concentrate sensing power along that direction. The paper evaluates the approach in an indoor scenario with a spherical target, using ray-tracing-generated channels, and reports that the proposed design attains near-optimal sensing SNR relative to the full-channel baseline in both LoS- and NLoS-dominant areas.

Significance. The core idea is timely and potentially valuable: if digital-twin ray tracing can provide the dominant sensing path direction without full sensing-channel knowledge, then ISAC beamforming can avoid the difficult pilot-based sensing-channel acquisition step. The manuscript is clearly written, the optimization formulation is standard, and the comparison against a full-channel SDR baseline and a LoS-direction baseline is appropriate. The simulation setup is reproducible in principle since it uses DeepMIMO and ray tracing. However, the central claim of near-optimality depends on two premises that are not tested: the dominant-partial-path heuristic and the assumption that the digital twin is an exact replica of the simulated environment. The evaluation also compares against an SDR-based baseline that is not guaranteed to be a true upper bound. These issues are load-bearing for the paper's main claim and require additional experiments or a reframed conclusion.

major comments (3)

[VI-B and V-A] The evaluation is circular with respect to the claim that the digital twin can approximate the real environment. The sensing channels used as ground truth are generated by ray tracing on the same EM 3D model that is also used to compute the digital twin's partial path information and dominant direction. That is, the simulation assumes a perfect digital twin; it never tests how prediction errors in geometry, material properties, or ray-tracing fidelity affect the selected direction and the resulting SNR. Please add experiments with a mismatched digital twin, e.g., perturbed object positions or material constants, or a different EM solver for ground truth, to quantify the sensitivity of the proposed approach to digital-twin inaccuracy.
[V-B, Eq. (8)] The dominant-partial-path heuristic is not justified by the path-gain model. Equation (8) gives the total path gain as α_l = α_l,jl β_l,1 β_l,2, yet the algorithm ranks paths using only β_l,1; it ignores the target's scattering amplitude α_l,jl and the post-target gain β_l,2. The paper states this as an intuitive assumption and notes it is reasonable without target shape/material information, but the simulation uses a 1-meter spherical target, whose scattering is aspect-independent. This is a favorable case that never challenges the ranking. Please evaluate the algorithm with aspect-dependent target models, such as flat plates or corner reflectors, or with randomly perturbed α_l,jl and β_l,2 values, and report how often the selected direction differs from the true dominant path and the corresponding SNR loss.
[IV-A] The 'full sensing channel' baseline is called an upper bound, but after semidefinite relaxation and the SVD-based rank-1 approximation in (17), the constructed beams are not guaranteed to be optimal for the original problem; only the relaxed objective value is an upper bound. Therefore the comparison in Section VI establishes near-optimality relative to an SDR-based benchmark, not necessarily relative to the true optimum of (12). Please either report the SDR relaxation gap for the simulated scenario, or replace the 'upper bound' terminology with 'SDR-based full-channel benchmark' and soften the near-optimal claim accordingly.

minor comments (6)

[II-C, Eq. (6) and Eq. (11)] There is a dimension mismatch in the sensing model: H_t should be of dimension N_r × N_t under the signal model y_t = H_t x, and the SNR expression in (11) should use ∥H_t f_u∥^2 and ∥H_t f_t∥^2 rather than H_t^H f_u and H_t^H f_t.
[IV-A, Eq. (14)] The notation Qu = hh^H is incomplete; it should be Q_u = h_u h_u^H to match the communication channel defined in Section II-B.
[I] The introduction contains typos such as 'his paper' and 'sensing sensing SNR'; please correct these in revision.
[VI-C, Fig. 3] The description of the beam patterns is confusing: the text says 'the main lobes of the communication beams point toward the directions of the communication user and the sensing target,' but Figure 3 appears to show one communication beam and one sensing beam. Please clarify which beam points to which direction and which beam forms the null.
[VI-C] The criterion for classifying the sensing-target area into LoS-dominant and NLoS-dominant regions is not defined. Please state how the partition is obtained, for example by comparing the LoS path power with the strongest NLoS path power.
[VI-C, Fig. 4] The term 'near-optimal' is used without a quantitative definition. Please report the median or mean SNR gap in dB between the proposed approach and the full-channel benchmark, in addition to the CDF plots.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the dominant-path heuristic is explicitly assumed and the beamforming optimization is independent of the ground-truth sensing channel once the direction is supplied.

full rationale

The paper's derivation chain is not circular at the equation level. The proposed design first obtains a dominant sensing path direction from partial ray tracing (Section V-B), then substitutes that direction into the beamforming optimization via Qt = a*(phi_star)a^T(phi_star) (Eq. 21) and solves (16a). This is an approximation, not an identity: the ground-truth sensing channel H in (6) includes target scattering αl = αl,jl βl,1 βl,2 (Eq. 8), while the digital twin provides only βl,1 and angles. The paper explicitly acknowledges the heuristic—'a sensing path with a higher partial path gain βl,1 is more likely to exhibit a greater total path gain'—and does not define the total gain in terms of the partial gain. The evaluation uses full ray tracing with a spherical target as ground truth and partial ray tracing to generate the twin's prediction; both share the same static environment model and simulator, which means the twin's environment approximation error is zero in simulation and the spherical target is a favorable scattering model. That is a limitation on external validity, but it is not a circular reduction because the predicted direction is not extracted from the ground-truth channel nor fitted to the sensing SNR. Self-citations ([2], [7], [9]–[11]) are contextual references to prior digital-twin and ISAC work; they are not invoked as unverified load-bearing evidence or as a uniqueness theorem. Therefore, under the quoted-equation standard, no circular step is present.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

The central claim depends on several domain assumptions: a point-reflector target, perfect self-interference cancellation, known communication channel, and a digital twin that is accurate in the simulation. The most fragile is the heuristic that the strongest partial path is the strongest total path, which is untested for non-isotropic targets.

assumptions (7)

domain assumption Sensing target is a single-point reflector
Section II-C: 'For the sensing target, we model it as a single-point reflector, as commonly adopted in the literature.'
domain assumption Perfect self-interference cancellation at the BS
Section II: 'we assume that the BS can perfectly cancel the self-interference'.
domain assumption Communication channel hu is known at the BS
Section III: 'we can assume that the communication channel is known at the BS for simplicity.'
domain assumption Digital twin ray tracing provides accurate path directions and partial gains
Section V-B: the digital twin is used to trace paths and compute βl,1, with no modeling error considered.
domain assumption The path with the highest partial path gain βl,1 is the dominant total path
Section V-B: 'a sensing path with a higher partial path gain βl,1 is more likely to exhibit a greater total path gain.' This is an unvalidated heuristic.
standard math Ray tracing channel model: sum of L paths with angles and gains
Equation (6), standard geometric channel model.
standard math Semidefinite relaxation and rank-1 approximation
Section IV-A, standard SDR technique.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Digital Twin Assisted Beamforming Design for Integrated Sensing and Communication Systems." pith.science (2026). https://pith.science/paper/PIYWSMIE

@misc{pith2026241207180,
  author       = {Pith},
  title        = {Pith review of: Digital Twin Assisted Beamforming Design for Integrated Sensing and Communication Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PIYWSMIE}},
  note         = {Machine review of arXiv:2412.07180}
}

read the original abstract

This paper explores a novel research direction where a digital twin is leveraged to assist the beamforming design for an integrated sensing and communication (ISAC) system. In this setup, a base station designs joint communication and sensing beamforming to serve the communication user and detect the sensing target concurrently. Utilizing the electromagnetic (EM) 3D model of the environment and ray tracing, the digital twin can provide various information, e.g., propagation path parameters and wireless channels, to aid communication and sensing systems. More specifically, our digital twin-based beamforming design first leverages the environment EM 3D model and ray tracing to (i) predict the directions of the line-of-sight (LoS) and non-line-of-sight (NLoS) sensing channel paths and (ii) identify the dominant one among these sensing channel paths. Then, to optimize the joint sensing and communication beam, we maximize the sensing signal-to-noise ratio (SNR) on the dominant sensing channel component while satisfying a minimum communication signal-to-interference-plus-noise ratio (SINR) requirement. Simulation results show that the proposed digital twin-assisted beamforming design achieves near-optimal target sensing SNR in both LoS and NLoS dominant areas, while ensuring the required SINR for the communication user. This highlights the potential of leveraging digital twins to assist ISAC systems.

Figures

Figures reproduced from arXiv: 2412.07180 by the authors.

**Figure 3.** This figure presents the beam patterns of the sensing and the [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

**Figure 4.** This figure presents the CDF of the sensing SNR with the minimum communication SINR of 10 dB. The proposed digital twin-aided approach achieves [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

12 extracted references · 9 canonical work pages

[1]

Integrated sensing and communications: Toward dual-functional wireless networks for 6G and beyond,

F. Liu, Y . Cui, C. Masouros, J. Xu, T. X. Han, Y . C. Eldar, and S. Buzzi, “Integrated sensing and communications: Toward dual-functional wireless networks for 6G and beyond,” IEEE J. on Sel. Areas in Commun. , vol. 40, no. 6, pp. 1728–1767, 2022

work page 2022
[2]

Integrated sensing and communication for 6G: Ten key machine learning roles,

U. Demirhan and A. Alkhateeb, “Integrated sensing and communication for 6G: Ten key machine learning roles,” IEEE Commun. Mag. , 2023

work page 2023
[3]

Joint transmit beamforming for multiuser MIMO communications and MIMO radar,

X. Liu, T. Huang, N. Shlezinger, Y . Liu, J. Zhou, and Y . C. Eldar, “Joint transmit beamforming for multiuser MIMO communications and MIMO radar,” IEEE Trans. on Signal Process , vol. 68, pp. 3929–3944, 2020

work page 2020
[4]

Multibeam opti- mization for joint communication and radio sensing using analog antenna arrays,

Y . Luo, J. A. Zhang, X. Huang, W. Ni, and J. Pan, “Multibeam opti- mization for joint communication and radio sensing using analog antenna arrays,” IEEE Trans. on V eh. Technol., vol. 69, no. 10, pp. 11 000–11 013, 2020

work page 2020
[5]

Joint transmit and receive beamforming design for integrated sensing and communication,

N. Zhao, Y . Wang, Z. Zhang, Q. Chang, and Y . Shen, “Joint transmit and receive beamforming design for integrated sensing and communication,” IEEE Commun. Letters , vol. 26, no. 3, pp. 662–666, 2022

work page 2022
[6]

Joint MIMO communications and sensing with hybrid beamforming architecture and OFDM waveform optimization,

S. D. Liyanaarachchi, T. Riihonen, C. B. Barneto, and M. Valkama, “Joint MIMO communications and sensing with hybrid beamforming architecture and OFDM waveform optimization,” IEEE Trans. on Wireless Commun., 2023

work page 2023
[7]

Cell-free joint sensing and communi- cation MIMO: A max-min fair beamforming approach,

U. Demirhan and A. Alkhateeb, “Cell-free joint sensing and communi- cation MIMO: A max-min fair beamforming approach,” in 2023 57th Asilomar Conference on Signals, Systems, and Computers . IEEE, 2023, pp. 381–386

work page 2023
[8]

Optimal transmit beamforming for integrated sensing and communication,

H. Hua, J. Xu, and T. X. Han, “Optimal transmit beamforming for integrated sensing and communication,” IEEE Trans. on V eh. Technol. , 2023

work page 2023

Show all 12 references

[9]

Real-time digital twins: Vision and research directions for 6G and beyond,

A. Alkhateeb, S. Jiang, and G. Charan, “Real-time digital twins: Vision and research directions for 6G and beyond,” IEEE Commun. Mag. , 2023

2023
[10]

Learnable wireless digital twins: Reconstructing electromagnetic field with neural representations,

S. Jiang, Q. Qu, X. Pan, A. Agrawal, R. Newcombe, and A. Alkhateeb, “Learnable wireless digital twins: Reconstructing electromagnetic field with neural representations,” arXiv preprint arXiv:2409.02564 , 2024

2024 arXiv
[11]

Large wireless model (LWM): A Foundation model for wireless channels,

S. Alikhani, G. Charan, and A. Alkhateeb, “Large wireless model (LWM): A Foundation model for wireless channels,” arXiv preprint arXiv:2411.08872, 2024

2024 arXiv
[12]

Semidefinite relaxation of quadratic optimization problems,

Z.-Q. Luo, W.-K. Ma, A. M.-C. So, Y . Ye, and S. Zhang, “Semidefinite relaxation of quadratic optimization problems,” IEEE Signal Process. Mag., vol. 27, no. 3, pp. 20–34, 2010

2010

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

Integrated sensing and communications: Toward dual-functional wireless networks for 6G and beyond,

F. Liu, Y . Cui, C. Masouros, J. Xu, T. X. Han, Y . C. Eldar, and S. Buzzi, “Integrated sensing and communications: Toward dual-functional wireless networks for 6G and beyond,” IEEE J. on Sel. Areas in Commun. , vol. 40, no. 6, pp. 1728–1767, 2022

work page 2022

[2] [2]

Integrated sensing and communication for 6G: Ten key machine learning roles,

U. Demirhan and A. Alkhateeb, “Integrated sensing and communication for 6G: Ten key machine learning roles,” IEEE Commun. Mag. , 2023

work page 2023

[3] [3]

Joint transmit beamforming for multiuser MIMO communications and MIMO radar,

X. Liu, T. Huang, N. Shlezinger, Y . Liu, J. Zhou, and Y . C. Eldar, “Joint transmit beamforming for multiuser MIMO communications and MIMO radar,” IEEE Trans. on Signal Process , vol. 68, pp. 3929–3944, 2020

work page 2020

[4] [4]

Multibeam opti- mization for joint communication and radio sensing using analog antenna arrays,

Y . Luo, J. A. Zhang, X. Huang, W. Ni, and J. Pan, “Multibeam opti- mization for joint communication and radio sensing using analog antenna arrays,” IEEE Trans. on V eh. Technol., vol. 69, no. 10, pp. 11 000–11 013, 2020

work page 2020

[5] [5]

Joint transmit and receive beamforming design for integrated sensing and communication,

N. Zhao, Y . Wang, Z. Zhang, Q. Chang, and Y . Shen, “Joint transmit and receive beamforming design for integrated sensing and communication,” IEEE Commun. Letters , vol. 26, no. 3, pp. 662–666, 2022

work page 2022

[6] [6]

Joint MIMO communications and sensing with hybrid beamforming architecture and OFDM waveform optimization,

S. D. Liyanaarachchi, T. Riihonen, C. B. Barneto, and M. Valkama, “Joint MIMO communications and sensing with hybrid beamforming architecture and OFDM waveform optimization,” IEEE Trans. on Wireless Commun., 2023

work page 2023

[7] [7]

Cell-free joint sensing and communi- cation MIMO: A max-min fair beamforming approach,

U. Demirhan and A. Alkhateeb, “Cell-free joint sensing and communi- cation MIMO: A max-min fair beamforming approach,” in 2023 57th Asilomar Conference on Signals, Systems, and Computers . IEEE, 2023, pp. 381–386

work page 2023

[8] [8]

Optimal transmit beamforming for integrated sensing and communication,

H. Hua, J. Xu, and T. X. Han, “Optimal transmit beamforming for integrated sensing and communication,” IEEE Trans. on V eh. Technol. , 2023

work page 2023