Robust Beamforming Design for Integrated Sensing and Communications with Mutual Coupling Effect

arXiv:2608.15556 · eess.SP · Submitted 2026-08-16 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Electrical Engineering. The day in electrical engineering: signals, systems and the hardware that carries them.

Rosa: I'm Rosa.

Dev: And I'm Dev, here with Taro.

Rosa: We're looking at the paper: "Robust Beamforming Design for Integrated Sensing and Communications with Mutual Coupling Effect".

Dev: Here's the gist of the paper. Integrated sensing and communications (ISAC) is a key technology for next-generation wireless networks, enabling communication and radar sensing over shared spectral and hardware resources.

Rosa: In practical multi-user multiple-input multiple-output (MU-MIMO) ISAC transmitters, however, mutual coupling (MC) between antenna elements distorts the array steering vector and each communication user (CU) channel, so that the sensing beampattern deviates from the desired one and the communication link to each user degrades.

Dev: To address this limitation, we propose a robust MC-compensated beamforming design that guarantees both the sensing and communication performance of MU-MIMO ISAC transmitters against the residual MC error.

Rosa: Problem Context Conventional ISAC beamforming relies on an ideal steering vector assuming no electromagnetic (EM) interaction between elements.

Paper summary: Rosa: So we’re diving into this paper now. The central idea of this work is tackling a real problem in integrated sensing and communications where you have to balance communication performance with radar sensing performance on the same hardware.

Dev: Specifically, when you use multi-user multiple-input multiple-output, or MU-MIMO, transmitters for ISAC—which is using shared resources for both—the mutual coupling between antenna elements distorts the system. This distortion messes up the desired sensing beampattern and degrades the communication link to each user.

Rosa: The thesis of this paper is to propose a robust MC-compensated beamforming design that specifically guarantees both the sensing and communication performance of these MU-MIMO ISAC transmitters against those residual mutual coupling errors.

Dev: They introduce a key idea: they place a residual error on the mutual coupling matrix of their physics model, which causes uncertainties in both the sensing steering vector and the communication channel.

Rosa: This means instead of just assuming an ideal array response, they center their uncertainty sets around this physics-based response, so each error bound only covers that residual modeling uncertainty through a single parameter.

Dev: The goal then becomes minimizing the worst-case sensing beampattern matching mean-squared error while simultaneously guaranteeing the worst-case signal to interference plus noise ratio for every single communication user.

Rosa: They achieve this by casting both of those difficult worst-case requirements into one common quadratic form and turning them into a convex semidefinite program, or SDP. That’s how they formulate the problem, P3.

Dev: The result is a design that is robust against the residual MC error in both radar sensing and communication performance under these specific constraints.

Rosa: And this matters because it means for real-world ISAC transmitters, you get a design that performs well even when the physical mutual coupling isn't perfectly captured by the model they used.

Dev: It’s about building systems that stay reliable when dealing with imperfect physical modeling of interference between antenna elements.

Taro: From an autonomy side, I’m thinking about how this uncertainty propagates into actual target detection or parameter estimation when you have these uncertainties in both domains simultaneously. If the residual error is significant, does it mean we need a different way to handle that in real-time?

Rosa: That's a good point. Taro, the paper doesn't explicitly detail those propagation scenarios; it focuses on establishing the worst-case performance bounds under these uncertainties as their main contribution.

Dev: I think the focus is really on showing that they can find a solution that meets both SINR and matching MSE constraints robustly against this specific type of error.

Conclusion: Rosa: So, looking at the final results of this paper by Maeng and Han, the title is Robust Beamforming Design for Integrated Sensing and Communications with Mutual Coupling Effect. It’s a pretty direct way to describe exactly what they did.

Dev: It boils down to this: they developed a method where you model the residual uncertainty in mutual coupling as a norm-bounded error on that matrix, which then drives both sensing and communication uncertainties.

Rosa: This approach means their design is centered on the physics-based response rather than an ideal one, making the uncertainty bounds much more focused.

Dev: So what this means for us in the real world is that we can expect better worst-case sensing pattern matching mean-squared error and higher communication signal-to-interference plus noise ratio than previous designs when mutual coupling is present.

Rosa: The authors are proving that you can get better worst-case performance guarantees by being smart about where you place your uncertainty in the design process.

Dev: Ultimately, they show that this is a way to achieve better performance by centering the uncertainty at the physical model rather than trying to center it around a perfect mathematical ideal.

Taro: For me, it’s about acknowledging that real hardware isn't perfect and designing systems that gracefully handle those inevitable physical distortions.

Rosa: Right, so this paper gives us a framework for building more resilient integrated systems by explicitly modeling the residual error from mutual coupling effects and using it to tighten our performance guarantees.

Jieon Maeng, Kawon Han

Ulsan National Institute of Science and Technology

eess.SP

Submitted: 2026-08-16

Updated: 2026-11-04

Comments: 5 pages, 5 figures, submitted to IEEE Wireless Communication Letters, 2026

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 83/100

The gist: Integrated sensing and communications (ISAC) is a key technology for next-generation wireless networks, enabling communication and radar sensing over shared spectral and hardware resources.

Key concepts

Mutual Coupling (MC)
MC occurs when closely spaced antenna elements electromagnetically interact with each other. This interaction distorts the intended signal patterns because the field radiated by one element affects its neighbors, causing the actual array response to deviate from the ideal theoretical response.
Robust MC-Compensated Design
This design improves upon standard MC compensation by modeling the uncertainty as a norm-bounded error on a physics-based model. Instead of assuming perfect knowledge of the physical model, it accounts for small, realistic errors in that model, ensuring performance guarantees even when the actual coupling differs slightly from the predicted one.
Worst-Case Guarantee
The design aims to ensure performance metrics—like sensing beampattern matching MSE and communication SINR—are met under the absolute worst possible realization of the residual MC error. This provides a strong safety margin for both radar sensing and data communication simultaneously.
Semidefinite Program (SDP)
The complex optimization problem, which involves minimizing performance while satisfying multiple worst-case constraints, is converted into a convex SDP. This mathematical framework allows the problem to be solved efficiently by replacing difficult non-convex constraints with linear matrix inequalities (LMIs).

Terminology

Summary

Integrated sensing and communications (ISAC) is a key technology for next-generation wireless networks, enabling communication and radar sensing over shared spectral and hardware resources. p. 0 In practical multi-user multiple-input multiple-output (MU-MIMO) ISAC transmitters, however, mutual coupling (MC) between antenna elements distorts the array steering vector and each communication user (CU) channel, so that the sensing beampattern deviates from the desired one and the communication link to each user degrades. p. 0 To address this limitation, we propose a robust MC-compensated beamforming design that guarantees both the sensing and communication performance of MU-MIMO ISAC transmitters against the residual MC error. p. 0

Problem Context

Conventional ISAC beamforming relies on an ideal steering vector assuming no electromagnetic (EM) interaction between elements. In practical ISAC transmitters, the mutual coupling (MC) effect between closely spaced elements becomes non-negligible. p. 0 This MC-induced distortion can be mitigated by MC-compensated beamforming which replaces the ideal array response with the one predicted by a physics-based MC model. p. 0 However, this model leaves a residual MC uncertainty. p. 0 Existing robust ISAC designs center their uncertainty sets at the ideal response, requiring error bounds to absorb the full MC-induced deviation in addition to the residual error. p. 0

Proposed Robust Design

The proposed robust design places a common norm-bounded residual error on the MC matrix of a physics-based model. This residual error induces both the sensing steering vector and communication channel uncertainties. p. 0 The design minimizes the worst-case beampattern matching mean-squared error (MSE) while guaranteeing the worst-case SINR of every CU. p. 0 This is achieved by casting both worst-case constraints into one common quadratic form and converting them into a convex semidefinite program (SDP). p. 1

Mathematical Formulation

The robust design (P3) minimizes the worst-case sensing beampattern matching MSE subject to the SINR constraint of every communication user and any residual within the error bound. To handle the min–max objective, an auxiliary variable Sm is introduced to upper-bound the worst-case matching error. p. 2 The problem is formulated as a convex SDP (P′3) by replacing semi-infinite constraints with linear matrix inequalities (LMIs) using Lemma 1. p. 1

Numerical Validation

Numerical results show that the proposed robust design attains both a lower sensing beampattern matching MSE and a higher communication SINR than those of the conventional designs. The margin widens with the mismatch, as shown in Fig. p. 4 4 where the robust design outperforms the MC-compensated one as the mismatch grows. p. 4 The complexity remains polynomial of the same order in Nt, higher only by a factor pM/(K+1) due to enforcing worst-case bounds at every angle and user. p. 3

Conclusion

In this paper, we have proposed robust MC-compensated beamforming for MU-MIMO ISAC which models the residual uncertainty of the MC model as a norm-bounded error on the MC matrix. The worst-case guarantees of the proposed design yield a lower matching MSE and a higher achieved SINR than the conventional designs in the presence of the residual MC error. p. 4

REFERENCES

arXiv:2608.15556v1 1,Abstract

--- Page 2 ---

I NTEGRATED sensing and communications (ISAC) is a key technology for next-generation wireless systems, allowing a single transceiver to simultaneously perform communication and radar sensing over shared spectral and hardware resources [1], [2]. In multi-user multiple-input multiple-output (MU-MIMO) ISAC systems, transmit beamforming shapes a desired sensing beampattern while guaranteeing the signalto-interference-plus-noise ratio (SINR) requirement of each communication user (CU) [3]. p. 0 Conventional ISAC beamforming is built on the ideal steering vector of the antenna array, which assumes no electromagnetic (EM) interaction between elements. p. 0 In the antenna arrays of practical ISAC transmitters, however, this assumption breaks down. p. 0 The mutual coupling (MC) effect between closely spaced elements becomes non-negligible, as the fields radiated by each element couple into the adjacent elements [4]. p. 0 This MC-induced distortion can be mitigated by MCcompensated beamforming, which replaces the ideal array response with the one predicted by a physics-based MC model [5], [6]. p. 0 Such a model, however, relies on idealized physical conditions that hold only approximately in practice, leaving a residual MC uncertainty. p. 0 Under this uncertainty, in MIMO radar sensing, MC distorts the transmit beampattern and degrades target detection and parameter estimation, while in MU-MIMO communication it reduces the beamforming gain at each user equipment (UE) and corrupts the channel state information (CSI), as illustrated in Fig. p. 0 1. Although MC-compensated designs replace the ideal array response with the one predicted by a physics-based MC model [5], [6], they evaluate the sensing beampattern and the channel at a fixed response, offering no robustness against the residual MC error. p. 0 Existing robust ISAC designs ensure the worst-case performance against CSI errors [7], [8] and target location uncertainty [9], [10]. p. 0 However, their uncertainty sets are centered at the ideal response, so the error bounds must absorb the full MC-induced deviation in addition to the residual error, making the design unnecessarily conservative. p. 0 Since the residual error lies on the MC matrix and distorts steering vectors and user channels, the uncertainty set should be centered at the MC model. p. 0 Such ISAC beamforming, robust to the residual MC error in both radar sensing and communication, has not yet been investigated. p. 0 To address this limitation, we propose a robust MCcompensated ISAC beamforming design that provides a worstcase performance guarantee for both radar sensing and MUMIMO communication under residual MC errors. p. 0 The main contributions of this paper are summarized as follows. p. 0 First, unlike the ideal-centered robust designs of [7]–[10], we place a common norm-bounded residual error on the MC matrix of a physics-based model [11], which induces the sensing steering vector and communication channel uncertainties. p. 0 Every uncertainty set is thus centered at the physics-based response, and each error bound covers only the residual MC error via an interpretable parameter. p. 0 Second, we minimize the worstcase beampattern matching mean-squared error (MSE) while guaranteeing the worst-case SINR of every CU. p. 0 We cast both worst-case constraints into one common quadratic form, convert them into linear matrix inequalities (LMIs), and develop a convex semidefinite program (SDP) that enforces both worstcase guarantees, at a complexity of the same polynomial order as the non-robust designs. p. 1

--- Page 3 ---

A. Communication Model Consider the MU-MIMO ISAC scenario in Fig. p. 1 p. 1 1, where a transmitter with an Nt-element patch antenna array serves K single-antenna CUs while sensing a radar target within the region-of-interest (ROI). p. 1 p. 1 Following the standard ISAC signal model [12], the transmitted signal is x = Wc c + Ws s, (1) where Wc = [wc,1,. p. 1 p. 1.., wc,K] ∈ C Nt×K collects the communication beamformers, c ∈ C K is the data-symbol vector of the K CUs, and Ws ∈ C Nt×Nt and s ∈ C Nt are the beamforming matrix and waveform of a dedicated sensing signal. p. 1 p. 1 The symbols and waveforms are zero-mean, unitpower, and mutually uncorrelated, E ccH = IK, E ssH = INt, E scH = 0. p. 1 p. 1 The channel of user p k follows the Rician model hk = κ/(κ + 1) hLoS,k + p 1/(κ + 1) hNLoS,k, where hLoS,k is the line-of-sight (LoS) component along the steering vector toward the user angle θk, hNLoS,k ∼ CN (0, INt) is the Rayleigh-scattered non-line-of-sight (NLoS) component, and κ is the Rician factor. p. 1 p. 1 The received signal of user k is yk = h H k x + nk with noise nk ∼ CN (0, σ2). p. 1 p. 1 Since the multiuser and sensing components act as interference, the SINR γk of user k is given by [12]. p. 1 p. 1

Sensing Model

For sensing, the transmit array is operated as a MIMO radar system. For a uniform linear array (ULA) with element spacing d, the ideal steering vector is a(θ) = h 1, e j 2πd λ0 sin θ,. p. 1.., e j 2πd λ0 (Nt−1) sin θ iT, where λ0 is the carrier wavelength and d = λ0/2 is assumed throughout this paper, with∥a(θ)∥2 = √Nt. p. 1 With the peruser covariance Rc,k = wc,kwH c k and the sensing covariance Rs = WsWH s, the total transmit covariance under (2) is R = E xxH = WcWH c + WsWH s = X K k=1 Rc,k + Rs, (5). p. 1 and the transmit beampattern is P(θ) = a H(θ) R a(θ). p. 0 We minimize the beampattern matching MSE between P(θ) and a desired sensing beampattern Pd(θ) over a grid of M angles M m=1, and utilize this MSE to formulate the objectives of all beamforming designs in the following sections. p. 1

Antenna Mutual Coupling Model

We model the MC of a patch-element ULA by the physics-based MC model of [11]. The MC is set by two admittances: the selfadmittance Yself ∈ C of an isolated patch, and the mutual admittance Ymut(p−q d) between two patches whose centers are separated by p − q d, where p, q ∈ 1. p. 1.., Nt index the array elements. p. 1 These assemble into the admittance matrix Y of the Nt-port array, relating the port currents to the port voltages, with [Y]pq = (Yself, p = q, Ymut(p − q d), p ≠ q). p. 1 Inverting it yields the open-circuit impedance matrix Zoc = Y−1, where the inversion automatically accounts for all multiple-scattering paths between elements. p. 1 Each port is then terminated with the load impedance ZL, giving the MC matrix of the load-terminated array as [4]. p. 1 Finally, we normalize Z by its average element gain g =∥Z∥F / √Nt as Z¯ = Z/g, so that the coupling alters only the pattern shape, not the average gain. p. 1 Replacing the ideal steering vector (4) by its coupled counterpart gives aMC(θ) = Z a ¯ (θ), on which MC-compensated beamforming evaluates the beampattern (6). p. 1 The same matrix acts on the channel [13], giving hMC,k = Z h¯ k. p. 1 Both multiplications are linear because the coupling acts on the element currents, which set the field radiated toward every angle and along every path alike. p. 1

Problem Formulation and Solution

The problem (P0) minimizes the sensing beampattern matching MSE subject to the SINR constraint of every CU. This problem is non-convex due to the quadratic SINR constraint (11b) and is relaxed to a convex problem by semidefinite relaxation (SDR) over the covariances of (5). p. 2 The MC-compensated design (P2) has the same form as (P1), but it evaluates the beampattern and the channel at a fixed array response and offers no guarantee against residual uncertainty. p. 2 We propose a robust MC-compensated design that places the residual error on the MC model via the normbounded error model of [15]. p. 2 This leads to the robust design (P′3) which minimizes worst-case performance subject to constraints derived from Lemma 1. p. 0

Computational Complexity

The complexity of the robust design (P′3) remains polynomial of the same order in Nt, higher only by a factor pM/(K+1) due to enforcing worst-case bounds at every angle and user. This is because each iteration costs the same order as non-robust designs, but the worst-case bounds impose one pair Φm, Ψm per angle and one Ξk per user, adding 2M+K blocks of size Nt+1 that raise the iteration count to O√MNt. p. 3

REFERENCES

arXiv:2608.15556v1 4,Problem Formulation and Solution

--- Page 4 ---

II. SYSTEM MODEL A. p. 1 Communication Model Consider the MU-MIMO ISAC scenario in Fig. 1, where a transmitter with an Nt-element patch antenna array serves K single-antenna CUs while sensing a radar target within the region-of-interest (ROI). Following the standard ISAC signal model [12], the transmitted signal is x = Wc c + Ws s, (1) where Wc = [wc,1,..., wc,K] ∈ C Nt×K collects the communication beamformers, c ∈ C K is the data-symbol vector of the K CUs, and Ws ∈ C Nt×Nt and s ∈ C Nt are the beamforming matrix and waveform of a dedicated sensing signal. The symbols and waveforms are zero-mean, unitpower, and mutually uncorrelated, E ccH = IK, E ssH = INt, E scH = 0. The channel of user p k follows the Rician model hk = κ/(κ + 1) hLoS,k + p 1/(κ + 1) hNLoS,k, where hLoS,k is the line-of-sight (LoS) component along the steering vector toward the user angle θk, hNLoS,k ∼ CN (0, INt) is the Rayleigh-scattered non-line-of-sight (NLoS) component, and κ is the Rician factor. The received signal of user k is yk = h H k x + nk with noise nk ∼ CN (0, σ2). Since the multiuser and sensing components act as interference, the SINR γk of user k is given by [12].

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Robust Beamforming Design for Integrated Sensing and Communications with Mutual Coupling Effect allows AI systems to operate reliably in practical multi-user multiple-input multiple-output (MU-MIMO) ISAC transmitters by guaranteeing both sensing and communication performance against residual mutual coupling errors.

  2. The proposed design minimizes the worstcase beampattern matching mean-squared error (MSE) for sensing while simultaneously guaranteeing the signal-to-interference-plus-noise ratio (SINR) for each communication user, leading to better radar target detection and more reliable data links.

  3. The system can handle residual mutual coupling by placing a common norm bounded residual error on the MC matrix of a physics based model, ensuring that uncertainty sets are centered at the physics based response rather than the ideal response, making it less conservative.

  4. AI systems equipped with this design can achieve lower matching MSE and higher SINR compared to conventional designs, with an advantage that widens as the residual error grows beyond the design bound eta.

Related papers