Pinching Antenna-Assisted ISAC with Waveguide Mode Selection

Abstract: Conventional pinching antenna (PA)-assisted integrated sensing and communication (ISAC) architectures typically assume static receiver locations or predetermined receive waveguides, thereby underutilizing the inherent spatial degrees of freedom. This paper proposes a novel mode-selectable PA-assisted ISAC framework to maximize the post-combining sensing signal-to-noise ratio while satisfying multi-user quality-of-service constraints by jointly optimizing the waveguide mode selection, transmit beamforming, and transmit/receive PA positions. To tackle the resulting mixed-integer nonconvex optimization problem, we develop a low-complexity block-coordinate descent algorithm that leverages a penalty-based majorization-minimization method to achieve high-quality suboptimal solutions. Numerical results demonstrate that the proposed design significantly outperforms both traditional PA and fixed-antenna benchmarks by synergistically harnessing spatial adaptability and modal reconfigurability. In particular, the mode-selectable design enables the coordinated optimization of transmit/receive operations and sensing-communication resource allocation, thereby maintaining sensing robustness under stringent communication requirements.

• The paper introduces a mode-selectable pinching-antenna ISAC framework that jointly assigns waveguides to transmission or reception, optimizes beamforming, and places antennas to maximize sensing SNR while meeting communication QoS.
• The proposed penalty-based MM and block-coordinate descent algorithm solves the mixed-integer nonconvex design with polynomial-time iterations, while rank-one beamforming solutions can be recovered without semidefinite-relaxation loss and optimal receive positions have a closed-form projection.
• Simulations with eight waveguides and three users show consistent sensing-SNR gains and stronger robustness to rising rate requirements than fixed mode splits, fixed receive positions, and conventional antenna arrays, although idealized channels and negligible waveguide loss limit practical conclusions.

Overview

This paper addresses a structural limitation in pinching-antenna system (PASS)-assisted integrated sensing and communication (ISAC): prior designs typically fix each waveguide's role (transmit or receive) and fix receive-antenna locations, leaving spatial degrees of freedom (DoF) unexploited. The authors propose a mode-selectable PA-assisted ISAC framework in which each of $N$ dielectric waveguides is dynamically configured as either a transmitting waveguide (TWG) or receiving waveguide (RWG), with at most one transmit PA (TPA) or receive PA (RPA) per waveguide. The design objective is to maximize the post-combining sensing SNR under per-user communication QoS constraints, jointly optimizing binary waveguide-mode selection $\bm{\tau}$, multiuser transmit beamforming, and TPA/RPA positions (2607.15547).

System model and problem formulation

The BS feeds $N$ waveguides of length $L$ at height $d$, serving $K$ single-antenna users while sensing a single stationary point target. A binary vector $\bm{\tau}$ partitions waveguides into TWGs ($\tau_n=1$) and RWGs ($\tau_n=0$), with the feasibility constraint $K \le \mathbf{1}^\top\bm{\tau} \le N-1$. Channels follow spherical-wave near-field models for the free-space segment and phase-shift coefficients within the dielectric waveguide; intra-waveguide power attenuation is neglected to characterize a performance upper bound. Sensing reuses the downlink communication waveform, and the post-combining SNR under normalized maximum-ratio combining is

$\bm{\tau}$0

which couples mode selection (through $\bm{\tau}$1), RPA placement (through $\bm{\tau}$2), beamforming, and TPA placement (through $\bm{\tau}$3). The resulting problem is mixed-integer nonconvex, with constraints on PA positions, total and per-waveguide power budgets, binary modes, and minimum user rates.

Optimization approach

The authors develop a block-coordinate descent (BCD) algorithm alternating between two blocks:

Beamforming and mode selection: The bilinear product $\bm{\tau}$4 is linearized via slack variables using the identity $\bm{\tau}$5, and the binary constraint is handled by a penalty on its D.C.-structured equivalent form. First-order Taylor majorization-minimization (MM) yields convex subproblems after semidefinite relaxation (SDR). Notably, a theorem establishes that a rank-one solution satisfying $\bm{\tau}$6 can always be constructed from the KKT conditions when the problem is feasible—so SDR incurs no loss here.

PA positioning: A lemma shows that, for each RWG, the optimal RPA location is the Euclidean projection of the target's x-coordinate onto the waveguide interval, i.e., $\bm{\tau}$7 for $\bm{\tau}$8, since receive path loss depends only on minimizing $\bm{\tau}$9. This closed-form result substantially reduces the positioning search space. For TPA positions, the remaining nonconvexity—arising from distance-dependent path loss, in-waveguide phases, and rank-one constraints on lifted matrices $N$0, $N$1—is addressed via penalty-based MM: D.C. constraints are convexified by Taylor lower bounds, rank-one conditions are recast as nuclear-norm minus spectral-norm penalties, and periodic phase-coupling terms are upper-bounded by Lipschitz-gradient surrogates with constants $N$2, $N$3. The overall BCD iteration converges to a suboptimal solution with polynomial-time complexity.

Numerical evaluation

Simulations use $N$4 m, $N$5 GHz, $N$6, $N$7, $N$8 W, noise floor $N$9 dBm, and $L$0 bit/s/Hz, averaged over random user/target placements in a $L$1 m² area. Three baselines are compared: fixed random mode split (Bench 1), fixed RPAs at the BS (Bench 2), and a conventional fixed antenna array (Bench 3).

Key findings include:

• Consistent sensing gain: the proposed design achieves the highest post-combining SNR across the entire power range, confirming that joint mode selection and RPA positioning provide effective DoF against two-way path loss.
• Regime-dependent baseline behavior: Bench 1 slightly outperforms Bench 2 at low power, but the ordering reverses at high power—RPA flexibility dominates low-power performance, whereas adaptive mode allocation becomes the bottleneck-relieving factor as power grows.
• QoS robustness: as $L$2 increases, all schemes degrade due to the sensing–communication trade-off, but the proposed design degrades only mildly and remains feasible over the full range, while Bench 1 becomes infeasible under stringent QoS and Bench 3 fails earliest.
• Fixed arrays are inadequate: Bench 3 is generally infeasible even under relaxed QoS, and its upper bound retains a persistent gap relative to PA-assisted schemes.

Limitations and open questions

The paper concedes several modeling assumptions that bound the applicability of its results. Intra-waveguide attenuation is neglected, so reported gains represent an upper bound rather than a hardware-realistic prediction. Perfect channel state information and known target location are assumed, with estimation errors explicitly deferred to future work. Only one PA per RWG is permitted, motivated by avoiding analytically intractable intra-waveguide re-radiation coupling, though multiple TPAs per TWG are claimed to be straightforward extensions. Sensing reuses the downlink waveform; dedicated sensing waveforms, which could expand DoF at the cost of optimization and receiver-processing complexity, remain unexplored. Direct TPA–RPA leakage, target-induced downlink scattering, and user-induced reflections are assumed negligible. Finally, the framework targets a single stationary point target under narrowband LoS propagation, leaving multi-target and non-stationary scenarios open.

Conclusion

The paper introduces waveguide-mode selection as a design dimension in PA-assisted ISAC and demonstrates, via a penalty-based MM/BCD solver with provable rank-one recovery and polynomial-time convergence, that jointly adapting transmission/reception roles alongside PA placement yields higher sensing SNR and markedly better robustness to tightening communication QoS than fixed-mode, fixed-position, or conventional fixed-array architectures. The central empirical claim—that adaptive waveguide utilization is the key mechanism balancing sensing robustness against communication performance—is supported by consistent simulation gains, subject to the idealized channel and hardware assumptions noted above.