arXiv:2609.38500
Q-SPARSE: Quantum Subspace Projection for Near-Field Angle-Range Spectrum Estimation
Mostafizur Rahaman Laskar, Sajad Daei, Gabor Fodor, Mikael Skoglund
Abstract
Near-field wavefront curvature enables joint angle-range localization, but exploiting it entails repeated covariance-eigenspace tests over a two-dimensional spherical manifold. We propose a quantum subspace-projection algorithm (Q-SPARSE) for super-resolution estimation of parameters in the near field, employing spectral signature similar to MUSIC. Q-SPARSE prepares each spherical steering hypothesis as a quantum state, applies quantum phase estimation (QPE) to covariance-generated evolution, and converts the resolved spectral mass into a localization spectrum. The proposed method does not need the full classical preparation of the eigenstates, which is a major bottleneck of the QPE algorithm, the main subroutine in quantum variants of MUSIC algorithms. Under an ideal signal-noise partition, its score equals the classical signal-subspace overlap exactly and thus preserves the maximizer of normalized near-field MUSIC. For a finite q number of qubit phase registers, we derive the score from the complete QPE kernel, and propose a nearest-bin selection scheme with soft-thresholding based on Marchenko-Pastur-calibrated logistic weighting decision criteria. Numerical results on IBM's QISKIT software platform are demonstrated for near-field array processing.
AI Summary
Q-SPARSE addresses joint angle–range estimation for near-field arrays, where MUSIC searches a two-dimensional spherical steering manifold after covariance eigenspace decomposition. It prepares each candidate steering vector as a quantum state, applies quantum phase estimation (QPE) to $U_R=\exp(i2\pi\widehat R/\alpha)$, and weights the measured phase bins to estimate signal-subspace membership without explicitly preparing covariance eigenvectors. Under an ideal signal–noise partition, the score is exactly $\langle a|P_S|a\rangle$, so its maximizers coincide with normalized near-field MUSIC. For finite phase-register width $q$, the paper derives the score using the full QPE leakage kernel. A logistic spectral weight uses the Marchenko–Pastur noise edge $\lambda_+=\sigma_n^2(1+\sqrt{N/L})^2$. With a flag measurement, $M$ shots give score variance at most $1/(4M)$. Conditional on efficient steering-state preparation and covariance block encoding, the stated search cost is $\widetilde O(G\epsilon^{-3}\operatorname{poly}(\log N))$ for $G$ hypotheses; generic dense synthesis instead costs $O(GN^2\epsilon^{-2}\log(1/\epsilon))$. Validation uses 2,000-scene numerical sweeps with $q=10$ and $M=10^4$, plus separate small-array Qiskit-Aer circuit simulations. Comparisons examine localization errors, detection, array size, and source loading; dense validation circuits reach roughly $10^4$ depth and $2.5\times10^3$ CX gates. The paper states that these bounds do not establish unconditional quantum speedup. Hard partitioning requires resolution below half the signal–noise eigengap; soft weighting requires resolution relative to its transition width. Colored noise needs additional treatment. Ideal performance cannot exceed matched MUSIC statistically, and any differences from logistic weighting also admit a classical matrix-function implementation.
Error found: The abstract advertises nearest-bin selection, but the algorithm specifies logistic weighting across all measured phase bins and gives no separate nearest-bin rule. The supplied numerical figures contain captions without plots or data, preventing verification of the reported confidence intervals and performance comparisons. The introduction describes validation on a real system, but the reported quantum-circuit validation uses QISKIT-Aer simulation rather than quantum hardware.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- Part of a journal submission and submitted to ICASSP 2027. Content may be updated post-review. Subject to journal policies regarding preprint hosting