Lightweight Single-Antenna Direction-of-Arrival Estimation for Curvilinear Trajectories in Mobile Embedded Systems

J\'er\'emie Voix; Lokman Sboui; Nour Zidi; Pascal Giard

REVIEW 2 major objections 4 minor 30 references

Lightweight Single-Antenna Direction-of-Arrival Estimation for Curvilinear Trajectories in Mobile Embedded Systems

T0 review · 2 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A single antenna moving along a curved path can estimate the direction to a known fixed beacon using only onboard radio ranging and compass headings, without GPS or optical tracking.

desk verdict The hardware range-domain bearing result is real and well-measured, but the curvilinear MUSIC simulation is built on a dimensionally inconsistent reconstruction objective (Eq. 8) that invalidates those simulation claims as they stand. read the letter →

arxiv 2608.12029 v1 pith:LBSF6F3J submitted 2026-08-12 eess.SP

Nour Zidi , Jérémie Voix , Pascal Giard , Lokman Sboui This is my paper · ORCID

classification eess.SP

keywords Direction-of-ArrivalvirtualantennaarrayMUSICtwo-wayranginginertialmeasurementunitUWBembeddedsystemscurvilineartrajectory

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

This paper claims that a single-antenna receiver moving along a curved path can estimate the direction of arrival of a known fixed beacon using only onboard two-way-ranging (TWR) distance measurements and inertial heading data, with no GPS, optical tracking, or motion capture. The receiver's motion synthesizes a virtual antenna array, and the paper shows how to reconstruct the array geometry from range circles and IMU displacements. In simulation, the curvilinear virtual-array multiple-signal-classification (MUSIC) estimator resolves multiple sources and stays accurate across the SNR range tested; in hardware, 100 consecutive rotating sweeps recover a 123 mm range modulation (against a measured 120 mm lever arm) and yield a single-sweep bearing precision of 8.2 degrees. The bearing computation consumes 144.9 mJ, about 3% of the cycle energy, which matters for battery-powered wearables that use direction to route audio in GPS-denied industrial environments.

What carries the argument

The load-bearing object is the virtual array: $M$ positions along the receiver's trajectory, treated as elements of an antenna array. Because no external tracker supplies those positions, the paper reconstructs them from TWR ranges and IMU headings by minimizing a combined cost of range-circle residuals and IMU displacement consistency (Eq. 8). For the range-domain hardware estimator, the key identity is the sinusoidal range modulation $d(\varphi_m)\approx d_0 - r\cos(\varphi_m - \theta)$, which maps the bearing $\theta$ to the phase of the range ripple; for the coherent estimator, the steering vector $\mathbf{a}(\alpha,\theta)=\exp\big(j k_w \mathbf{R}_M^{\top}\mathbf{u}(\theta)\big)$ maps candidate directions onto the reconstructed geometry. The first-order error model $\sigma_{\hat{\theta}}\approx \sigma_{\mathrm{eff}}/(\sqrt{M}\,W(\theta))$ then ties bearing uncertainty to the effective cross-range aperture, showing why motion direction relative to the source matters.

What would settle it

Repeat the circular experiment with the anchor placed at a known but deliberately offset surveyed position; a correct reconstruction should shift the recovered bearing by exactly the geometric angle of that offset, while a biased reconstruction would produce a different shift. A stronger check is free-form motion with a motion-capture system as ground truth, comparing reconstructed virtual-element positions to captured positions: an unbiased estimator should show zero-mean position errors following the $\sigma_{\mathrm{read}}/\sqrt{N}$ averaging law.

Watch

Extended reading notes

Core claim

The central discovery is that direction-of-arrival estimation can be moved from a physical antenna array to the motion of a single antenna, provided the motion is measured well enough to reconstruct the synthetic aperture. The paper formulates a multiple-signal-classification (MUSIC) estimator for arbitrary curvilinear trajectories whose virtual-element coordinates are recovered by jointly enforcing TWR range circles centered on the known beacon and IMU displacement increments. It then validates the range-domain measurement chain experimentally: rotating the tag at a 120 mm lever arm produces a sinusoidal range modulation whose phase is the beacon bearing, and phase-aligned averaging across 100 sweeps recovers a 123 mm amplitude with a single-sweep bearing precision of 8.2 degrees. Absolute world-frame accuracy is limited by a systematic magnetometer drift of about 2.0 degrees per sweep, so the hardware results characterize precision rather than absolute accuracy; the authors are explicit that phase-coherent MUSIC remains a simulation-only result.

Load-bearing premise

The method assumes the fixed beacon's position is known accurately in advance; if that surveyed position is wrong, every reconstructed virtual-array coordinate and every bearing estimate inherits a bias that onboard IMU data cannot detect or correct.

Editorial extensions

If this is right

A wearable node with one UWB radio and an IMU can compute bearing to a pre-surveyed beacon entirely onboard, eliminating the need for GPS or external tracking hardware in indoor industrial settings.
Range-domain bearing remains viable even when the single-sweep modulation amplitude is comparable to ranging noise, because phase-aligned accumulation across sweeps recovers the modulation.
In simulation, the reconstructed curvilinear array supports MUSIC multi-source resolution for separations around 8 degrees with six or more virtual elements, and outlier gating roughly halves strong-NLoS RMSE.
The measured computational cost of 144.9 mJ per bearing estimate, about 3% of cycle energy, would allow frequent directional updates without dominating a wearable's battery budget.

Reading between the lines

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

The 8.2-degree precision and the near-unity per-bin amplitude SNR suggest the range-domain method is suited to sector-level spatial audio rather than fine angular resolution; reaching sub-degree accuracy would likely require phase-coherent processing, which the paper leaves for future work.
If the constant-rotation-rate heading correction (Eq. 26) is replaced by a general motion model, the same range-domain mechanism should extend to free-form trajectories, but this extension is not demonstrated here.
The known-beacon assumption could in principle be relaxed: with two or more anchors, the TWR circles would bootstrap receiver and beacon positions simultaneously, removing the pre-surveyed infrastructure requirement.
The measured noise floor (about 151 mm after 128 exchanges) implies a design rule for choosing motion patterns: the cross-range aperture $W(\theta)$ must be large enough that $\sigma_{\mathrm{eff}}/(\sqrt{M}\,W(\theta))$ meets the target angular accuracy.

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

The stress-test note is correct, and this is the first thing you need to know. Eq. (8) minimizes (||r_m - r_tx||^2 - dhat_m)^2, which mixes squared distance with one-way range. Even with perfect ranging, the true positions don't minimize the objective (the residual is (d^2 - d)^2, not zero). So the TWR+IMU reconstruction used for the curvilinear virtual array — Algorithm 1, Figure 4, Figure 9, Table II — is not a valid least-squares fit to the measurement model. The MUSIC simulation results that depend on those reconstructed coordinates are unreliable. The paper is honest that MUSIC is simulation-only, but the simulation itself rests on this flawed step.

What survives is the hardware validation. The circular experiment bypasses Eq. (8) and fits Eq. (27) directly to the range sequence, so the 123 mm recovered modulation vs. the 120 mm lever arm is a genuine cross-validation. The measurement discipline is good: static sigma_read of 151 mm, the 1/sqrt(N) averaging law confirmed by the N=32 vs N=128 factor of 2.01, 100 consecutive sweeps retained, and a clean board-level energy measurement (144.9 mJ, 3% of cycle). The single-antenna TWR+IMU bearing estimator is a legitimate system-level step beyond the cited linear-motion and dual-antenna works, and the authors are appropriately careful about the magnetometer drift limiting absolute accuracy.

Soft spots beyond Eq. (8): the 8.2-degree figure is a self-consistency precision metric, not an absolute error bound — the paper says so, but it's easy to over-read. The linearization bound Eq. (15) is asserted without proof; minor. No code or data are provided, which is a real omission for a paper whose simulations are now suspect. The citation pattern is fair and relevant.

Bottom line: the hardware contribution deserves serious consideration, but the manuscript cannot be accepted with the current Eq. (8). The simulation must be redone with a dimensionally consistent constraint (either (||r_m - r_tx|| - dhat_m)^2 or (||r_m - r_tx||^2 - dhat_m^2)^2), and the authors should deposit data/code to make the corrected reconstructions verifiable. I would send this to peer review, but with the reconstruction flaw flagged as a major revision requirement. I wouldn't cite it in its current form.

Referee Report

2 major / 4 minor

Summary. The paper proposes a single-antenna direction-of-arrival framework in which receiver motion creates a virtual aperture. Onboard two-way-ranging (TWR) to a known fixed beacon and IMU headings are used to reconstruct the trajectory, and a phase-coherent curvilinear virtual-array MUSIC formulation is evaluated numerically. Hardware experiments validate a range-domain TWR-IMU bearing estimator on a motorized circular platform over 100 sweeps, reporting a 123 mm recovered range-modulation amplitude against a 120 mm physical lever arm, a single-sweep bearing precision of 8.2 degrees, and 144.9 mJ per embedded bearing-estimation cycle. The paper explicitly limits hardware validation to range-domain estimation and states that phase-coherent MUSIC remains simulation-only.

Significance. The hardware measurement chain is a genuine system-level contribution with credible external validation: the static sigma_read around 151 mm, the 1/sqrt(N) averaging law confirmed by the N=32 versus N=128 factor of 2.01, the recovered 123 mm modulation amplitude against the independently measured 120 mm lever arm, and the careful per-phase energy breakdown all support the range-domain estimator. However, the general curvilinear MUSIC claim rests on trajectory reconstruction via Eq. (8), which is dimensionally inconsistent as written. Until that reconstruction is corrected and the simulation chain is re-run with realistic reconstruction errors, the broader feasibility claim for arbitrary curvilinear trajectories is not established.

major comments (2)

[Section III-A, Eqs. (8)-(9), Algorithm 1] The reconstruction objective is dimensionally inconsistent. The first term of Eq. (8) is (||r_m - r_tx||^2 - dhat_m)^2, which compares a squared distance (units m^2) with a one-way range (units m). If the range observation were perfect, the true element position would satisfy ||r_m - r_tx|| = dhat_m, leaving a residual (dhat_m^2 - dhat_m)^2 rather than zero, so the true trajectory is not the minimizer of Eq. (8). The correct circle constraint is (||r_m - r_tx|| - dhat_m)^2 or (||r_m - r_tx||^2 - dhat_m^2)^2. Consequently, the statistical weighting in Eq. (9), mu = sigma_read^2 / sigma_dr^2, is only consistent with a linear range residual; for a squared-distance residual, the variance depends on approximately 2 d sigma_read and the weight must be re-derived. Since Algorithm 1 and Figures 4, 9, and Table II use coordinates reconstructed from Eq. (8), this error invalidates the reported curvilinear-reconstruction and MUSIC simulation results. The hardware experiment fits Eq. (27) directly and does not exercise Eq. (8), so it cannot validate the general curvilinear reconstruction.
[Section IV, Figures 4 and 9] The link between realistic TWR-IMU reconstruction error and MUSIC performance is not established. Figure 4 reports a median element-position error of about 25 cm for an approximately 8 m trajectory, whereas the MUSIC study uses a compact 3.5-lambda (about 17.5 cm) aperture and Figure 9 perturbs virtual-element positions only up to roughly lambda/4 = 1.25 cm. No figure reports the distribution of the reconstructed-coordinate error for the exact compact array used in Figures 5-8, 10, and 11. Without this, the reader cannot tell whether phase-coherent MUSIC on the reconstructed compact aperture is feasible at the measured TWR noise level (sigma_read about 151 mm) and segment lengths involved. Please report MUSIC DOA RMSE using the actual reconstructed coordinates for the simulation geometry with the stated SR1020/IMU noise, or state explicitly which position-error statistics were assumed.

minor comments (4)

[Section IV, Algorithm 1] The IMU displacement-noise and heading-error statistics used to generate noisy observations in Algorithm 1 are never stated; reporting sigma_dr and the heading-noise model is necessary for reproducibility.
[Section V-D, Conclusion] The 8.2-degree value is computed from the standard deviation of consecutive DOA differences after removing the mean increment, so it is a relative precision metric rather than an absolute accuracy metric. The conclusion says this, but the abstract should explicitly avoid the word 'accuracy' in that sentence.
[Abstract] The sentence 'single-sweep bearing precision is 8.2 degrees absolute world-frame accuracy is limited by systematic BNO055 magnetometer drift' appears to be missing a semicolon or period after '8.2 degrees'.
[Section V-C and Table IV] The 'per-bin amplitude SNR approximately 0.9' is stated without a defining formula; specifying SNR = r / sigma_bin would help the reader connect this value to the 123 mm amplitude and 136 mm per-bin RMSE.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the hardware validation is anchored to an independently measured 120 mm lever arm, and the simulated MUSIC study is a mismatch analysis rather than a self-referential fit. The Eq. (8) unit mismatch is a correctness risk, not a circular reduction.

full rationale

The central hardware feasibility claim is validated against an external physical measurement: the phase-aligned range modulation is fitted at 123 mm and compared with the independently measured 120 mm antenna lever arm (Section V-C), so the recovered amplitude is not an input assumed into the derivation. The 8.2 degree bearing precision is admittedly a self-consistency statistic computed from consecutive sweep differences after removing the mean magnetometer drift, and the paper explicitly states that these experiments characterize estimator precision rather than absolute world-frame accuracy (Section V-D); using an estimator's own outputs to estimate its precision is not circular because no external quantity is being predicted from a fitted version of itself. The MUSIC portion is explicitly scope-limited: Section II-B.6 states that the MUSIC formulation is validated only through numerical simulations using phase-coherent snapshots and reconstructed curvilinear-array geometries, and the simulations inject the true trajectory as ground truth while constructing steering vectors from the reconstructed geometry (Algorithm 1). That is an array-manifold mismatch analysis, not an equation reduced to its own input. The skeptic's Eq. (8) unit mismatch is a real correctness concern: the first term compares squared Euclidean distance (m^2) with a one-way range (m), so a perfect range measurement would not zero the residual at the true position. However, this is a modeling flaw, not circularity: the reconstructed positions are not assumed in the measurement model, and the true DOA is supplied independently in simulation. No self-citation chain is load-bearing: reference [1] provides background for the RAVE application, and the cited ranging application note and MUSIC references are external. On the circularity axis, the paper is self-contained and its headline hardware check is externally anchored.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The central feasibility claim rests primarily on the known-anchor assumption, the far-field plane-wave model, and the heading-alignment correction. The hardware chain is supported by independent measurements such as the static ranging noise and the 120 mm physical lever arm. The main unverified axiom is the linearization error bound, which is stated without proof and may underestimate accumulated error in heading-based trajectory integration.

free parameters (4)

Range modulation amplitude r = 123 mm (compared to 120 mm measured lever arm)
Fitted from the phase-aligned average of 100 sweeps using Eq. (27). Close agreement with the independent physical lever arm supports the model, but the value is fitted, not predicted ab initio.
Averaging count N_avg = 128
Protocol parameter chosen by design. Static measurements confirm the 1/sqrt(N) noise reduction law, so it is a validated design choice rather than an ad hoc fit.
Angular bin population thresholds = 4 observations per bin; 8 of 12 bins per sweep
Retention rules are stated a priori and all 100 sweeps passed, so they do not constitute post hoc selection, but they are hand-chosen thresholds.
Nominal simulation settings M, N_snap, theta0 = M=6, N_snap=16, theta0=30 degrees
Chosen for the compact curvilinear VAA; the paper sweeps over these settings, so they are not hidden fits to the conclusion.

assumptions (6)

domain assumption The fixed beacon position r_tx is known a priori; TWR range circles are centered at this known location.
Invoked in Section II-A and Eq. (8). If the anchor position is inaccurate, the TWR constraint biases all reconstructed virtual-element coordinates and hence the DOA estimate.
domain assumption Far-field narrowband plane-wave propagation: d_{m,k} is approximated by rho_k minus u^T(theta_k) r_m, so the bulk phase drops and the steering vector takes the form of Eq. (12).
Used throughout the MUSIC formulation in Section III-B. A near-field extension is explicitly deferred to future work in Section VI.
standard math The linearization error bound ||r_hat_m minus r_true_m|| <= kappa_max ell_max^2 / 8 (Eq. 15) is valid as stated.
The bound is stated without derivation. For heading-based Euler integration, the error may accumulate over segments, so this axiom carries unverified weight in Table II.
domain assumption First-order position-uncertainty model delta p_m = b + beta J pbar_m + epsilon_m with epsilon_m i.i.d. Gaussian captures the dominant trajectory reconstruction errors.
Underlies the range-domain bearing-error model Eq. (25). A common translation b drops out of the cross-range aperture, while heading bias beta remains as a systematic error.
domain assumption TWR readout noise after N_avg averaging behaves as sigma_read approximately sigma_raw / sqrt(N_avg) with independent elementary errors (Eq. 7).
Verified by static captures, including the factor of 2.01 for N=32 versus N=128, but it assumes stationarity during the rotating sweep.
domain assumption Heading observations are aligned to range-window centers using the linear rotation-rate correction phi_corr = phi_cache + omega (t_age - T_avg/2) in Eq. (26).
Assumes constant rotation rate within each averaging window. Motorized rotation makes this plausible, but free-form motion would violate it.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Lightweight Single-Antenna Direction-of-Arrival Estimation for Curvilinear Trajectories in Mobile Embedded Systems." pith.science (2026). https://pith.science/paper/LBSF6F3J

@misc{pith2026260812029,
  author       = {Pith},
  title        = {Pith review of: Lightweight Single-Antenna Direction-of-Arrival Estimation for Curvilinear Trajectories in Mobile Embedded Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LBSF6F3J}},
  note         = {Machine review of arXiv:2608.12029}
}

read the original abstract

Accurate direction-of-arrival (DOA) estimation is valuable for spatially selective communication in noisy industrial environments. This work investigates a lightweight single-antenna framework in which receiver motion forms a virtual aperture. The receiver uses onboard inertial measurement unit (IMU) headings and two-way-ranging (TWR) measurements to a known fixed beacon, avoiding GPS, optical tracking, and high-precision external tracking of the mobile receiver. The curvilinear virtual-array MUSIC formulation is evaluated numerically using phase-coherent narrowband snapshots over arbitrary trajectories. Hardware experiments validate a range- domain TWR-IMU bearing estimator using corrected and averaged ranging observations. In a campaign of 100 consecutive four-revolution sweeps, all trials are retained. Phase-aligned accumulation recovers a 123 mm range-modulation amplitude, in close agreement with the measured 120 mm antenna lever arm. The single-sweep bearing precision is 8.2{\deg} absolute world-frame accuracy is limited by systematic BNO055 magnetometer drift in the motorized setup. The embedded bearing-estimation computation consumes 144.9 mJ per estimate, approximately 3 % of the measured cycle energy in the tested configuration. These results establish the feasibility of onboard range-domain bearing estimation and motivate future experimental validation of phase-coherent MUSIC during free-form mobile trajectories.

Figures

Figures reproduced from arXiv: 2608.12029 by the authors.

**Figure 2.** Communication process between transmitter and receiver. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Trajectory in the 2-D curved plane. C. Curvilinear Motion Linearization The curvilinear trajectory is approximated by piecewiselinear segments, as shown in [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (10 more)

**Figure 4.** Figure 4: Position reconstruction for a large-scale random curvilinear trajectory [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: Influence of SNR on MUSIC DOA estimation for three trajectory types at [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]

**Figure 6.** Figure 6: DOA RMSE versus snapshot count Nsnap ∈ {4,8,16,64,100,512} at SNR= 10 dB, f = 6 GHz, M = 6, θ0 = 30◦ , and NMC = 1000. Shaded regions denote bootstrap 95% confidence intervals. A. Influence of Signal-to-Noise Ratio [PITH_FULL_IMAGE:figures/full_fig_p006_6.png]

**Figure 8.** Figure 8: DOA RMSE versus SNR for three simulated propagation conditions [PITH_FULL_IMAGE:figures/full_fig_p007_8.png]

**Figure 9.** Figure 9: Local MUSIC DOA RMSE versus per-axis virtual-element posi [PITH_FULL_IMAGE:figures/full_fig_p007_9.png]

**Figure 11.** Figure 11: DOA RMSE versus SNR for the proposed curvilinear-VAA MUSIC [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]

**Figure 13.** Figure 13: Distribution of 1000 averaged range readings at a fixed 5 m distance [PITH_FULL_IMAGE:figures/full_fig_p009_13.png]

**Figure 14.** Figure 14: Phase-aligned average of the centered range observations across the [PITH_FULL_IMAGE:figures/full_fig_p010_14.png]

**Figure 15.** Figure 15: Zoomed current trace around the transition from ranging to embedded DOA estimation. The complete measured cycle lasts 39.7 s; only the selected [PITH_FULL_IMAGE:figures/full_fig_p011_15.png]

**Figure 16.** Figure 16: Cumulative tag and anchor energy over the zoomed DOA [PITH_FULL_IMAGE:figures/full_fig_p011_16.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 30 canonical work pages

[1]

Integration of a Distance Sen- sitive Wireless Communication Protocol to Hearing Protectors Equipped with In-Ear Microphones,

R. E. Bou Serhal, T. H. Falk, and J. V oix, “Integration of a Distance Sen- sitive Wireless Communication Protocol to Hearing Protectors Equipped with In-Ear Microphones,”Meetings Acoust., vol. 19, no. 1, May 2013

work page 2013
[2]

DOA Estimation of Coherent Signals for Uniform Linear Arrays With Mutual Coupling,

B. Liao and S. C. Chan, “DOA Estimation of Coherent Signals for Uniform Linear Arrays With Mutual Coupling,” inIEEE Int. Symp. on Circuits and Syst. (ISCAS), May 2011, pp. 377–380

work page 2011
[3]

An Improved AoA Estimation Algorithm for BLE System in the Presence of Phase Noise,

D. Xiao, S. Hu, K. Kang, and H. Qian, “An Improved AoA Estimation Algorithm for BLE System in the Presence of Phase Noise,”IEEE Trans. on Consum. Electron., vol. 69, no. 3, pp. 400–407, 2023

work page 2023
[4]

Robust Nonlinear Array Interpolation for Direction of Arrival Estimation of Highly Correlated Signals,

M. A. M. Marinho, F. Antreich, S. Caizzone, J. P. C. L. D. Costa, A. Vinel, and E. P. D. Freitas, “Robust Nonlinear Array Interpolation for Direction of Arrival Estimation of Highly Correlated Signals,”Signal Process., vol. 144, pp. 19–28, March 2018

work page 2018
[5]

Adaptive Beamforming Algorithm Based on Residual Neural Networks,

L. Chen, W. Wei, and D. Xia, “Adaptive Beamforming Algorithm Based on Residual Neural Networks,”Circuits, Systems, and Signal Process., vol. 44, no. 1, pp. 556–574, January 2025

work page 2025
[6]

High-Precision Fast Direction-of-Arrival Estimation Method for Planar Array,

S. Li, L. Li, B. Liu, Y . Song, M. Li, J. Ren, and W. Jiang, “High-Precision Fast Direction-of-Arrival Estimation Method for Planar Array,”Space: Science & Technol., vol. 3, 2023

work page 2023
[7]

Emitter Localization Using a Single Moving Observer Based on UKF,

Y . Zhu, S. Xie, M. Yang, and M. Zuo, “Emitter Localization Using a Single Moving Observer Based on UKF,” inIEEE Int. Conf. Commun. Technol. (ICCT). IEEE, October 2017, pp. 1157–1161

work page 2017
[8]

DOA Estimation with Non-Uniform Linear Arrays: A Phase-Difference Projection Approach,

H. Chen, T. Ballal, and T. Y . Al-Naffouri, “DOA Estimation with Non-Uniform Linear Arrays: A Phase-Difference Projection Approach,” vol. 10, no. 11, pp. 2435–2439, November 2021

work page 2021

Show all 30 references

[9]

A Sparse-Based Off-Grid DOA Estimation Method for Coprime Arrays,

W. Si, F. Zeng, C. Hou, and Z. Peng, “A Sparse-Based Off-Grid DOA Estimation Method for Coprime Arrays,”Sensors, vol. 18, no. 9, p. 3025, September 2018

2018
[10]

Coprime array parameters optimization for DOA estimation,

F. A. Salman and B. M. Sabbar, “Coprime array parameters optimization for DOA estimation,”Iraqi Journal of Information and Communication Technology, vol. 4, no. 2, pp. 23–32, 2021

2021
[11]

A novel dft-based doa estimation by a virtual array extension using simple multiplications for fmcw radar,

B.-J. Kim, S. Kim, and J.-W. Lee, “A novel dft-based doa estimation by a virtual array extension using simple multiplications for fmcw radar,” Sensors, vol. 18, no. 5, p. 1560, 2018

2018
[12]

Virtual Multiantenna Array for Estimating the Direction of a Transmitter: System, Bounds, and Experimental Results,

F. Quitin, P. D. Doncker, F. Horlin, and W. P. Tay, “Virtual Multiantenna Array for Estimating the Direction of a Transmitter: System, Bounds, and Experimental Results,” vol. 67, no. 2, pp. 1510–1520, February 2018

2018
[13]

Direction-of-Arrival Estimation Using Virtual Dual-Antenna Receivers: Algorithms and Controlled Ex- periments,

Y . Agram, J. Cheng, and F. Quitin, “Direction-of-Arrival Estimation Using Virtual Dual-Antenna Receivers: Algorithms and Controlled Ex- periments,” inIEEE Veh. Technol. Conf. (VTC), June 2023, pp. 1–5

2023
[14]

Direction-of-Arrival Estimation Based on Deep Neural Networks With Robustness to Array Imperfec- tions,

Z.-M. Liu, C. Zhang, and P. S. Yu, “Direction-of-Arrival Estimation Based on Deep Neural Networks With Robustness to Array Imperfec- tions,”IEEE Transactions on Antennas and Propagation, vol. 66, no. 12, pp. 7315–7327, Dec. 2018

2018
[15]

DeepMUSIC: Multiple Signal Classification via Deep Learning,

A. M. Elbir, “DeepMUSIC: Multiple Signal Classification via Deep Learning,”IEEE Sensors Lett., vol. 4, no. 4, pp. 1–4, Apr. 2020

2020
[16]

DOA Estimation Method Based on Improved Deep Convolutional Neural Network,

F. Zhao, G. Hu, C. Zhan, and Y . Zhang, “DOA Estimation Method Based on Improved Deep Convolutional Neural Network,”Sensors, vol. 22, no. 4, p. 1305, 2022

2022
[17]

Coarray Tensor Direction-of-Arrival Estimation,

H. Zheng, C. Zhou, Z. Shi, Y . Gu, and Y . D. Zhang, “Coarray Tensor Direction-of-Arrival Estimation,”IEEE Trans. on Signal Process., vol. 71, pp. 1128–1142, 2023

2023
[18]

A Hybrid Dynamic Subarray Architecture for Efficient DOA Estimation in THz Ultra-Massive Hybrid MIMO Systems,

Y . Tian, J. Ren, T. Wu, W. Liu, C. Yuen, M. Debbah, N. Al-Dhahir, M. C. Valenti, H. C. So, and Y . C. Eldar, “A Hybrid Dynamic Subarray Architecture for Efficient DOA Estimation in THz Ultra-Massive Hybrid MIMO Systems,”IEEE Trans. on Signal Process., 2026, accepted for publi...

2026 arXiv
[19]

DoA Estimation for Hybrid Receivers: Full Spatial Coverage and Successive Refinement,

A. Abdelbadie, M. Mostafa, S. Bameri, R. H. Gohary, and D. Thomas, “DoA Estimation for Hybrid Receivers: Full Spatial Coverage and Successive Refinement,”IEEE Trans. on Signal Process., vol. 72, pp. 4730–4744, 2024

2024
[20]

Efficient Methods for Unambiguous Direction-of-Arrival Estimation with Co-Prime Linear Arrays,

C. Ashok, “Efficient Methods for Unambiguous Direction-of-Arrival Estimation with Co-Prime Linear Arrays,” Ph.D. dissertation, Anna University, Chennai, India, Dec. 2022. [Online]. Available: https: //signalprocessingsociety.org/sites/default/files/uploads/newsletter/docs/ Ash...

2022
[21]

A Novel Self-Calibrated UWB-Based Indoor Localization System for Consumer Electronics Applications,

W. Ahmad, M. U. Hadi, M. B. Janjua, and S. Mumtaz, “A Novel Self-Calibrated UWB-Based Indoor Localization System for Consumer Electronics Applications,”IEEE Trans. on Consum. Electron., vol. 70, no. 1, pp. 1672–1684, 2024

2024
[22]

An Optimized Location-Based System for the Improvement of Consumer Services,

A. Khan, I. Ullah, and M. Akbar, “An Optimized Location-Based System for the Improvement of Consumer Services,”IEEE Trans. on Consum. Electron., vol. 70, no. 2, pp. 4998–5006, 2024

2024
[23]

SR10x0 Ranging Application Note (Rev. 0.2),

S. Microsystems, “SR10x0 Ranging Application Note (Rev. 0.2),” February 2022. [Online]. Available: https://www.sparkmicro.com/wp- content/uploads/2022/03/SR10x0 Ranging AN v0.2.pdf

2022
[24]

MUSIC and Improved MUSIC Algorithm to Estimate Direction of Arrival,

P. Gupta and S. P. Kar, “MUSIC and Improved MUSIC Algorithm to Estimate Direction of Arrival,” inInt. Conf. on Commun. and Signal Process. (ICCSP), April 2015, pp. 757–761

2015
[25]

Performance Analysis of MUSIC Algorithm for DOA Estimation with Varying ULA Parameters,

M. W. T. S. Chowdhury and M. Mastora, “Performance Analysis of MUSIC Algorithm for DOA Estimation with Varying ULA Parameters,” inInt. Conf. on Comput. and Inf. Technol. (ICCIT), December 2020, pp. 1–5

2020
[26]

Subspace Leakage Analysis and Improved DOA Estimation With Small Sample Size,

M. Shaghaghi and S. A. V orobyov, “Subspace Leakage Analysis and Improved DOA Estimation With Small Sample Size,”IEEE Trans. on Signal Process., vol. 63, no. 12, pp. 3251–3265, June 2015

2015
[27]

ESPRIT - Estimation of Signal Parameters Via Rotational Invariance Techniques,

R. Roy and T. Kailath, “ESPRIT - Estimation of Signal Parameters Via Rotational Invariance Techniques,” vol. 37, no. 7, pp. 984–995, 1989

1989
[28]

High-resolution frequency-wavenumber spectrum analysis,

J. Capon, “High-resolution frequency-wavenumber spectrum analysis,” Proceedings of the IEEE, vol. 57, no. 8, pp. 1408–1418, Aug. 1969

1969
[29]

H. L. Van Trees,Optimum Array Processing: Part IV of Detection, Estimation, and Modulation Theory, ser. Detection, Estimation, and Modulation Theory. New York, NY , USA: Wiley, 2002, vol. 1

2002
[30]

SR1010 and SR1020 Wireless UWB Transceivers,

S. Microsystems, “SR1010 and SR1020 Wireless UWB Transceivers,” April 2025. [Online]. Available: https://www.sparkmicro.com/product/ sr1010-sr1020-transceivers/

2025

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

Integration of a Distance Sen- sitive Wireless Communication Protocol to Hearing Protectors Equipped with In-Ear Microphones,

R. E. Bou Serhal, T. H. Falk, and J. V oix, “Integration of a Distance Sen- sitive Wireless Communication Protocol to Hearing Protectors Equipped with In-Ear Microphones,”Meetings Acoust., vol. 19, no. 1, May 2013

work page 2013

[2] [2]

DOA Estimation of Coherent Signals for Uniform Linear Arrays With Mutual Coupling,

B. Liao and S. C. Chan, “DOA Estimation of Coherent Signals for Uniform Linear Arrays With Mutual Coupling,” inIEEE Int. Symp. on Circuits and Syst. (ISCAS), May 2011, pp. 377–380

work page 2011

[3] [3]

An Improved AoA Estimation Algorithm for BLE System in the Presence of Phase Noise,

D. Xiao, S. Hu, K. Kang, and H. Qian, “An Improved AoA Estimation Algorithm for BLE System in the Presence of Phase Noise,”IEEE Trans. on Consum. Electron., vol. 69, no. 3, pp. 400–407, 2023

work page 2023

[4] [4]

Robust Nonlinear Array Interpolation for Direction of Arrival Estimation of Highly Correlated Signals,

M. A. M. Marinho, F. Antreich, S. Caizzone, J. P. C. L. D. Costa, A. Vinel, and E. P. D. Freitas, “Robust Nonlinear Array Interpolation for Direction of Arrival Estimation of Highly Correlated Signals,”Signal Process., vol. 144, pp. 19–28, March 2018

work page 2018

[5] [5]

Adaptive Beamforming Algorithm Based on Residual Neural Networks,

L. Chen, W. Wei, and D. Xia, “Adaptive Beamforming Algorithm Based on Residual Neural Networks,”Circuits, Systems, and Signal Process., vol. 44, no. 1, pp. 556–574, January 2025

work page 2025

[6] [6]

High-Precision Fast Direction-of-Arrival Estimation Method for Planar Array,

S. Li, L. Li, B. Liu, Y . Song, M. Li, J. Ren, and W. Jiang, “High-Precision Fast Direction-of-Arrival Estimation Method for Planar Array,”Space: Science & Technol., vol. 3, 2023

work page 2023

[7] [7]

Emitter Localization Using a Single Moving Observer Based on UKF,

Y . Zhu, S. Xie, M. Yang, and M. Zuo, “Emitter Localization Using a Single Moving Observer Based on UKF,” inIEEE Int. Conf. Commun. Technol. (ICCT). IEEE, October 2017, pp. 1157–1161

work page 2017

[8] [8]

DOA Estimation with Non-Uniform Linear Arrays: A Phase-Difference Projection Approach,

H. Chen, T. Ballal, and T. Y . Al-Naffouri, “DOA Estimation with Non-Uniform Linear Arrays: A Phase-Difference Projection Approach,” vol. 10, no. 11, pp. 2435–2439, November 2021

work page 2021