Pith. sign in

REVIEW 3 major objections 4 minor 36 references

Joint Synchronization and Sensing in Networked ISAC via Structured Canonical Polyadic Decomposition

T0 review · 3 major / 4 minor · reviewed 2026-08-01 · deepseek-v4-flash

Pith's one-line read SCPD, a structured tensor decomposition, jointly synchronizes networked ISAC and estimates target parameters, asymptotically reaching the Cramér–Rao bound.

desk verdict Useful joint sync-and-sensing tensor framework, but the headline CRB claim is not actually established: the paper's own two-step solver breaks the MLE premise. read the letter →

arxiv 2607.18680 v1 pith:2HIVOPJ5 submitted 2026-07-21 eess.SP

classification eess.SP
keywords networkedISACsynchronizationcanonicalpolyadicdecompositionVandermondestructureCramér–Raoboundtargettrackingadaptivebeamformingtensor
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

Networked integrated sensing and communication (ISAC) lets multiple base stations cooperate to sense targets, but this requires accurate time and frequency synchronization. The paper proposes SCPD (structured canonical polyadic decomposition), which decomposes the received sensing tensor into Vandermonde-structured factor matrices so that each multipath component is separated from the others. From the separated components, SCPD jointly estimates timing and carrier-frequency offsets between each base-station pair, along with target delays, Doppler shifts, and angles. The paper further shows how estimates from different pairs can be associated and fused to track target trajectories and velocities, with an adaptive beamformer steering the next snapshot. The paper claims SCPD asymptotically attains the Cramér–Rao bound, the theoretical lower bound on unbiased estimator variance.

What carries the argument

The central object is the structured canonical polyadic decomposition (SCPD): a CPD of the received signal tensor in which every factor matrix is constrained to be Vandermonde, meaning each column is a geometric sinusoid with a distinct generator (frequency). The solver is a constrained alternating least-squares (CALS) algorithm that alternates between least-squares updates and a projection onto the Vandermonde set. The Vandermonde generators carry the physical parameters: delay plus offset, Doppler plus CFO, and angles; bistatic reciprocity between the two links of a base-station pair is used to cancel offsets and recover target parameters.

What would settle it

Compute the error covariance of the (22)-(25) two-step estimate for a single Vandermonde column in Gaussian noise and compare it with the Cramér–Rao bound; if the covariance exceeds the bound, the claim fails. Also, a Monte Carlo with artificially correlated residuals would reveal whether the empirical CRB tracking in Figs. 2–6 is an artifact of the i.i.d. noise assumption.

Watch

Extended reading notes

Core claim

The central claim is that enforcing Vandermonde structure on the factor matrices of a canonical polyadic decomposition of the received sensing signal separates the multipath components of the channel, and that this separation enables one tensor factorization to estimate both synchronization offsets (timing offset and carrier frequency offset per base-station pair) and target parameters (bistatic delay, Doppler shift, and angles) with an accuracy that asymptotically reaches the Cramér–Rao bound. The paper also establishes identifiability conditions under which the decomposition is unique, and it extends the pairwise estimates to a tracking algorithm that recovers 2D trajectories and velocitie

Load-bearing premise

The claim that SCPD asymptotically attains the Cramér–Rao bound rests on the assumption that the errors left after its approximate two-step (least-squares plus Vandermonde projection) update are independent and identically distributed Gaussian; if that assumption fails, the estimator is no longer maximum likelihood and the CRB argument does not follow.

Editorial extensions

If this is right

  • Synchronization and sensing happen in a single decomposition step, removing the need for a separate calibration phase or dedicated synchronization pilots.
  • Because multipath components are separated, the inter-path interference that corrupts compressed-vector methods like SOE-MP is avoided.
  • If the Cramér–Rao bound is attained, no unbiased estimator of these parameters can achieve lower error variance, so the method is statistically optimal.
  • The tracking and adaptive beamforming loop keeps a network continuously focused on targets across snapshots, with a consistency check to discard outliers.
  • The identifiability conditions give a priori guarantees on when the decomposition is unique, and therefore when the parameter estimates are unambiguous.

Reading between the lines

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

  • The CRB-optimality claim is only as strong as the MLE premise; the paper's two-step solver (unconstrained least squares followed by Vandermonde projection) is only shown to be an MLE under an i.i.d. Gaussian error assumption that the structured least-squares step does not obviously guarantee, so a direct covariance analysis of the solver is a natural next check.
  • The same Vandermonde-constrained tensor decomposition could apply to other multi-dimensional harmonic retrieval problems, such as MIMO channel estimation or monostatic radar with multiple snapshots, wherever the measurement tensor is a sum of sinusoids.
  • The paper assumes disjoint bandwidth parts per base station to avoid interference; an extension where base stations share the same bandwidth would need to handle cross-link interference inside the tensor model.
  • The tracking algorithm's outlier rejection and consistency check are heuristic; a probabilistic data-association or Bayesian filter could be tested against them in the same simulation setup.
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.

Referee Report

3 major / 4 minor

Summary. The paper proposes a tensor-based framework for joint time-frequency synchronization and multi-target sensing in networked ISAC. Each BS pair forms a third-order tensor from received MIMO-OFDM pilot signals, decomposes it via a structured canonical polyadic decomposition (SCPD) with Vandermonde factor constraints, and reads off offset and target parameters using bistatic reciprocity. The authors derive identifiability conditions for SCPD, compute a Slepian–Bangs CRB, and present a multi-target tracking algorithm with adaptive beamforming. Simulations compare SCPD against SOE-MP, ESPRIT-LS, and CP-VDM, reporting RMSE gains and success rates.

Significance. If the claims are upheld, the framework would be a useful contribution: it addresses inter-path interference in networked synchronization, exploits Vandermonde structure in a principled way, and fuses estimates from multiple BS pairs for trajectory and velocity tracking. The paper is explicit about the signal model, the reciprocity relations, and the CRB computation, which makes the derivations checkable. The simulation study is extensive and compares against three sensible baselines. However, the central theoretical claim—that SCPD asymptotically achieves the CRB—is not established by the arguments in the manuscript; the evidence offered is empirical and conditional on trial success. This is a load-bearing gap because the abstract and introduction foreground the CRB claim.

major comments (3)
  1. [Sec. IV-A, Eqs. (21)-(25)] The paper claims that the projection step (24) 'constitutes a maximum likelihood estimator' under i.i.d. circular-Gaussian entries of the estimation error matrix, and uses this to assert SCPD asymptotically attains the CRB. This premise is not valid for the algorithm's own output. In subproblem (21)-(22), the unconstrained LS estimate has error covariance proportional to (D^H D)^{-1} ⊗ I, where D is a Khatri-Rao product of Vandermonde factors. Since (D^H D)^{-1} is generally non-diagonal, the entries of the estimate fed into (23)-(24) are statistically correlated and non-identically distributed. Thus (24) is not the MLE of the original measurement model, and the chain 'CALS → MLE → asymptotic CRB' is not established. The paper itself labels the solution to (21) as 'approximate yet efficient'; no theorem shows that the fixed point of the alternating projections is the joint MLE or that it
  2. [Sec. VI-A, Figs. 2-6] The RMSE curves are averaged only over successful Monte Carlo trials, while the CRB is an unconditional performance bound. The text states: 'the curves in Figs. 2-6 are smoothed by averaging only over successful trials and excluding the failed ones.' Table I reports success rates well below 100% (e.g., 93.73% for SCPD at L=2, SNR=-25 dB). Conditional averaging removes the estimation outliers that contribute to the mean squared error, so the comparison against the CRB is biased in favor of the proposed method. To support 'SCPD approaches the CRB,' the authors should report unconditional RMSE or explicitly frame the curves as conditional on successful detection/association, and discuss how the CRB comparison should be interpreted in that case.
  3. [Sec. VI-A, text after Fig. 2] The statement 'Under mild regularity conditions, maximum likelihood estimates are asymptotically unbiased and capable of reaching the CRB' is invoked to bridge the simulation results to the asymptotic claim. This inference depends entirely on the MLE status of SCPD, which is the same unproven premise as in Section IV-A. The citation [21] is to a tensor-based channel estimation paper and does not supply a theorem for the CALS algorithm here. The empirical curves alone cannot establish the asymptotic claim; they are consistent with the algorithm being a good estimator but not with CRB attainment in the unconditional sense.
minor comments (4)
  1. [Sec. IV-A, Algorithm 1] The stopping criterion |ε_{r+1}-ε_r|/ε_r < δ can be undefined when ε_r=0; a safeguard such as checking ε_r > 0 or using an absolute tolerance would make the algorithm description more robust.
  2. [Eq. (26)] The notation M(·) is defined as the mean, but in (26) it is applied to a vector difference of frequency estimates. Since the text later uses V(·) for variance, it would help to state explicitly that the mean is taken over the L target indices.
  3. [Sec. IV-B, Lemma 1] The uniqueness conditions in (28) are stated as existence conditions on integer pairs {P_t,Q_t}. It would improve readability to give an example of how such pairs are chosen for the standard tensor dimensions, since the proof relies on the spatial smoothing construction without providing a practical selection rule.
  4. [Sec. VI-A, Table I] The success-rate table is informative, but the definition of 'successful trial' is tied to the bistatic-range check in (51). This is a reasonable practical criterion, yet it is not the same as 'the estimator converged to the true parameter'; the conditional RMSE should be interpreted accordingly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SCPD derives estimates from factor matrices and compares to an independently computed CRB; the MLE premise issue is a correctness gap, not circularity.

full rationale

The paper's central derivation is not circular. The SCPD estimator in Algorithm 1 solves (18) by alternating constrained least squares, and the final parameter estimates in (26) are read off the estimated Vandermonde factor matrices using the forward model (17) and the bistatic reciprocity (5). No parameter is fitted to its own prediction, and no output quantity is defined in terms of the claimed result. The CRB in (37)-(43) is a standard Slepian-Bangs computation from the Gaussian measurement model (1); it is independent of the algorithm, so comparing RMSE curves against it is an external benchmark, not a self-fulfilling reduction. The identifiability Lemma 1 is argued from the Vandermonde structure and spatial smoothing, not imported from a same-author uniqueness theorem. The self-citations [18] and [19] motivate the tensor signal model and the pairwise-synchronization idea, but they are not load-bearing for the CRB claim or uniqueness. The main validity concern is the MLE premise in Section IV-A: (24) is called an MLE under i.i.d. Gaussian estimation error, but the error of the LS estimate (22) has covariance proportional to (D^H D)^-1 tensor I, which is colored for the Khatri-Rao Vandermonde D used in Algorithm 1. That makes the chain 'SCPD is MLE, hence asymptotic CRB' unsupported. This is a correctness gap in the theoretical claim, not circularity: the paper does not define the CRB in terms of the estimator, nor does it fit parameters to its own predictions. Smoothing RMSE only over successful trials while plotting the unconditional CRB biases the empirical comparison, but it is not a constructional equivalence. Therefore no circular step meets the standard of 'Eq. X = Eq. Y by construction', and the circularity score is 0.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The central claim rests on the standard ISAC tensor model (17) plus two domain assumptions the authors flag only lightly: exact bistatic reciprocity (5) and known non-zero pilots. The CRB/MLE claim additionally leans on a Gaussian-error assumption that the algorithm's own two-step design violates. Algorithm hyperparameters (J_o=1, η=15 m) are hand-set without sensitivity analysis; L is assumed known. No invented entities. The uniqueness theorem reduces to known Vandermonde-CPD theory [20].

free parameters (3)
  • J_o (assumed number of outlier measurements) = 1
    Hand-chosen input to Algorithm 2 (Section VI simulation settings). Affects the trimming in (52); not fitted or justified from data. No sensitivity analysis is given.
  • η (consistency-check displacement threshold) = 15 m
    Hand-chosen in Section VI settings; controls when tracking results are replaced by the previous snapshot. No sensitivity analysis is given.
  • CALS stop parameters R, δ = R=20, δ=10^-4
    Convergence settings in Section VI; results may depend on these at low SNR, and no convergence or sensitivity study is provided.
assumptions (6)
  • domain assumption Bistatic reciprocity of TOs and CFOs: Δτ_j1 = −Δτ_j2, Δν_j1 = −Δν_j2 (eq. 5)
    Load-bearing for all offset estimates in (26) and for the parameter reduction in the CRB (Section IV-C). Cited from [10],[15]; treated as an exact 'inherent' property. Tolerances/imperfections are not modeled.
  • domain assumption Known, non-zero pilot symbols s_{i,n,k} for element-wise division in (13)/(16)
    The CPD model (16) is formed by dividing the received tensor by the pilot tensor; zero or low-power pilot symbols would amplify noise and break the model.
  • domain assumption Number of targets L is known a priori
    L is an input to Algorithms 1 and 2; the CPD rank and all permutation/data-association searches (27), (47), (49) assume it is known. No model-order selection is provided.
  • domain assumption Joint i.i.d. circular Gaussian noise model for the MLE and CRB claims
    Used in Section IV-C (38) for the CRB and in Section IV-A for the MLE claim. The paper itself notes the MLE claim requires i.i.d. errors on the factor-matrix estimate, which the two-step procedure does not deliver.
  • domain assumption Disjoint bandwidth parts per BS so transmissions can be separated by BWP index
    Inherited from [3] (Section II). Enables each BS to separate all links; requires receive capability over all BWPs simultaneously (implied full-duplex / wideband listening).
  • standard math Vandermonde-structure identifiability machinery (spatial smoothing) of Sørensen & De Lathauwer
    Lemma 1 is proven by invoking the spatial-smoothing technique and equations (28)-(35) of [20]; correctness of the lemma therefore imports the correctness of [20].

how reviews work

0 comments
Cite this review

Pith. "Pith review of Joint Synchronization and Sensing in Networked ISAC via Structured Canonical Polyadic Decomposition." pith.science (2026). https://pith.science/paper/2HIVOPJ5

@misc{pith2026260718680,
  author       = {Pith},
  title        = {Pith review of: Joint Synchronization and Sensing in Networked ISAC via Structured Canonical Polyadic Decomposition},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2HIVOPJ5}},
  note         = {Machine review of arXiv:2607.18680}
}
read the original abstract

Networked integrated sensing and communication (ISAC) offers significant potential for next-generation wireless systems. By exploiting spatial diversity through the cooperation of multiple base stations (BSs), this architecture expands coverage and achieves enhanced sensing performance. However, accurate sensing in networked ISAC requires time-frequency synchronization among BSs. Existing synchronization methods for networked ISAC suffer from inter-path interference caused by sensing channel compression. To address this problem, this paper proposes a structured canonical polyadic decomposition (SCPD) algorithm that effectively separates the multipath components of the sensing channel. Benefiting from this separation, SCPD achieves joint network-level synchronization and multi-target parameter estimation. We establish theoretical identifiability conditions for SCPD and show that it asymptotically achieves the Cram\'{e}r-Rao bound. Furthermore, by incorporating parameters estimated from different BS pairs, we propose a multi-target tracking algorithm designed for the continuous operation of the system. The proposed algorithm tracks both the trajectories and velocities of moving targets by leveraging geometric diversity. Utilizing tracking results from the previous snapshot, an adaptive beamforming scheme is also developed to improve tracking performance in the next snapshot. Simulation results demonstrate that the proposed algorithms achieve superior accuracy and outlier robustness for both synchronization and sensing in networked ISAC, outperforming traditional approaches.

Figures

Figures reproduced from arXiv: 2607.18680 by the authors.

Figure 1
Figure 1. A networked ISAC system with multi-BSs collaboratively performing [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. RMSE of parameter estimation versus SNR in the scenario where [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. RMSE of parameter estimation versus SNR in the scenario where [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: RMSE of parameter estimation versus antenna number [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: RMSE of parameter estimation versus OFDM symbol number [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: RMSE of parameter estimation versus subcarrier number [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Trajectory tracking results in the scenario where SNR [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Velocity tracking results in the scenario where SNR [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Trajectory tracking results in the scenario where SNR [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Velocity tracking results in the scenario where SNR [PITH_FULL_IMAGE:figures/full_fig_p011_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 1 linked inside Pith

  1. [19]

    Pairwise base station synchronization via joint tensor decomposition for networked ISAC,

    L. Chen, Y . Liang, and H. Li, “Pairwise base station synchronization via joint tensor decomposition for networked ISAC,”IEEE Wireless Commun. Lett., vol. 15, pp. 1990–1994, 2026

  2. [21]

    Tensor-based channel estimation for dual-polarized massive MIMO systems,

    C. Qian, X. Fu, N. D. Sidiropoulos, and Y . Yang, “Tensor-based channel estimation for dual-polarized massive MIMO systems,”IEEE Trans. Signal Process., vol. 66, no. 24, pp. 6390–6403, 2018

  3. [1]

    Cooperative ISAC networks: Opportunities and challenges,

    K. Meng, C. Masouros, A. P. Petropulu, and L. Hanzo, “Cooperative ISAC networks: Opportunities and challenges,”IEEE Wireless Com- mun., vol. 32, no. 3, pp. 212–219, 2025

  4. [2]

    Cooperative multi-static ISAC networks: A unified design framework for active and passive sensing,

    Y . Yang, Z. Li, J. Zhao, Q. Wu, Z. Wei, W. Chen, and W. Jia, “Cooperative multi-static ISAC networks: A unified design framework for active and passive sensing,”IEEE Trans. Wireless Commun., vol. 25, pp. 5401–5415, 2026

  5. [3]

    Target localization in cooperative ISAC systems: A scheme based on 5G NR OFDM signals,

    Z. Zhang, H. Ren, C. Pan, S. Hong, D. Wang, J. Wang, and X. You, “Target localization in cooperative ISAC systems: A scheme based on 5G NR OFDM signals,”IEEE Trans. Commun., vol. 73, no. 5, pp. 3562– 3578, 2025

  6. [4]

    Cooperative multistatic sensing using 5G NR PRS signal,

    A. Sahoo, A. Sonny, S. Blandino, T. Ropitault, D. Griffith, T. Nguyen, and N. Golmie, “Cooperative multistatic sensing using 5G NR PRS signal,” inProc. IEEE Globecom Workshops, Taipei, Taiwan, 2025

  7. [5]

    Integrated sensing and communication signals toward 5G-A and 6G: A survey,

    Z. Wei, H. Qu, Y . Wang, X. Yuan, H. Wu, Y . Du, K. Han, N. Zhang, and Z. Feng, “Integrated sensing and communication signals toward 5G-A and 6G: A survey,”IEEE Internet Things J., vol. 10, no. 13, pp. 11 068– 11 092, 2023

  8. [6]

    Cooperative integrated sensing and communication in 6G: From operators perspective,

    X. Wang, Z. Han, R. Xi, G. Liu, L. Han, J. Jin, Y . Xue, L. Ma, Y . Wang, T. Jianget al., “Cooperative integrated sensing and communication in 6G: From operators perspective,”IEEE Wireless Commun., vol. 32, no. 1, pp. 52–59, 2025

Show all 36 references
  1. [7]

    Joint location and velocity estimation and fundamental CRLB analysis for cell-free MIMO-ISAC,

    G. Xia, P. Xiao, Q. Luo, B. Ji, Y . Zhang, and H. Zhou, “Joint location and velocity estimation and fundamental CRLB analysis for cell-free MIMO-ISAC,”IEEE Trans. Commun., vol. 74, pp. 7357–7374, 2026

  2. [8]

    Coordinated transmit beamforming for networked ISAC with imperfect CSI and time synchronization,

    X. Yang, Z. Wei, J. Xu, Y . Fang, H. Wu, and Z. Feng, “Coordinated transmit beamforming for networked ISAC with imperfect CSI and time synchronization,”IEEE Trans. Wireless Commun., vol. 23, no. 12, p. 18019–18035, 2024

  3. [9]

    Wireless time synchroniza- tion for multi-static sensing in JCAS systems,

    N. M. Idrees, C. Desset, and A. Bourdoux, “Wireless time synchroniza- tion for multi-static sensing in JCAS systems,” inProc. IEEE Radar Conf., 2025, pp. 888–893

  4. [10]

    Over-the-air time-frequency synchronization in distributed ISAC systems,

    K. Han, K. Meng, and C. Masouros, “Over-the-air time-frequency synchronization in distributed ISAC systems,” 2025, arXiv:2503.08920

  5. [11]

    On the feasibility of distributed beamforming in wireless networks,

    R. Mudumbai, G. Barriac, and U. Madhow, “On the feasibility of distributed beamforming in wireless networks,”IEEE Trans. Wireless Commun., vol. 6, no. 5, pp. 1754–1763, 2007

  6. [12]

    Time synchronization for ZigBee networks,

    D. Cox, E. Jovanov, and A. Milenkovic, “Time synchronization for ZigBee networks,” inProc. 37th Southeastern Symp. Syst. Theory, 2005, pp. 135–138

  7. [13]

    Network synchronization for dense small cell networks,

    K. J. Zou, K. W. Yang, M. Wang, B. Ren, J. Hu, J. Zhang, M. Hua, and X. You, “Network synchronization for dense small cell networks,” IEEE Trans. Commun., vol. 22, no. 2, pp. 108–117, 2015

  8. [14]

    Sensing in bistatic ISAC systems with clock asynchronism: A signal processing perspective,

    K. Wu, J. Pegoraro, F. Meneghello, J. A. Zhang, J. O. Lacruz, J. Widmer, F. Restuccia, M. Rossi, X. Huang, D. Zhang, G. Caire, and Y . J. Guo, “Sensing in bistatic ISAC systems with clock asynchronism: A signal processing perspective,”IEEE Signal Process. Mag., vol. 41, no. 5,...

  9. [15]

    N. J. Willis,Bistatic Radar, vol. 2. Chennai, India: SciTech Publishing, 2005

  10. [16]

    Low- rank tensor decomposition-aided channel estimation for millimeter wave MIMO-OFDM systems,

    Z. Zhou, J. Fang, L. Yang, H. Li, Z. Chen, and R. S. Blum, “Low- rank tensor decomposition-aided channel estimation for millimeter wave MIMO-OFDM systems,”IEEE J. Sel. Areas Commun., vol. 35, no. 7, pp. 1524–1538, 2017

  11. [17]

    Integrated sensing and communication with massive MIMO: A unified tensor approach for channel and target parameter estimation,

    R. Zhang, L. Cheng, S. Wang, Y . Lou, Y . Gao, W. Wu, and D. W. K. Ng, “Integrated sensing and communication with massive MIMO: A unified tensor approach for channel and target parameter estimation,” IEEE Trans. Wireless Commun., vol. 23, no. 8, pp. 8571–8587, 2024

  12. [18]

    Tensor-based target sens- ing for resource-irregular ISAC systems,

    L. Chen, L. Ge, X. Jiang, Z. Jiang, and H. Li, “Tensor-based target sens- ing for resource-irregular ISAC systems,”IEEE Trans. Signal Process., vol. 74, pp. 605–621, 2026

  13. [20]

    Blind signal separation via tensor decomposition with Vandermonde factor: Canonical polyadic decompo- sition,

    M. Sørensen and L. D. Lathauwer, “Blind signal separation via tensor decomposition with Vandermonde factor: Canonical polyadic decompo- sition,”IEEE Trans. Signal Process., vol. 61, no. 22, pp. 5507–5519, 2013

  14. [22]

    Joint range-velocity- azimuth estimation for OFDM-based integrated sensing and communi- cation,

    Z. Hu, Q. Ye, Y . Huang, S. Hu, and G. Yang, “Joint range-velocity- azimuth estimation for OFDM-based integrated sensing and communi- cation,”IEEE Trans. Wireless Commun., vol. 23, no. 10, pp. 12 933– 12 948, 2024

  15. [23]

    ISAR sensing based on MUSIC algorithm in integrated sensing and commu- nications,

    R. Zhang, Z. Wang, Z. Wei, Y . Liu, Z. Xiong, and Z. Feng, “ISAR sensing based on MUSIC algorithm in integrated sensing and commu- nications,”IEEE Trans. Cogn. Commun. Netw., vol. 11, no. 5, pp. 3292– 3305, 2025

  16. [24]

    An ESPRIT-based moving target sensing method for MIMO-OFDM ISAC systems,

    Y . Xiang, Y . Gao, X. Yang, S. Kang, and M. Shao, “An ESPRIT-based moving target sensing method for MIMO-OFDM ISAC systems,”IEEE Commun. Lett., vol. 27, no. 12, pp. 3205–3209, 2023

  17. [25]

    Integrated sensing and communication with mmWave massive MIMO: A compressed sampling perspective,

    Z. Gao, Z. Wan, D. Zheng, S. Tan, C. Masouros, D. W. K. Ng, and S. Chen, “Integrated sensing and communication with mmWave massive MIMO: A compressed sampling perspective,”IEEE Trans. Wireless Commun., vol. 22, no. 3, pp. 1745–1762, 2022

  18. [26]

    Sparse orthogonal matching pursuit-based parameter estimation for integrated sensing and communications,

    N.-S. Duong, K.-H. Ngo, T.-M. Dinh, and V .-L. Nguyen, “Sparse orthogonal matching pursuit-based parameter estimation for integrated sensing and communications,” inProc. IEEE INFOCOM Conf. Comput. Commun. Workshops, 2025, pp. 1–6

  19. [27]

    Moving target detection using distributed MIMO radar in clutter with nonhomogeneous power,

    P. Wang, H. Li, and B. Himed, “Moving target detection using distributed MIMO radar in clutter with nonhomogeneous power,”IEEE Trans. Signal Process., vol. 59, no. 10, pp. 4809–4820, 2011

  20. [28]

    Target localization geometry gain in distributed MIMO radar,

    M. Sadeghi, F. Behnia, R. Amiri, and A. Farina, “Target localization geometry gain in distributed MIMO radar,”IEEE Trans. Signal Process., vol. 69, pp. 1642–1652, 2021

  21. [29]

    Distributed localization of target for MIMO radar with widely separated directional transmitters and omnidirectional receivers,

    K. Xiong, G. Cui, W. Yi, S. Wang, and L. Kong, “Distributed localization of target for MIMO radar with widely separated directional transmitters and omnidirectional receivers,”IEEE Trans. Aerosp. Electron. Syst., vol. 59, no. 3, pp. 3171–3187, 2022

  22. [30]

    Joint detection and localiza- tion in distributed MIMO radars employing waveforms with imperfect auto-and cross-correlation,

    Y . Lai, L. Venturino, E. Grossi, and W. Yi, “Joint detection and localiza- tion in distributed MIMO radars employing waveforms with imperfect auto-and cross-correlation,”IEEE Trans. Veh. Technol., vol. 72, no. 12, pp. 16 524–16 537, 2023

  23. [31]

    S. M. Kay,Fundamentals of Statistical Signal Processing: Estimation Theory, Hoboken, NJ, USA: Prentice-Hall, Inc., 1993

  24. [32]

    Higher-order SVD-based subspace estimation to improve the parameter estimation accuracy in multidimensional harmonic retrieval problems,

    M. Haardt, F. Roemer, and G. D. Galdo, “Higher-order SVD-based subspace estimation to improve the parameter estimation accuracy in multidimensional harmonic retrieval problems,”IEEE Trans. Signal Process., vol. 56, no. 7, pp. 3198–3213, 2008

  25. [33]

    Using the matrix pencil method to estimate the parameters of a sum of complex exponentials,

    T. K. Sarkar and O. Pereira, “Using the matrix pencil method to estimate the parameters of a sum of complex exponentials,”IEEE Antennas Propag. Mag., vol. 37, no. 1, pp. 48–55, 1995

  26. [34]

    Unitary PUMA algorithm for estimating the frequency of a complex sinusoid,

    C. Qian, L. Huang, H.-C. So, N. D. Sidiropoulos, and J. Xie, “Unitary PUMA algorithm for estimating the frequency of a complex sinusoid,” IEEE Trans. Signal Process., vol. 63, no. 20, pp. 5358–5368, 2015

  27. [35]

    An algorithm for tracking multiple targets,

    D. Reid, “An algorithm for tracking multiple targets,”IEEE Trans. Autom. Control, vol. 24, no. 6, pp. 843–854, 2003

  28. [36]

    Robust LOS identification for passive multi-target localization in multipath obstructed environments,

    Y . Liang and H. Li, “Robust LOS identification for passive multi-target localization in multipath obstructed environments,”IEEE Trans. Signal Inf. Process. Netw., vol. 11, pp. 1030–1043, 2025

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.