Received Signal Strength Based Wireless Source Localization with Inaccurate Anchor Position

Guojun Han; Jing Chen; Yang Liu; Yonghua Wang; Zheng Xue

REVIEW 4 major objections 6 minor 45 references

Received Signal Strength Based Wireless Source Localization with Inaccurate Anchor Position

T0 review · 4 major / 6 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read The paper proposes an RSS-based source localization method that stays accurate when anchor positions are only known within a bound, using a worst-case semidefinite relaxation and joint source-and-anchor rounding.

desk verdict The paper's core rounding idea is worth attention, but the derivation changes the estimator mid-stream and the main SDP has a sign error, so the reported simulation results cannot be tied to the stated method without code or a corrected formulation. read the letter →

arxiv 1908.03202 v5 pith:JKX55R44 submitted 2019-08-09 eess.SP

Yang Liu , Guojun Han , Yonghua Wang , Zheng Xue , Jing Chen This is my paper · ORCID

classification eess.SP

keywords RSSlocalizationanchorpositionuncertaintyrobustestimationmin-maxoptimizationsemidefiniteprogrammingroundingalgorithmworst-casedesignreceivedsignalstrength

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

Received-signal-strength (RSS) localization usually assumes the anchor nodes' positions are exact, but real deployments often have only roughly surveyed anchors. This paper tries to close that gap with an estimator that treats each anchor error as bounded by $\zeta$ and assumes nothing else about its distribution. The authors derive a worst-case, min-max approximation of the maximum-likelihood objective, relax it into a semidefinite program (a convex optimization problem over matrix variables) that is solvable without a starting point, and add a rounding algorithm that searches candidate source locations and candidate anchor locations together. In Monte Carlo simulations with both random and planned anchor layouts, the combined method achieves the lowest root-mean-square error among the compared RSS, SOCP, and distance-based estimators across anchor error, measurement noise, and anchor count. If the claim holds, a practical system can localize accurately with imperfect anchor positions and no knowledge of the anchor error distribution.

What carries the argument

The load-bearing object is the min-max reformulation of the RSS objective: replace the original sum of squared log-ratio terms by $\max_i \left|\log_{10}\left((\|x-\hat z_i\|-\delta_i)^2/\beta_i^2\right)\right|$ after a first-order Taylor expansion of the anchor error. This makes the worst case over anchor perturbations reduce to a known bound $\zeta$ on $\delta_i$, so the nonconvex problem becomes a semidefinite program with variables $x$, scalar $k$, distance vector $l$, and matrices $X=xx^T$, $L=ll^T$ relaxed to convex matrix constraints. The second mechanism is the joint rounding algorithm: it uses $X^* - x^* x^{*T}$ as a covariance to sample source candidates and uniformly samples anchor candidates inside each $\zeta$-ball, selecting the pair with the lowest RSS residual. The combination is what converts a loose convex relaxation into a usable point estimate.

What would settle it

Run a dense grid search over a small two-dimensional configuration with noiseless RSS, known anchor error bound $\zeta$, and anchor positions that satisfy $\|x-\hat z_i\|>\zeta$; compare the global minimum of the original sum-of-squared log-ratio objective (Eq. (8)) with the output of the SDP plus rounding. If for some geometry the rounded SDP solution lies farther from the true source than the grid's best point by more than the grid resolution, the relaxation demonstrably solves a different problem than the announced robust estimator.

Watch

Extended reading notes

Core claim

The central claim is that inaccurate anchor positions need not be modeled statistically for robust RSS source localization: a bounded worst-case model suffices. The paper starts from the log-normal shadowing model, writes the maximum-likelihood estimate as minimizing the sum of squared log-ratio terms, then Taylor-expands the anchor perturbation (Eq. (9)) and converts the sum into a max over anchors of absolute log-ratios (Eq. (13)). This converts the uncertainty set into an additive bound $\pm\zeta$ on each distance ratio, allowing a semidefinite relaxation (Eq. (25)) with linear matrix inequality constraints. The final step is a rounding algorithm that samples candidate source positions from the relaxed covariance and candidate anchors uniformly inside each error ball, then picks the combination whose simulated RSS residuals are smallest. The paper's simulations report this 'r-r' pipeline as the best performer in root-mean-square error among all compared methods.

Load-bearing premise

The method stands on the assumption that the first-order Taylor expansion of anchor error plus the switch to a max-over-anchors absolute log-ratio objective leaves the optimum close enough to the original robust maximum-likelihood estimate that the relaxed problem still solves the localization task.

Editorial extensions

If this is right

Field deployments can relax the requirement of precisely surveyed anchors: a conservative bound $\zeta$ on anchor error is enough to run the estimator.
Since the SDP relaxation is convex, the method needs no initialization, removing the local-minimum dependence that makes the ML baseline impractical.
The relaxation also yields a covariance-type uncertainty measure $X^* - x^* x^{*T}$, so the estimator can report a rough confidence region along with the point estimate.
The joint rounding over source and anchor candidates is what recovers accuracy under random anchor layouts; planned layouts already give a low error before rounding because the source tends to lie inside the region enclosed by the anchors (the convex hull).
Across the simulated range of anchor error, measurement noise, and anchor count, the combined pipeline is reported to have the smallest RMSE and the narrowest error distribution among the compared methods.

Reading between the lines

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

The same worst-case relaxation could be extended to unknown path-loss parameters or transmit power, since the paper notes propagation-parameter self-estimation as future work; the $\zeta$-ball treatment of anchor error is a template for other bounded model errors.
The joint rounding algorithm could be repurposed as an anchor-calibration tool: when the source position is known, the selected candidate anchors form a corrected anchor map, a use the paper does not explore.
Because the method replaces the sum-of-squares ML objective with an $\ell^\infty$ (max) objective, it likely trades statistical efficiency for robustness; comparing full minimax SDP estimates against the $\ell^\infty$ version on the same data would quantify that trade.
The uniform sampling of anchor candidates inside the $\zeta$-ball means accuracy should improve as more candidate anchors per anchor are drawn; a testable extension is to measure RMSE as a function of the sample count $N$ to find the point of diminishing returns.

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

Quick take: the paper has a plausible core idea, but the derivation does not support the estimator it announces, and the sign error in the main SDP means the reported simulations may not be solving the stated problem.

What is actually new: the bounded-error min-max formulation for RSS localization with unknown anchor positions, and the rounding algorithm (Alg. 4) that searches over candidate anchor positions and source positions together. That rounding idea is a reasonable extension of the usual randomization/grid-search rounding, and selecting the best combination by fitting the measured RSS vector is sensible. The paper also does the standard empirical work well: it sweeps anchor error bound, noise variance, and number of anchors; it separates random and planned anchor layouts; and it includes boxplots. The related work is broad and the writing is readable.

Soft spots, in order of importance. First, Eqs. (12) to (13) quietly replace the sum over anchors of squared log-ratios with a max over anchors of absolute log-ratios. That is a different estimator—a Chebyshev-type fit rather than the announced ML fit. No argument is given that the two objectives share a minimizer or comparable behavior, so the abstract's phrase "maximum likelihood estimator" is not supported. Second, Eq. (25) has signs that do not match Eq. (18): the first constraint uses tr(X) + 2x^T z_i with a plus, while ||x - z_i||^2 = tr(X) - 2x^T z_i + z_i^T z_i with a minus. The printed SDP constrains ||x + z_i||^2 up to rounding terms. Also X is stated in R^{M x M} after being defined in R^2. If the simulations used the printed signs, they solved a different, likely wrong problem; if they used corrected signs, the manuscript is mis-specified. No code or data are provided to disambiguate. Third, the abstract and introduction promise a real experiment, but the full text stops at simulations. Fourth, the closest robust-RSS baselines, especially [22] and [29], are cited but not compared numerically, so the "outperforms all" claim is under-supported.

The simulation study is internally consistent, but it validates the method on data generated from the same log-normal model used to derive it; that is common for method papers and not damning by itself, but it makes the missing code and data more important. The core heuristic could still work; the problem is that the link between the announced estimator and the actual SDP is broken on the page.

Who will get value: researchers working on RSS localization and convex relaxations. If the derivation is corrected and code released, the rounding algorithm and bounded-error formulation would be worth citing. As it stands, I would not rely on the numerical claims, but the paper is serious and the issues are repairable. A serious referee should see it. Recommendation: send to peer review with major revision required—correct the derivation, release code and data, either run the promised real experiment or remove the claim, and compare directly with robust RSS baselines.

Referee Report

4 major / 6 minor

Summary. The manuscript studies RSS-based source localization when anchor positions are known only up to a bounded error norm ζ. Starting from a log-normal path-loss model, the authors propose a min-max ('robust ML') estimator, transform it by Taylor expansion into a scalar radial uncertainty model, relax the resulting nonconvex problem to an SDP by dropping rank-one constraints, and develop three rounding schemes, including one that also refines anchor candidate positions. Simulations under random and planned anchor deployments compare the proposed Robust-RSS estimators with existing SDP, SOCP, distance-based, and ML methods. The paper claims that the proposed estimator with anchor-aware rounding (r-r) achieves the lowest RMSE under anchor position uncertainty and RSS measurement noise.

Significance. If correct, the paper would provide a practical contribution: an SDP-based RSS localization method that requires no distributional assumption on anchor errors and that can be solved efficiently with standard SDP solvers. The paper also contains useful practical discussion of numerical scaling and of why RSS-based methods degrade under anchor error. The strongest asset is the explicit worst-case formulation and the attempt to use RSS consistency in rounding. However, the derivation contains objective substitutions and sign/dimension inconsistencies that make the solved problem not exactly the announced estimator, and no code or data is provided. The significance is therefore conditional on correcting these technical issues.

major comments (4)

[3.2, Eqs. (11)-(13)] Eq. (11) defines the robust estimator as min_x max_{|δ_i|≤ζ} ∑_{i=1}^M ( log10( (||x−ẑ_i||−δ_i)^2 / β_i^2 ) )^2. Eq. (13) is min_x max_i | log10( (||x−ẑ_i||−δ_i)^2 / β_i^2 ) |, which replaces the sum over anchors with a maximum over anchors and drops the square; this is a Chebyshev (l∞) fit, not the ML estimator announced in Eq. (8). The text says only that the l∞-norm is used 'to facilitate the design of a convex estimator,' but no argument is given that the two objectives share a minimizer or have comparable risk. Since every subsequent constraint and all simulation results inherit Eq. (13), the abstract's 'maximum likelihood estimator' claim is not supported. Either supply a proof or justification, or re-label the estimator as a robust Chebyshev/minimax estimator throughout.
[3.2, Eqs. (22) and (25)] There is a sign error relative to Eq. (18). Eq. (18) gives ||x−ẑ_i||^2 = tr(X) − 2x^T ẑ_i + ẑ_i^T ẑ_i. The constraints in Eqs. (22) and (25) use tr(X) + 2x^T ẑ_i + ẑ_i^T ẑ_i ± 2ζ l_i + ζ^2. As printed, these constraints bound quantities involving ||x + ẑ_i||^2, not the intended squared distance to the inaccurate anchor. If the simulations used the printed signs, the reported RMSE is for a different SDP. In addition, Eq. (25) states X∈R^{M×M} while Eqs. (19) and (23) require X = xx^T ∈ S^2; the Schur complement (X x; x^T 1) is only well-formed for X∈S^2. These corrections are essential before the numerical results can be interpreted.
[3.2, rank-one relaxation] The paper drops X = xx^T and L = ll^T with no analysis of the relaxation gap. The rounding algorithms depend on X* − x*x*^T being a meaningful covariance/error ellipsoid, and the 'ro' baseline is claimed to be the relaxed solution of Eq. (16); neither is justified without evidence on tightness. Please report the ranks of X* and L* and the objective gap, or provide a theoretical tightness argument. If the relaxation is loose, explain how the rounding algorithms compensate, since otherwise the reported gains from rounding are not explained.
[4, validation] All numerical results are generated from the same log-normal model (Eq. (1)) used in the derivation, so the experiments test self-consistency but not robustness to model mismatch. The Introduction announces real experiment results in 'Section V,' but the manuscript contains no real experiment. Either include real measurements or delete that claim and temper statements about practical deployment. This matters because the paper's title and conclusions emphasize robustness under anchor uncertainty.

minor comments (6)

[3.3, Alg. 3] The while condition 'whileds =||xo− x∗||)≤ 3σd do' is malformed; it should state a proper loop condition involving ds and the distance to x*.
[Notation throughout] The symbol L is overloaded: Eq. (20) uses L = ll^T, while Alg. 4 uses L for the RSS vector; β_i^2 and k also appear with inconsistent subscripts in Eqs. (28)-(30).
[Figs. 4 and 6] The captions' abbreviated label mapping is inconsistent (e.g., 'M,S,D,O,P,R,G,N' includes 'ml' twice and does not match the legend order).
[3.3, Alg. 4] The notation p←(M N) for M-permutations of N and the complexity of enumerating all combinations is not discussed; the choice of N and tt is left unspecified.
[1, organization] The Introduction states that Section V will present numerical and real experiment results, but the experiments are in Section IV and no real results are reported; renumber and correct.
[4, Eq. (28)] Equation (28) contains stray text ('...,M' and missing parentheses) that makes the SOCP-RSS constraints ambiguous.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the derivation is a sequence of explicit relaxations and approximations, and the in-sample simulations plus measurement-based rounding are standard estimation practice rather than fitted inputs renamed as predictions.

full rationale

The paper's central derivation chain is transparent: Eq. (8) poses a min-max robust ML problem, Eq. (9) uses a Taylor expansion, Eqs. (12)-(13) replace the sum-of-squared residual objective with a max-of-absolute-residual objective, and Eqs. (15)-(25) relax the nonconvex problem into an SDP. Each of these is an explicit modeling or relaxation step, not a definition of the output in terms of the output. Even where the algebra is questionable (e.g., the signs in Eq. (25) appear inconsistent with Eq. (18)), that is a correctness or verifiability concern, not circularity, because the SDP is not constructed from the quantity it is supposed to predict. The rounding algorithm Alg. 4 selects a candidate source-anchor combination by minimizing the squared RSS residual against the measured RSS vector L; this is the standard role of measurements in estimation, not a hidden parameter fit. The simulations are generated from the same log-normal shadowing model used in the derivation, which limits external validation, but it does not make the RMSE comparison circular: the compared methods are also model-based, and the proposed estimator could in principle perform poorly under the stated model. The author self-citations ([18], [19], [42], [43]) appear only as related-work pointers on distance-matrix localization and sensor deployment, and they are not load-bearing premises in the derivation. No unique-theorem or ansatz-by-citation pattern is present. Overall, no step in the paper reduces by construction to its own inputs, so the circularity score is 0.

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

The central method has no physical invented entities. It relies on the log-normal RSS model, bounded anchor error, a first-order Taylor approximation, and the unproved tightness of the SDP relaxation. Several algorithm tuning constants such as tt, N, and grid step sizes are hand-chosen and are not specified with values in the manuscript.

free parameters (3)

Number of randomization trials tt
Algorithm 1 and Algorithm 4 draw tt candidate source positions. No value or sensitivity analysis is given, and the accuracy of the rounding step depends on this hand-chosen constant.
Number of anchor candidates N
Algorithm 4 samples N possible positions per anchor within the error disk. The complexity and accuracy of the final rounding step depend on this hand-chosen integer, which is never specified.
Grid search step sizes ds and Delta-ds
Algorithm 3 starts with ds=0.0001 and increment Delta-ds=0.001 times sigma_d. These values are hand-chosen and not justified, and they affect the search resolution.

assumptions (6)

domain assumption Log-normal shadowing path loss model in Eq (1), with known transmit power, reference path loss, and path loss exponent gamma.
All derivations and simulations assume this RSS model. The paper does not validate it against real measurements.
domain assumption Anchor position errors are bounded as ||Delta_i|| <= zeta, with no specified distribution, in Eq (3).
The min-max formulation in Eq (8) is built entirely on this bounded-error model.
domain assumption First-order Taylor expansion of ||x - z_i|| in Eq (9), neglecting o(||Delta_i||).
The transformation from Eq (8) to Eq (11) depends on this local linearization. The paper provides no error bound for the neglected terms.
domain assumption The assumption ||x - hat-z_i|| - zeta > 0 for all i, stated in Section 3.2.
The robust constraints in Eq (15) require that the source remains outside a zeta-neighborhood of every anchor. The SDP relaxation does not explicitly enforce this condition.
ad hoc to paper The rank-one constraints X = xx^T and L = ll^T in Eq (22) can be dropped without destroying solution quality.
The relaxation to Eq (25) relies on the heuristic tightness of the SDP relaxation. No theoretical tightness proof is given; only simulation results are offered.
ad hoc to paper The covariance matrix X* - x*x*^T used in randomization approximates the error distribution of the source estimate.
Algorithms 1 and 4 sample candidate source positions from this Gaussian form. The paper does not justify this distributional approximation.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Received Signal Strength Based Wireless Source Localization with Inaccurate Anchor Position." pith.science (2026). https://pith.science/paper/JKX55R44

@misc{pith2026190803202,
  author       = {Pith},
  title        = {Pith review of: Received Signal Strength Based Wireless Source Localization with Inaccurate Anchor Position},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JKX55R44}},
  note         = {Machine review of arXiv:1908.03202}
}

read the original abstract

Received signal strength (RSS)-based wireless localization is easy to implement at low cost. In practice,exact positions of anchors may not be available. This paper focuses on determining the location of a source in the presence of inaccurate positions of anchors based on RSS directly. We first use Taylor expansion and a min-max approach to get a maximum likelihood estimator of the coordinates of the source. Then we propose a relaxed semi-definite programming model to circumvent the non-convexity. We also propose a rounding algorithm considering both inaccurate source location and inaccurate anchor locations.Simulation results together with analysis are presented to validate the proposed method.

Figures

Figures reproduced from arXiv: 1908.03202 by the authors.

**Figure 1.** Illustration of rounding algorithm Alg.4. The circle denotes the real source location. Plus sign denotes the solution get from Eq. [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗

**Figure 2.** Two types of anchor placement. In (a), anchors are randomly placed. In (b) anchors are placed by design. It is obvious that in (b) the source [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗

**Figure 3.** Numerical simulation results of different methods with various value of anchor location error [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Figures from the paper (5 more)

**Figure 4.** Figure 4: The boxplot of estimation errors. (a) is under ’bad’ anchor placement as shown in Fig.2a. (b) is under ’good’ anchor placement as show [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]

**Figure 5.** Figure 5: Numerical simulation results of different methods with various value of RSS measurement noise [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

**Figure 6.** Figure 6: The boxplot of estimation errors corresponding with different level of noise. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]

**Figure 7.** Figure 7: Numerical simulation results of different methods with various number of anchors [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

**Figure 8.** Figure 8: The boxplot of estimation errors corresponding with different number of anchors. (a) is under ’bad’ anchor placement as shown in Fig.2a. (b) [PITH_FULL_IMAGE:figures/full_fig_p010_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

45 extracted references · 43 canonical work pages

[22]

Robust differential received signal strength- based localization,

Y. Hu and G. Leus, “Robust differential received signal strength- based localization,” IEEE T ransactions on Signal Processing, vol. 65, no. 12, pp. 3261–3276, 2017

work page 2017
[29]

Rss-based localization in wireless sensor networks using convex relaxation: Noncooperative and cooperative schemes,

S. Tomic, M. Beko, and R. Dinis, “Rss-based localization in wireless sensor networks using convex relaxation: Noncooperative and cooperative schemes,” IEEE T ransactions on Vehicular T echnology , vol. 64, no. 5, pp. 2037–2050, May 2015

work page 2015
[1]

Smartphone- based indoor positioning using ble ibeacon and reliable lightweight ﬁngerprint map,

T.-M. T. Dinh, N.-S. Duong, and K. Sandrasegaran, “Smartphone- based indoor positioning using ble ibeacon and reliable lightweight ﬁngerprint map,” IEEE Sensors Journal , vol. 20, no. 17, pp. 10 283–10 294, 2020

work page 2020
[2]

Cedar: A cost-effective crowdsensing system for detecting and localizing drones,

G. Yang, X. Shi, L. Feng, S. He, Z. Shi, and J. Chen, “Cedar: A cost-effective crowdsensing system for detecting and localizing drones,” IEEE T ransactions on Mobile Computing, vol. 19, no. 9, pp. 2028–2043, Jan. 2020

work page 2020
[3]

Target detection and localization meth- ods using compartmental model for internet of things,

S. Kumar and S. K. Das, “Target detection and localization meth- ods using compartmental model for internet of things,” IEEE T ransactions on Mobile Computing, vol. 19, no. 9, pp. 2234–2249, Jan. 2020

work page 2020
[4]

Adversarial learning-enabled automatic wiﬁ indoor radio map construction and adaptation with mobile robot,

H. Zou, C.-L. Chen, M. Li, J. Yang, Y. Zhou, L. Xie, and C. J. Spanos, “Adversarial learning-enabled automatic wiﬁ indoor radio map construction and adaptation with mobile robot,” IEEE Internet of Things Journal, vol. 7, no. 8, pp. 6946–6954, Aug. 2020

work page 2020
[5]

Smartloc: Smart wireless indoor localization empowered by machine learning,

L. Li, X. Guo, and N. Ansari, “Smartloc: Smart wireless indoor localization empowered by machine learning,” IEEE T ransactions on Industrial Electronics , vol. 67, no. 8, pp. 6883–6893, Aug. 2020

work page 2020
[6]

Spear: Source position estimation for anchor po- sition uncertainty reduction,

M. Angjelichinoski, D. Denkovski, V . Atanasovski, and L. Gavrilovska, “Spear: Source position estimation for anchor po- sition uncertainty reduction,” IEEE Communications Letters, vol. 18, no. 4, pp. 560–563, Apr. 2014

work page 2014

Show all 45 references

[7]

Cramer-rao lower bounds of rss-based localization with anchor position uncertainty,

M. Angjelichinoski and D. Denkovski, “Cramer-rao lower bounds of rss-based localization with anchor position uncertainty,” IEEE T ransactions on Information Theory , vol. 61, no. 5, pp. 2807–2834, May 2015

2015
[8]

Rss distribution-based passive localization and its application in sensor networks,

C. Liu, D. Fang, Z. Yang, H. Jiang, X. Chen, W. Wang, T. Xing, and L. Cai, “Rss distribution-based passive localization and its application in sensor networks,” IEEE T ransactions on Wireless Communications, vol. 15, no. 4, pp. 2883–2895, April 2016

2016
[9]

Distributed algorithm for target localiza- tion in wireless sensor networks using rss and aoa measurements,

S. Tomic and M. Beko, “Distributed algorithm for target localiza- tion in wireless sensor networks using rss and aoa measurements,” Pervasive and Mobile Computing , vol. 37, pp. 63–77, 2017

2017
[10]

Source localization in wireless sensor networks from signal time-of-arrival measurements,

E. Xu, Z. Ding, and S. Dasgupta, “Source localization in wireless sensor networks from signal time-of-arrival measurements,” IEEE T ransactions on Signal Processing, vol. 59, no. 6, pp. 2887–2897, 2011

2011
[11]

Localization via ultra-wideband radios,

S. Gezici, Z. Tian, G. B. Giannakis, H. Kobayashi, A. F. Molisch, H. V . Poor, and Z. Sahinoglu, “Localization via ultra-wideband radios,” IEEE Signal Processing Magazine , vol. 22, no. 4, pp. 70–84, 2005

2005
[12]

Uav-based passive geolocation based on channel estimation,

Q. Liang and S. W. Samn, “Uav-based passive geolocation based on channel estimation,” in 2010 IEEE Globecom Workshops , Dec 2010, pp. 1821–1825

2010
[13]

Tdoa approach for target localization based on improved genetic algorithm,

M. Rosi ´c, M. Simie, and P . Luki ´c, “Tdoa approach for target localization based on improved genetic algorithm,” in 2016 24th T elecommunications Forum (TELFOR), Nov 2016, pp. 1–4

2016
[14]

Source localization using tdoa and fdoa mea- surements based on modiﬁed cuckoo search algorithm,

M. Abd El Aziz, “Source localization using tdoa and fdoa mea- surements based on modiﬁed cuckoo search algorithm,” Wireless Networks, vol. 23, no. 2, pp. 487–495, 2017

2017
[15]

Second-order cone relaxation for tdoa-based localization under mixed los/nlos condi- tions,

W. Wang, G. Wang, F. Zhang, and Y. Li, “Second-order cone relaxation for tdoa-based localization under mixed los/nlos condi- tions,” IEEE Signal Processing Letters, vol. 23, no. 12, pp. 1872–1876, 2016

2016
[16]

Doa acoustic source localization in mobile robot sensor networks,

R. Levorato and E. Pagello, “Doa acoustic source localization in mobile robot sensor networks,” in 2015 IEEE International Confer- ence on Autonomous Robot Systems and Competitions , April 2015, pp. 71–76

2015
[17]

Sectorized antenna-based doa estimation and localization: Advanced algorithms and measurements,

J. Werner and J. Wang, “Sectorized antenna-based doa estimation and localization: Advanced algorithms and measurements,” IEEE Journal on Selected Areas in Communications , vol. 33, no. 11, pp. 2272–2286, 2015

2015
[18]

Local patches alignment embed- ding based localization for wireless sensor networks,

Y. Liu, J. Chen, and Y.-j. Zhan, “Local patches alignment embed- ding based localization for wireless sensor networks,” Wireless Personal Communications, vol. 70, no. 1, pp. 373–389, 2013

2013
[19]

Wireless sensor network localization based on semi-supervised local subspace alignment,

J. Chen and Y. Liu, “Wireless sensor network localization based on semi-supervised local subspace alignment,” in Applied Mechanics and Materials, vol. 385, 2013, pp. 1618–1621

2013
[20]

Second-order cone programming relaxation of sensor network localization,

P . Tseng, “Second-order cone programming relaxation of sensor network localization,” SIAM Journal On Optimization, vol. 18, no. 1, pp. 156–185, 2007

2007
[21]

Robust relative location estimation in wireless sensor networks with inexact position problems,

B.-S. C. Wei-Yu Chiu, “Robust relative location estimation in wireless sensor networks with inexact position problems,” IEEE T ransactions on Mobile Computing, vol. 11, no. 6, pp. 935–946, 2012

2012
[23]

Source localization using tdoa and fdoa measurements in the presence of receiver location errors: Analysis and solution,

K. C. Ho and X. N. Lu, “Source localization using tdoa and fdoa measurements in the presence of receiver location errors: Analysis and solution,” IEEE T ransactions on Signal Processing, vol. 55, no. 2, pp. 684–696, 2007

2007
[24]

Effect in range difference position estimation due to stations’ position errors,

D. E. Manolakis and M. E. Cox, “Effect in range difference position estimation due to stations’ position errors,” IEEE T ransactions on Aerospace and Electronic Systems , vol. 34, no. 1, pp. 329–334, 1998

1998
[25]

Second Order Cone Programming for Sensor Network Localization with Anchor Position Uncertainty,

G. Naddafzadeh-Shirazi, M. B. Shenouda, and L. Lampe, “Second Order Cone Programming for Sensor Network Localization with Anchor Position Uncertainty,” IEEE T ransactions On Wireless Com- munications, vol. 13, no. 2, pp. 749–763, FEB 2014

2014
[26]

Efﬁcient convex relaxation methods for robust target localization by a sensor network using time differences of arrivals,

K. Yang, G. Wang, and Z.-Q. Luo, “Efﬁcient convex relaxation methods for robust target localization by a sensor network using time differences of arrivals,” IEEE T ransactions on Signal Processing, vol. 57, no. 7, pp. 2775–2784, 2009

2009
[27]

Range-Based Rigid Body Localization With a Calibration Emitter for Mitigating Anchor Position Un- certainties,

B. Hao, K. C. Ho, and Z. Li, “Range-Based Rigid Body Localization With a Calibration Emitter for Mitigating Anchor Position Un- certainties,” IEEE T ransactions On Wireless Communication , vol. 18, no. 12, pp. 5734–5748, DEC 2019

2019
[28]

On the performance limits of cooperative localization in wireless sensor networks with strong sensor position uncertainty,

Y. Xiong, N. Wu, and H. Wang, “On the performance limits of cooperative localization in wireless sensor networks with strong sensor position uncertainty,” IEEE Communications Letters , vol. 21, no. 7, pp. 1613–1616, Jul. 2017

2017
[30]

Geometric interpretation of theoretical bounds for rss-based source localization with uncertain anchor positions,

D. Denkovski, M. Angjelichinoski, V . Atanasovski, and L. Gavrilovska, “Geometric interpretation of theoretical bounds for rss-based source localization with uncertain anchor positions,” Digital Signal Processing , vol. 68, pp. 167–181, Sep. 2017. 12

2017
[31]

Goldsmith, Wireless Communications

A. Goldsmith, Wireless Communications . Cambridge University Press, 2005

2005
[32]

Rank-one semideﬁnite programming solutions for mobile source localization in sensor networks,

X. Wu, H. Qi, and N. Xiong, “Rank-one semideﬁnite programming solutions for mobile source localization in sensor networks,” IEEE T ransactions on Network Science and Engineering , vol. 8, no. 1, pp. 638–650, Jan. 2021

2021
[33]

CVX: Matlab software for disciplined convex programming, version 2.1,

M. Grant and S. Boyd, “CVX: Matlab software for disciplined convex programming, version 2.1,” http://cvxr.com/cvx, Mar. 2014

2014
[34]

Solving semideﬁnite- quadratic-linear programs using sdpt3,

R. H. T ¨ut ¨unc ¨u, K. C. Toh, and M. J. Todd, “Solving semideﬁnite- quadratic-linear programs using sdpt3,” Mathematical Program- ming, vol. 95, no. 2, pp. 189–217, Feb 2003

2003
[35]

Relative location estimation in wireless sensor networks,

N. Patwari and A. Hero, “Relative location estimation in wireless sensor networks,” IEEE T ransactions On Signal Processing , vol. 51, no. 8, pp. 2137–2148, AUG 2003

2003
[36]

Received Signal Strength-Based Wireless Localization via Semideﬁnite Programming: Noncooper- ative and Cooperative Schemes,

R. W. Ouyang and A. K.-S. Wong, “Received Signal Strength-Based Wireless Localization via Semideﬁnite Programming: Noncooper- ative and Cooperative Schemes,” IEEE T ransactions On Vehicular T echnology, vol. 59, no. 3, pp. 1307–1318, MAR 2010

2010
[37]

The effect of multipath and nlos on toa ranging error and energy based on uwb,

X. Li and J. He, “The effect of multipath and nlos on toa ranging error and energy based on uwb,” in 2016 IEEE International Con- ference on Consumer Electronics-T aiwan (ICCE-TW) , May 2016, pp. 1–2

2016
[38]

https://www.ibm.com/analytics/cplex-optimizer

“https://www.ibm.com/analytics/cplex-optimizer.”
[39]

A novel algorithm for dis- tributed localization in wireless sensor networks,

M. Naraghi-Pour and G. C. Rojas, “A novel algorithm for dis- tributed localization in wireless sensor networks,” ACM T ransac- tion on Sensor Networks , vol. 11, no. 9, pp. 1–25, Sep. 2014

2014
[40]

A distributed support vector ma- chine learning over wireless sensor networks,

W. Kim and M. S. Stankovic, “A distributed support vector ma- chine learning over wireless sensor networks,” IEEE T ransactions on Cybernetics, vol. 45, no. 11, pp. 2599–2611, Nov. 2015

2015
[41]

Distributed event localization via alternat- ing direction method of multipliers,

C. Zhang and Y. Wang, “Distributed event localization via alternat- ing direction method of multipliers,” IEEE T ransactions on Mobile Computing, vol. 17, no. 2, pp. 348–361, Feb. 2018

2018
[42]

Optimization of sensor deployment for localization accuracy improvement,

R. Wang, B. Shen, and Y. Liu, “Optimization of sensor deployment for localization accuracy improvement,” in 2016 IEEE International Conference on Consumer Electronics-China (ICCE-China) , 2017

2016
[43]

To improve localization accuracy: A two-objective optimization method,

R. Wang, C. Ye, B. Shen, and Y. Liu, “To improve localization accuracy: A two-objective optimization method,” in 2018 13th IEEE Conference on Industrial Electronics and Applications (ICIEA) , 2018, pp. 2510–2513

2018
[44]

A general framework for sensor placement in source localization,

B. Spinelli, L. E. Celis, and P . Thiran, “A general framework for sensor placement in source localization,” IEEE T ransactions on Network Science and Engineering, vol. 6, no. 2, pp. 86–102, Apr. 2019

2019
[45]

Informative path planning for location ﬁngerprint collection,

Y. Wei, C. Frincu, and R. Zheng, “Informative path planning for location ﬁngerprint collection,” IEEE T ransactions on Network Science and Engineering , vol. 7, no. 3, pp. 1633–1644, Jul. 2020

2020

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [22]

Robust differential received signal strength- based localization,

Y. Hu and G. Leus, “Robust differential received signal strength- based localization,” IEEE T ransactions on Signal Processing, vol. 65, no. 12, pp. 3261–3276, 2017

work page 2017

[2] [29]

Rss-based localization in wireless sensor networks using convex relaxation: Noncooperative and cooperative schemes,

S. Tomic, M. Beko, and R. Dinis, “Rss-based localization in wireless sensor networks using convex relaxation: Noncooperative and cooperative schemes,” IEEE T ransactions on Vehicular T echnology , vol. 64, no. 5, pp. 2037–2050, May 2015

work page 2015

[3] [1]

Smartphone- based indoor positioning using ble ibeacon and reliable lightweight ﬁngerprint map,

T.-M. T. Dinh, N.-S. Duong, and K. Sandrasegaran, “Smartphone- based indoor positioning using ble ibeacon and reliable lightweight ﬁngerprint map,” IEEE Sensors Journal , vol. 20, no. 17, pp. 10 283–10 294, 2020

work page 2020

[4] [2]

Cedar: A cost-effective crowdsensing system for detecting and localizing drones,

G. Yang, X. Shi, L. Feng, S. He, Z. Shi, and J. Chen, “Cedar: A cost-effective crowdsensing system for detecting and localizing drones,” IEEE T ransactions on Mobile Computing, vol. 19, no. 9, pp. 2028–2043, Jan. 2020

work page 2020

[5] [3]

Target detection and localization meth- ods using compartmental model for internet of things,

S. Kumar and S. K. Das, “Target detection and localization meth- ods using compartmental model for internet of things,” IEEE T ransactions on Mobile Computing, vol. 19, no. 9, pp. 2234–2249, Jan. 2020

work page 2020

[6] [4]

Adversarial learning-enabled automatic wiﬁ indoor radio map construction and adaptation with mobile robot,

H. Zou, C.-L. Chen, M. Li, J. Yang, Y. Zhou, L. Xie, and C. J. Spanos, “Adversarial learning-enabled automatic wiﬁ indoor radio map construction and adaptation with mobile robot,” IEEE Internet of Things Journal, vol. 7, no. 8, pp. 6946–6954, Aug. 2020

work page 2020

[7] [5]

Smartloc: Smart wireless indoor localization empowered by machine learning,

L. Li, X. Guo, and N. Ansari, “Smartloc: Smart wireless indoor localization empowered by machine learning,” IEEE T ransactions on Industrial Electronics , vol. 67, no. 8, pp. 6883–6893, Aug. 2020

work page 2020

[8] [6]

Spear: Source position estimation for anchor po- sition uncertainty reduction,

M. Angjelichinoski, D. Denkovski, V . Atanasovski, and L. Gavrilovska, “Spear: Source position estimation for anchor po- sition uncertainty reduction,” IEEE Communications Letters, vol. 18, no. 4, pp. 560–563, Apr. 2014

work page 2014