Pith. sign in

REVIEW 4 major objections 4 minor 29 references

Hybrid CNN-LSTM based Indoor Pedestrian Localization with CSI Fingerprint Maps

T0 review · 4 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read This paper claims spatial Wi-Fi phase maps, a hybrid CNN-LSTM, and a particle filter localize pedestrians to 0.36 m in dynamic settings and 0.17 m in static ones.

desk verdict The map-based CSI fingerprint idea is new, but the headline accuracy numbers are not reproducible from the paper's own experiments. read the letter →

arxiv 2412.13601 v2 pith:ZI5KVVDJ submitted 2024-12-18 cs.CV cs.AI

Muhammad Emad-ud-din This is my paper · ORCID
classification cs.CVcs.AI
keywords indoorlocalizationWi-Fifingerprintingchannelstateinformation(CSI)CSIphasesanitizationhybridCNN-LSTMparticlefilterpedestriantrajectoryOFDMsubcarriers
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 is trying to establish that ordinary Wi-Fi channel measurements, which usually fluctuate too much for reliable positioning, can be rearranged into a spatial image and learned from as a sequence. The image, called a CSI Fingerprint Map, places 270 sanitized phase values from a single observation at the pixel corresponding to its physical location, and the hybrid convolutional-and-recurrent network learns both spatial fingerprint structure and short-term temporal dependencies between consecutive map slices. A particle filter then enforces a human walk model by generating many trajectory hypotheses and keeping the one a real pedestrian could have walked. On a 140 square meter corridor with one access point and one receiver, the reported average errors are 0.36 meters in a moderately dynamic environment and 0.17 meters in a static one, well below the 1 to 4 meter range typical of earlier Wi-Fi fingerprinting. If this holds, sub-meter indoor pedestrian localization becomes possible with existing Wi-Fi infrastructure and without cameras, IMUs, or magnetometers.

What carries the argument

The central object is the CSI Fingerprint Map: a 2D spatial grid whose pixel at each reference location is a vector of 270 sanitized CSI phase values from 30 OFDM subcarriers across 9 MIMO antenna pairs. Successive maps are indexed so that all pixels in map i are treated as one time step before map i+1, which is what lets the recurrent network see the trajectory as a sequence. The argument is carried by three supporting mechanisms: a phase-sanitization and denoising step that makes fingerprints temporally stable, an Information Adaptive Sliding Window that feeds the CNN spatial patches of increasing neighborhood size, and a particle-filter hypothesis-selection module that enforces a human gait model and discards location sequences that violate realistic walking speed, stride, or heading constraints.

What would settle it

One decisive check is to train the same model on the same CSI Fingerprint Maps but with the index order of the maps randomly shuffled; if the localization error stays the same, the LSTM is not using temporal structure and the central mechanism is a spatial prior. A corroborating check is to rerun the evaluation on longer test walks measured against independent motion-capture ground truth instead of sequential walk labels.

Watch

Extended reading notes

Core claim

The central claim is that treating CSI phase observations as pixels in a time-ordered 2D map turns Wi-Fi fingerprinting from a per-location classification problem into a spatial-sequence learning problem. Each map pixel carries 270 sanitized phase values from nine MIMO antenna pairs, and a sliding window of pixels (1x1, 2x2, or 3x3) forms proposals whose size adapts to how much surrounding spatial information is available. A CNN extracts spatial features from these map patches, an LSTM models short-term dependencies between consecutive maps, and the output is a 37-entry confidence vector over grid locations. A multi-hypothesis particle filter, driven by a foot-position human walk model, then weights candidate trajectories and selects the most likely one, reducing mean distance error by about 38 percent on average and down to as low as 0.07 meters for some parameter combinations. The author frames the result as a proof of concept that sparse Wi-Fi observations, one access point, and moderate noise are enough for fine-grained pedestrian localization and heading estimation.

Load-bearing premise

The method depends on treating map slices built from observations taken at 36 locations at different times as consecutive moments of one walk; if that artificial time order is not physically meaningful, the temporal modeling that the whole hybrid architecture depends on has nothing real to learn.

Editorial extensions

If this is right

  • The reported error bars put Wi-Fi-only localization in the sub-meter range previously associated with camera-based or sensor-fused systems.
  • Because the pipeline needs only one access point and one receiver, deployment could piggyback on existing Wi-Fi infrastructure rather than requiring a dense grid of anchors.
  • The particle-filter step not only removes gross misclassifications but also produces pedestrian heading estimates from Wi-Fi alone, removing the need for IMU or magnetometer input.
  • The post-filtering accuracy improvement implies that most residual CNN-LSTM errors are implausible walking jumps, which is why enforcing a human walk model can cut mean distance error by roughly 38 percent on average.

Reading between the lines

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

  • An extension the paper leaves implicit is that the same spatial map representation could be trained with RSSI values instead of CSI phase; holding the architecture fixed would isolate how much of the gain comes from the spatial arrangement and how much from the richer phase feature.
  • A decisive ablation the paper does not report is shuffling the temporal order of the map slices during training; if localization accuracy does not degrade, the LSTM is effectively learning a spatial prior while the claimed temporal dependency is an artifact of the collection design.
  • Because the data come from a single corridor and a single walk pattern, whether the learned spatial-temporal weights transfer to other floor plans, device placements, and walking styles is an untested consequence of treating fingerprints as maps.
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

4 major / 4 minor

Summary. The paper proposes a Wi-Fi CSI fingerprinting localization system built around a "CSI Fingerprint Map" (a 2D+channel image formed by re-indexing CSI phase observations by map location), a hybrid CNN-LSTM classifier, and a particle-filter hypothesis-selection module. The authors report comparisons against DeepFi, ConFi, and a temporal LSTM baseline, and claim average RMSE of 0.36 m in a moderately dynamic environment and 0.17 m in a static environment.

Significance. If the claimed sub-meter accuracy were supported by the experiments, this would be a meaningful step toward fine-grained Wi-Fi localization with sparse infrastructure. The paper contributes a two-month dataset, a spatial-map representation, and an evaluation against established baselines. However, the authors do not provide code or data, and as detailed below the headline results are not reproducible from the reported experiments, which substantially limits the significance of the contribution.

major comments (4)
  1. [Abstract and §5.4 / Table 3] The abstract's RMSE values (0.36 m moderately dynamic, 0.17 m static) are not reproducible from the experimental record. Table 3 reports mean distance error before and after PF-HSM for 12 parameter combinations, with after-PF values from 0.07 m to 0.75 m; no row or aggregation yields 0.36 m, 0.17 m is not among the reported values, and no static test protocol is described anywhere in §5. The paper also alternates between 'mean error' (Tables 2 and 3) and 'RMSE' (abstract), so the headline metric is not the measured quantity. The stated 38% improvement would turn the Table 2 CNN-LSTM value of 0.79 m into approximately 0.49 m, not 0.36 m. The central claim of the paper is therefore not derivable from the paper's own experimental results.
  2. [§4.2.1, Eq. (3)] The 'time-ordered' property of the CSI Fingerprint Maps is a construction, not a physical temporal sequence. A single device cannot collect simultaneous observations at 36 locations; the text itself acknowledges that 'one must gather observations at each location at different times.' The re-indexing that creates M_i therefore does not provide a real temporal step between M_i and M_{i+1}, and the LSTM's temporal modeling lacks a physical basis. The paper must either demonstrate that this re-indexing preserves meaningful temporal structure or abandon the LSTM temporal claim.
  3. [Table 3] Configuration selection is performed on the reported test results. The best after-PF error (0.07 m) is achieved for grid size 1 m, 3×3 sliding window, 1 m/s, while other configurations give errors up to 0.75 m; no a priori selection rule or cross-validation procedure is stated. Reporting the minimum of a grid search over grid size, window size, and walking speed as the method's accuracy is circular unless the evaluation protocol separates model selection from testing.
  4. [§5.4 and §6] The paper claims in the conclusion that the method 'delivers estimated pedestrian heading without the use of any additional sensors,' but the heading accuracy evaluation in §5.4 is performed only with a simulated human walk motion model, not with real walking data. No heading ground truth from real pedestrian trajectories is presented. This overclaim is not supported by the reported experiments.
minor comments (4)
  1. [§4.2.4] The particle-filter measurement model equations contain apparent errors: d0 uses (rx - lx) for both the cos and sin terms, omitting (ry - ly); the normalization factor 1/(√(2π)(1 - c)) in p_L and p_R is not derived. These issues hinder reproducibility.
  2. [Throughout] The manuscript has numerous presentation issues: Table 2 is referred to as Table 1 in the text ('The mean errors of all compared methods is presented in Table 1'), Figure 3 is mislabeled as 'Figure 3 Trajectory to grid-cell mapping,' equations are unnumbered, and the formatting of Table 3 is garbled.
  3. [§5.2 and §5.3] The methodology does not state how many test trajectories were collected, their lengths, or how the reported mean errors are aggregated over trajectories and parameter combinations. Without this information, the comparison in Table 2 cannot be independently reproduced.
  4. [§3.3] The denoising rule (removing observations with CSI phase values beyond 2σ for any subcarrier) is described only qualitatively; the fraction of removed observations, their spatial/temporal distribution, and the effect on the resulting fingerprints are not reported, so the impact of denoising on the final accuracy cannot be assessed.

Circularity Check

2 steps flagged · score 6.0 of 10

The abstract's headline accuracies are a post hoc selection from a test-set parameter sweep, and the temporal ordering that motivates the LSTM is imposed by the map indexing rather than measured.

  1. fitted input called prediction [Abstract; Section 5.4, Table 3]
    "The experimental results show marked improvement with an average RMSE of 0.36 m in a moderately dynamic and 0.17 m in a static environment. ... PF-HSM reduces the Mean Distance Error to as low as 0.07 m for a specific combination of parameters."

    The abstract's single accuracy values are not derivable from the reported experiments. Section 5.4 presents a 12-row sweep over grid-size, sliding-window size, and walking speed, with after-PF mean distance errors from 0.07 to 0.75 m, and gives no selection rule or held-out test protocol. The tables report 'Mean Distance Error,' while the abstract reports 'RMSE,' and no static-environment test is described. The only way to obtain the headline numbers is to select or aggregate favorable cells of the test-set sweep; reporting a post hoc best/aggregated cell as the method's accuracy is equivalent to fitting the method to the test set and then renaming the fitted value as a prediction.

  2. self definitional [Section 4.2.1 (CSI Fingerprint Map)]
    "all pixel observations FP_{x,y}^i in M_i are taken at the previous time step when compared to FP_{x,y}^{i+1} in M_{i+1}. The reason behind why this order is maintained between subsequent observations , M_i, is that this enables us to feed these time-ordered maps to a sequential neural network such as LSTM."

    The temporal ordering that justifies the LSTM is manufactured by indexing, not observed. M_i is defined as the collection of fingerprints with observation index i at each location, and the dataset was collected with one device at 36 locations, so those observations cannot be simultaneous. The claim that every pixel of M_i precedes the corresponding pixel of M_{i+1} is therefore true by definition of the index i, not by measurement. The 'time-ordered maps' property is thus built into the input representation, and the temporal relationship the hybrid CNN-LSTM is said to exploit is an artifact of that construction rather than a temporal dependency derived from pedestrian motion.

full rationale

The paper is not circular in the narrow sense that the CNN-LSTM classifier is trained on real CSI data and compared against DeepFi and ConFi on a common test stream. Those comparisons have independent content. However, two load-bearing parts of the paper's central narrative partially reduce to construction. First, the abstract's headline accuracies (0.36 m dynamic, 0.17 m static) do not appear in any table, no static test protocol is given, and the only nearby evidence is a grid search over three parameters with after-PF errors ranging from 0.07 to 0.75 m and no stated selection rule; reporting such a selected/aggregated value as the method's accuracy is a post hoc fit to the test set. Second, the time-ordered CSI Fingerprint Map property that enables the LSTM is imposed by re-indexing independent per-location observations, so the temporal relationship exploited by the model is defined into the representation. These are partial circularities in the reporting and input construction. The numerical comparison against prior methods is not itself circular, which keeps the score at 6 rather than higher.

Assumptions & free parameters 7 free parameters · 5 assumptions · 0 invented entities

The central claim rests on several hand-chosen parameters and assumptions: a Gaussian noise model, the temporal stability of sanitized phase, an artificial time-ordering of maps collected asynchronously, and a human walk model taken from prior work. The free parameters are many, and the headline accuracy appears to be selected from the best combination in a grid search. No new physical entities are introduced.

free parameters (7)
  • Training grid size = 1 m and 2 m tested; best results reported with 2 m
    Table 3 reports results for both 1 m and 2 m grid sizes; the best accuracy (0.07 m) uses 2 m. The grid size is chosen after evaluating on the test set.
  • Information Adaptive Sliding Window Size = 1x1, 2x2, 3x3 tested; 3x3 used in best results
    The sliding window size is a method parameter evaluated in Table 3; the paper reports the best results with 3x3 windows but does not specify which setting produced the abstract's headline numbers.
  • Pedestrian walking speed = 1 m/s and 3 m/s tested
    The walking speed assumption is varied in Table 3 and affects the particle filter and trajectory hypothesis generation; no rationale for choosing 3 m/s in the headline result is given.
  • LSTM L2 regularization factor for forget gate = 2X the default of 0.001
    Section 4.2.3 states that a 2X weight gave optimal results after testing 1X, 2X, and 4X. This is a tuned hyperparameter.
  • Number of CNN layers and LSTM hidden units = Not specified; chosen by tuning
    Section 4.2.3 states that increasing CNN layers reduced accuracy and increasing LSTM hidden units caused overfitting, but exact values used in the final model are not reported.
  • Denoising threshold = 2 sigma
    Section 3.3 removes CSI phase values beyond the 2 sigma bound, based on an assumed Gaussian noise distribution. This threshold is chosen by hand.
  • Particle filter parameters (n, f, h, etc.) = n >= 3, f not specified, h not specified
    Section 4.2.4 defines the parameters but does not report their values, so the filter behavior is not reproducible.
assumptions (5)
  • domain assumption Sanitized CSI phase values are temporally stable over 44 days
    Section 3.3 and Figure 3 show stability for one location over two months, but the paper generalizes this to all locations and times. No formal study of long-term CSI stability is provided, as the text admits.
  • domain assumption CSI phase noise follows a Gaussian distribution
    Section 3.3 states: 'We thus assume similar noise distribution and bounds for CSI phase values' based on RSSI literature. This assumption is used to justify the 2 sigma denoising rule.
  • ad hoc to paper Time-ordered CSI fingerprint maps are a meaningful temporal sequence
    Section 4.2.1 assumes that observations from different locations, collected at different physical times, can be indexed to form a consistent temporal sequence. A single device cannot sample all locations simultaneously, so this ordering is a construction without physical justification.
  • standard math The human walk model from [20] is applicable to this environment
    The particle filter relies on the bipedal gait model from prior work [20]. The paper does not validate the model on the actual test corridor, and it uses a simulated version [21] for validation.
  • domain assumption CSI features provide spatial diversity at the 1 m grid resolution
    The entire fingerprinting approach assumes distinct locations yield distinct CSI phase patterns. This is supported in the literature but not tested formally within this paper's environment.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Hybrid CNN-LSTM based Indoor Pedestrian Localization with CSI Fingerprint Maps." pith.science (2026). https://pith.science/paper/ZI5KVVDJ

@misc{pith2026241213601,
  author       = {Pith},
  title        = {Pith review of: Hybrid CNN-LSTM based Indoor Pedestrian Localization with CSI Fingerprint Maps},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZI5KVVDJ}},
  note         = {Machine review of arXiv:2412.13601}
}
read the original abstract

The paper presents a novel Wi-Fi fingerprinting system that uses Channel State Information (CSI) data for fine-grained pedestrian localization. The proposed system exploits the frequency diversity and spatial diversity of the features extracted from CSI data to generate a 2D+channel image termed as a CSI Fingerprint Map. We then use this CSI Fingerprint Map representation of CSI data to generate a pedestrian trajectory hypothesis using a hybrid architecture that combines a Convolutional Neural Network and a Long Short-Term Memory Recurrent Neural Network model. The proposed architecture exploits the temporal and spatial relationship information among the CSI data observations gathered at neighboring locations. A particle filter is then employed to separate out the most likely hypothesis matching a human walk model. The experimental performance of our method is compared to existing deep learning localization methods such ConFi, DeepFi and to a self-developed temporal-feature based LSTM based location classifier. The experimental results show marked improvement with an average RMSE of 0.36 m in a moderately dynamic and 0.17 m in a static environment. Our method is essentially a proof of concept that with (1) sparse availability of observations, (2) limited infrastructure requirements, (3) moderate level of short-term and long-term noise in the training and testing environment, reliable fine-grained Wi-Fi based pedestrian localization is a potential option.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

29 extracted references · 27 canonical work pages

  1. [1]

    Introduction People localization is an essential component of many applications like indoor security camera systems, activity classification, and elderly people who need to be monitored. Apart from Wi-Fi based localization, alternate options like camera-based people tr acking generally fail to deliver because of problems with maintaining line-of-sight wit...

    work page

  2. [2]

    Literature Review For coarse-grained localization, RSSI based fingerprinting has not been able to show convincing accuracy , primarily because of lack of space diversity in RSSI measure. Both RSSI and CSI channel quality indicators contain unpredictable variance over time and space due to the presence of people and existing obstacles, which create RF phen...

    work page 2019

  3. [3]

    discriminant features

    Preliminaries and CSI Phase Information 3.1. Channel State Information CSI records the variation in an 802.11n standard channel, experienced during propagation. A wireless signal may experience variance caused due to possible RF phenomena like the multipath effect, fading, shadowing, reflection, refraction, interference and delay distortion. Without a cha...

    work page

  4. [4]

    time-ordered maps

    Hybrid CNN LSTM based Pedestrian Localization 4.1. Dataset Collection A CSI observation dataset was collected in the corridors of Emerging Technologies Building at Texas A&M University. Each training observation consists of 30 sanitized CSI Phase values for 9 MIMO channel pairs between an access point (AP) and a Wi-Fi device as shown in Figure 4. Thus, ea...

    work page

  5. [5]

    Experiment Setup We use a low-power pocket held embedded PC as a Wi-Fi device for collecting test data set

    Results and Performance Evaluation 5.1. Experiment Setup We use a low-power pocket held embedded PC as a Wi-Fi device for collecting test data set. This device is installed with a power optimized Ubuntu LTS configuration; thus, the device can last upwards of 9 hours of data collection activity. The device has an Intel Wi-Fi Link 5300 NIC wireless LAN card...

    work page

  6. [6]

    A temporally stable and diverse sanitized CSI phase value signal is used for fingerprinting

    Conclusion The proposed method delivers robust results in terms of pedestrian localization and heading. A temporally stable and diverse sanitized CSI phase value signal is used for fingerprinting. The novelty of the proposed method lies in representation of the signal to 2D mapping domain and then exploiting this representation to be used in location clas...

    work page

  7. [7]

    We list down the proposed method in brief steps listed below. a. We shape the collected dataset into a 2D+channels CSI Map format. This map 𝑀𝑖 has each pixel associated to a location (𝑥, 𝑦) where training observations were gathered during the pre -requisite fingerprinting collection stage. This step explains the notion of Information Adaptive Sliding Wind...

    work page

  8. [8]

    Smartphone -Based Traveled Distance Estimation Using Individual Walking Patterns for Indoor Localization

    Kang J, Lee J, Eom DS. Smartphone -Based Traveled Distance Estimation Using Individual Walking Patterns for Indoor Localization. Sensors (Basel). 2018;18(9):3149. Published 2018 Sep 18. (As an example of GRU implementation) doi:10.1109/icc.2018.8422562

    work page arXiv 2018

Show all 29 references
  1. [9]

    DeepML: Deep LSTM for indoor localization with smartphone magnetic and light sensors,

    X. Wang, Z. Yu, and S. Mao, “DeepML: Deep LSTM for indoor localization with smartphone magnetic and light sensors,” in Proc. IEEE ICC 2017, Kansas City, MO, May 2018. 2

    2017

  2. [10]

    CSI -Based Fingerprinting for Indoor Localization: A Deep Learning Approach,

    X. Wang, L. Gao, S. Mao and S. Pandey, "CSI -Based Fingerprinting for Indoor Localization: A Deep Learning Approach," in IEEE Transactions on Vehicular Technology, vol. 66, no. 1, pp. 763-776, Jan. 2017

    2017

  3. [11]

    Chung, C

    J. Chung, C. Gulcehre, K. Cho, and Y. Bengio. Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv:1412.3555, 2014

    2014 arXiv

  4. [12]

    BiLoc: Bi -Modal Deep Learning for Indoor Localization with Commodity 5GHz WiFi,

    X. Wang, L. Gao and S. Mao, "BiLoc: Bi -Modal Deep Learning for Indoor Localization with Commodity 5GHz WiFi," in IEEE Access, vol. 5, pp. 4209-4220, 2017

    2017

  5. [13]

    Analysis of Sources of Large Positioning Errors in Deterministic Fingerprinting,

    J. Torres-Sospedra and A. Moreira, “Analysis of Sources of Large Positioning Errors in Deterministic Fingerprinting,” Sensors, vol. 17, no. 12, p. 2736, 2017

    2017

  6. [14]

    Yao, S., Hu, S., Zhao, Y., Zhang, A., & Abdelzaher, T. (2017). DeepSense. Proceedings of the 26th International Conference on World Wide Web - WWW 17. doi:10.1145/3038912.3052577

    2017

  7. [15]

    Wang, X., Yu, Z., & Mao, S. (2018). DeepML: Deep LSTM for Indoor Localization with Smartphone Magnetic and Light Sensors. 2018 IEEE International Conference on Communications (ICC)

    2018

  8. [16]

    CiFi: Deep convolutional neural networks for indoor localization with 5 GHz Wi -Fi

    Wang, Xuyu et al. “CiFi: Deep convolutional neural networks for indoor localization with 5 GHz Wi -Fi.” 2017 IEEE International Conference on Communications (ICC) (2017): 1-6

    2017

  9. [17]

    Wi -Fi Crowdsourced Fingerprinting Dataset for Indoor Positioning

    Lohan, Elena Simona, Torres -Sospedra, Joaquín, Leppäkoski, Helena, Richter, Philipp, Peng, Zhe, Huerta, Joaquín "Wi -Fi Crowdsourced Fingerprinting Dataset for Indoor Positioning", MDPI Data journal

  10. [18]

    Weblink: http://indoorlocplatform.uji.es/databases/all/

  11. [19]

    Multi -person localization via RF body reflections

    Adib, F.; Kabelac, Z.; Katabi, D. Multi -person localization via RF body reflections. In Proceedings of the 12th USENIX Conference on Networked Systems Design and Implementation, Oakland USA, pp.279-292, May 2015

    2015

  12. [20]

    Ambiguity function analysis of wireless LAN transmissions for passive radar

    Colone, F.; Woodbridge, K.; Guo, H.; Mason, D.; Baker, C.J. Ambiguity function analysis of wireless LAN transmissions for passive radar. IEEE Trans. Aerosp. Electron. Syst, 2011, vol. 47, no. 1, pp. 240-264

    2011

  13. [21]

    This motion model is easily configurable, and bounds can be set on normally distributed pedestrian walk phase, speed, stride and heading direction

    was implemented to generate human-like motion trajectories. This motion model is easily configurable, and bounds can be set on normally distributed pedestrian walk phase, speed, stride and heading direction . The simulated noisy data along with the ground truth was fed to PF -...

  14. [22]

    Khalili, A.; Soliman, A. A. Track before mitigate: aspect dependence-based tracking method for multipath mitigation. Electronics Letters, 2016, vol. 52, no. 4, pp. 316-317

    2016

  15. [23]

    A.; Asaduzzaman, M

    Khalili, A.; Soliman, A. A.; Asaduzzaman, M. Quantum particle filter: a multiple mode method for low delay abrupt pedestrian motion tracking. Electronics Letters, 2015, vol. 51, no. 16, pp. 1251-1253

    2015

  16. [24]

    DeepFi: Deep learning for indoor fingerprinting using channel state information,

    X. Wang, L. Gao, S. Mao and S. Pandey, "DeepFi: Deep learning for indoor fingerprinting using channel state information," 2015 IEEE Wireless Communications and Networking Conference (WCNC), New Orleans, LA, 2015, pp. 1666-1671

    2015

  17. [25]

    ConFi: Convolutional Neural Networks Based Indoor Wi -Fi Localization Using Channel State Information,

    H. Chen, Y. Zhang, W. Li, X. Tao and P. Zhang, "ConFi: Convolutional Neural Networks Based Indoor Wi -Fi Localization Using Channel State Information," in IEEE Access, vol. 5, pp. 18066-18074, 2017

    2017

  18. [26]

    PADS: Passive Detection of Moving Targets with Dynamic Speed using PHY Layer Information

    Qian K, Wu C, Yang Z, et al. PADS: Passive Detection of Moving Targets with Dynamic Speed using PHY Layer Information. In: Proc. IEEE ICPADS’14. Hsinchu, Taiwan

  19. [27]

    Deep Learning Convolutional Neural Networks for Radio Identification,

    S. Riyaz, K. Sankhe, S. Ioannidis and K. Chowdhury, "Deep Learning Convolutional Neural Networks for Radio Identification," in IEEE Communications Magazine, vol. 56, no. 9, pp. 146-152, Sept. 2018

    2018

  20. [28]

    Estimation of pedestrian distribution in indoor environments using multiple pedestrian tracking,

    M. Emaduddin and D. A. Shell, "Estimation of pedestrian distribution in indoor environments using multiple pedestrian tracking," in IEEE Intl. conf. on Robotics and Automation, 2009

    2009

  21. [29]

    Straub, J. (2010). Pedestrian Indoor Localization and Tracking using a Particle Filter combined with a learning Accessibility Map (Dissertation). Technische Universität München, Munich, Germany

    2010

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.