Pith. sign in

REVIEW 3 major objections 6 minor 51 references

Solar-CSK: Decoding Color Coded Visible Light Communications using Solar Cells

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read An array of spectrally diverse solar cells, read by a bidirectional LSTM trained on differences from pilot anchors, can decode color-shift-keyed visible light across distance and ambient light without retraining.

desk verdict A credible proof-of-concept for solar-cell CSK decoding, undermined by an overreaching no-retraining claim. read the letter →

arxiv 2505.10226 v1 pith:56ZARNUX submitted 2025-05-15 cs.NI

Yanxiang Wang , Yihe Yan , Jiawei Hu , Cheng Jiang , Brano Kusy , Ashraf Uddin , Mahbub Hassan , Wen Hu This is my paper · ORCID
classification cs.NI
keywords VisibleLightCommunicationColorShiftKeying(CSK)solarcellreceivertandemcellsmulti-materialarraybidirectionalLSTMdifferentialanchorfeaturesenergyharvesting
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

The paper tries to establish that solar cells, which normally respond to all wavelengths at once, can be turned into receivers for color-coded visible light communication (CSK) without the color filters that would sharply reduce energy harvesting. Its proposal is to replace a single broadband solar cell with an array of cells made from different materials, mimicking the layered wavelength selectivity of tandem solar cells, and to decode the resulting coarse color signatures with a bidirectional LSTM fed on anchor-relative differences. On a seven-cell commercial prototype, this anchor-based decoder reports bit error rates well below conventional channel estimation across transmitter-receiver distances from 25 to 50 cm and ambient light from 0 to 1154 lux, with no retraining when the channel changes. A sympathetic reader would care because the result points toward IoT devices that simultaneously harvest light energy and receive color-modulated data from ordinary RGB LEDs.

What carries the argument

The engine of the method is the differential shift from anchors. Each packet begins with anchor symbols; for every incoming sample $d_i$ and each anchor reading $a_j$, the receiver computes $\Delta_{i,j} = |d_i - a_j| / a_j$. Because the anchors travel through the same channel as the payload, this ratio cancels common multiplicative distortions from distance and ambient level. These $\Delta$ vectors are fed as a time sequence to a two-layer bidirectional LSTM with 64 hidden units per direction, followed by a fully connected classifier that outputs symbol probabilities. The multi-material cell array is what makes the input features color-sensitive in the first place: cells with different absorption curves give different voltage signatures for red, green, and blue, unlike an all-silicon array.

What would settle it

Train the LSTM on the published distances and lux levels, then run the same decoder, without retraining, at 60 cm and 70 cm, under 1500 lux or a colored ambient source, and with the receiver in motion; if the CSK-4 bit error rate rises above 1% or the anchor-relative feature vectors for the same symbol at 50 cm and 70 cm separate markedly, the no-retraining claim is refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that partial wavelength selectivity, not narrowband filtering, is sufficient to demodulate CSK. A receiver built from seven off-the-shelf solar cells with different absorption profiles—one polycrystalline, one amorphous, and five organic—produces distinct voltage signatures for different transmitted colors, and a bidirectional LSTM that compares each incoming sample with pilot anchors can classify those signatures. In every distance and ambient-light scenario tested, this anchor-based machine-learning decoder outperforms a classical least-squares channel-estimation decoder and an LSTM fed with raw samples. The paper further claims that because the anchors travel through the same channel as the payload, the decoder needs no retraining or fine-tuning when the channel changes, enabling immediate deployment in new environments.

Load-bearing premise

The scheme assumes that the anchor-relative ratio between a received sample and its anchor stays informative for every channel it will meet; the experiments only test channels inside the trained 25–50 cm and 0–1154 lux ranges, so the 'immediate deployment in diverse environments' claim rests on that ratio remaining discriminative at longer distances, stronger or colored ambient light, and with a moving receiver.

Editorial extensions

If this is right

  • Color-shift-keying receivers no longer need color filters: a small array of differently absorbing solar cells supplies enough coarse spectral information to separate CSK symbols while the cells keep harvesting energy.
  • Because every packet carries its own anchors, the decoder can move between the tested distances and ambient light levels with no recalibration or fine-tuning, cutting deployment overhead for IoT nodes.
  • The anchor-differential LSTM beats both least-squares channel estimation and an LSTM on raw samples in every condition tested, so the differential input, not the neural network alone, drives the generalization.
  • Using seven cells and four anchors, the prototype sustains a 500 Hz baud rate over 25–50 cm with a 2 kHz ADC, and 4-CSK remains usable under 1154 lux ambient light.

Reading between the lines

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

  • A natural stress test the paper leaves implicit is to vary the color of the ambient light, not just its intensity; a colored source might shift the anchor and data features differently and could break the ratio invariance even inside the tested lux range.
  • If the anchor-ratio invariance holds, the same cell array could serve as a general-purpose spectral signature sensor—for indoor localization or gesture recognition—while still harvesting energy, since the multi-material responses are already the input features.
  • The 500 Hz ceiling appears set by solar-cell response time and the 2 kHz ADC, not by the decoding architecture, so faster or optimized tandem cells would likely raise the data rate without changing the decoder.
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 / 6 minor

Summary. The manuscript presents Solar-CSK, a VLC receiver that uses an array of seven spectrally diverse commercial solar cells as a stand-in for tandem solar cells, and decodes IEEE 802.15.7 Color Shift Keying (CSK) signals from an RGB LED. The proposed demodulator is a bidirectional LSTM fed with 'differential shift from anchors' features (Eqs. 8-10). The experimental section reports baud-rate sweeps, cell-count and anchor-count studies, anchor-selection strategies, and leave-one-out evaluations across distances of 25-50 cm and ambient light levels of 0-1154 lux, comparing channel estimation, a raw-input LSTM, and the anchor-input LSTM. The central claim is that anchor-based ML outperforms channel estimation in all tested conditions and requires no retraining or fine-tuning in new environments.

Significance. The work is a credible first demonstration of CSK decoding using solar cells, addressing a genuine limitation of broadband photovoltaic receivers. The experimental design is mostly sound: the LSTM is tested on held-out distances and lux values, the comparison includes a conventional channel-estimation baseline, and the anchor-selection analysis yields a useful practical guideline. The strongest contribution is the empirical evidence that spectral diversity across off-the-shelf cells, combined with a learned decoder, can separate CSK symbols without color filters at modest data rates. The main weakness is that the 'no retraining for diverse environments' claim goes beyond the evidence: the leave-one-out evaluation only covers interpolation within a small grid, and the anchor feature of Eq. (9) does not theoretically cancel additive ambient light. These issues are fixable with additional experiments or a more careful statement of the supported claims.

major comments (3)
  1. [§6.5.1, §6.5.2] The leave-one-out evaluation only demonstrates interpolation, not the 'immediate deployment in diverse environments' claim. The six distances (25-50 cm) and four lux levels (0-1154 lux) all lie inside the training range, and no combined distance-by-ambient condition is tested. To support the no-retraining claim, the authors should either add out-of-range test conditions (e.g., distances of 60-80 cm, lux levels above 1154, and joint distance/ambient variations) or explicitly restrict the claim to interpolation within the tested range.
  2. [§4.3.2, Eq. (9)] The differential feature does not cancel additive ambient light. If each solar-cell reading is approximated as y = g·s + c, then Δ_ij = |d_i - a_j|/a_j = |s_i - s_a|/(s_a + c/g), which depends on the ambient-to-signal ratio c/g. The manuscript's argument that anchor and data symbols 'experience identical channel distortions' justifies invariance only under a common multiplicative gain. Since Section 6.5.2 treats ambient light as a first-class variable and reports color pulling, the claimed no-retraining robustness to ambient changes needs either an analytical bound on the feature's sensitivity to additive perturbations or an explicit extrapolation experiment with high and combined ambient levels.
  3. [Figs. 23-25] The key comparative plots (anchor selection, distance sweep, ambient sweep) report no error bars or confidence intervals, while Figs. 19-20 do. The statement that ML using anchors 'significantly surpasses' channel estimation in all cases would be substantially strengthened by reporting means and variances across the cross-validation folds or repeated trials, especially for the BER curves that cross the 10^-2 line.
minor comments (6)
  1. [§4.3.2, Eq. (9)] The notation f_abs uses |d_i - a_j|/a_j; if an anchor reading a_j can be near zero the feature is unbounded. Please state the practical range of anchor values or add a small regularization term in the denominator.
  2. [§4, Fig. 12] There are small typographical errors: 'diver circuit' should be 'driver circuit' in Section 4, and the Fig. 12 caption 'Optial power' should be 'Optical power.'
  3. [§5] Please provide details of the LSTM input segmentation: the sequence length T, how data symbols and anchor symbols are aligned in the input features, and how many anchor examples are included per packet. This information is needed to reproduce the results.
  4. [§6.5.2] The 'color pulling' effect at 1154 lux is described qualitatively. Quantifying the shift, for example as a displacement in CIE xy coordinates or as a change in per-cell voltage ratios, would make the claim more concrete and testable.
  5. [§2.4, §4.2] The prototype is an array of discrete cells rather than an integrated tandem cell, which the paper acknowledges. To avoid overstating the connection, consider softening 'tandem solar cells' in the title and abstract to 'spectrally diverse solar cell array' where the prototype is meant.
  6. [§5, §6] No statement is made about code or data availability. Given the empirical nature of the work, releasing the data and training scripts would improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the anchor-differential LSTM is evaluated on held-out distances and lux levels; Eq. (9) is a hand-designed feature transform, not a fitted prediction, and author self-citations are used only as background/inspiration.

full rationale

The paper's core claim is empirical: a BiLSTM fed with anchor-differential solar-cell features can decode CSK symbols under unseen distances and ambient-light levels. The leave-one-out protocol in Section 6.5 trains on some distance/lux values and tests on a held-out value inside the tested range; the LSTM weights and hyperparameters are not fit to the test conditions, and the anchor measurements are pilot symbols transmitted in each packet under the same channel, so the differential features are computed from test-time data rather than from the training labels. Equation (9), f_abs = |d_i - a_j| / a_j, is a manually chosen normalization and is not derived from, nor used to predict, the trained model's outputs. Author self-citations (Spectral-Loc [38], SolarGest [22]) are cited as inspiration or as examples of diverse-material solar cells, not as the mathematical or experimental basis for the CSK decoding result; no uniqueness theorem from prior author work is invoked. The paper's 'no retraining' claim is supported only by interpolation over 25-50 cm and 0-1154 lux, and Eq. (9) does not remove additive ambient light, but those are limitations of evidence and robustness, not cases where a prediction reduces to its own input by construction. No fitted parameter is renamed as a prediction, and no load-bearing step is defined in terms of the target result. Accordingly, no circularity is found.

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

The central contribution is empirical; free parameters are calibration values and hyperparameters rather than constants in a derivation. The main ad hoc assumption is the transferability of the multi-cell array to real tandem cells, which is asserted but not validated with an actual tandem device.

free parameters (5)
  • LED channel efficiency calibration coefficients = e_R=1.00, e_G=0.45, e_B=0.75 (relative to red)
    Measured with a true color sensor to equalize optical power across RGB channels (Eq. 2, Section 4.1). These coefficients are specific to the FD-3RGB-Y2 LED and would need recalibration for other LEDs.
  • Per-cell amplifier gains = Not specified
    Section 5: 'the gain values are individually calibrated and adjusted for each solar cell' to ensure uniform sensitivity. These gains are fitted to the specific cells and affect the input features to the LSTM.
  • LSTM hyperparameters = 2 layers, 64 hidden units per direction, dropout 0.2, initial learning rate 1e-3, Adam
    Chosen by hand; standard choices for sequence classification. Not claimed as optimal, but part of the ML method.
  • Number of anchors = 4
    Selected based on Fig. 21, which shows BER saturates at 4 anchors for 4-CSK at 45 cm. This value is then used for all other experiments, so the choice is data-dependent and may not be optimal for other symbol counts or channel conditions.
  • Default baud rate = 500 Hz
    Selected from Fig. 19 as the highest baud rate with BER below 1%. This limits the claimed data rate.
assumptions (5)
  • standard math CIE 1931 color space and color matching functions correctly describe perceived color.
    Used to define CSK constellation points in Section 2.1.
  • standard math IEEE 802.15.7 CSK modulation standard defines valid symbol constellations.
    Used as the modulation scheme; constellation points are taken from the standard (Section 2.2).
  • ad hoc to paper A discrete array of different-material solar cells approximates the spectral selectivity of true tandem (multi-junction) solar cells.
    Section 2.4 and Section 4.2: the prototype uses separate cells because commercial tandem cells are unavailable. Actual tandem cells typically provide a single series-connected output, not per-layer voltage outputs; the paper does not demonstrate that a true two-terminal tandem cell would yield the same multi-dimensional features.
  • domain assumption Anchor symbols and data symbols experience identical channel conditions within a packet.
    Section 4.3.2: 'Since both data packet symbols and anchor symbols experience identical channel conditions, their relative differences remain consistent.' This holds when the channel is static over the packet duration; fast-varying channels or moving receivers would break the differential invariance.
  • domain assumption The differential ratio feature |d_i - a_j| / a_j is approximately invariant across the tested channel changes (distance, ambient light).
    This is the load-bearing generalization assumption. The leave-one-out experiments provide evidence within the tested ranges, but only for interpolated conditions; the form of f_abs is chosen ad hoc.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Solar-CSK: Decoding Color Coded Visible Light Communications using Solar Cells." pith.science (2026). https://pith.science/paper/56ZARNUX

@misc{pith2026250510226,
  author       = {Pith},
  title        = {Pith review of: Solar-CSK: Decoding Color Coded Visible Light Communications using Solar Cells},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/56ZARNUX}},
  note         = {Machine review of arXiv:2505.10226}
}
read the original abstract

Visible Light Communication (VLC) provides an energy-efficient wireless solution by using existing LED-based illumination for high-speed data transmissions. Although solar cells offer the advantage of simultaneous energy harvesting and data reception, their broadband nature hinders accurate decoding of color-coded signals like Color Shift Keying (CSK). In this paper, we propose a novel approach exploiting the concept of tandem solar cells, multi-layer devices with partial wavelength selectivity, to capture coarse color information without resorting to energy-limiting color filters. To address the residual spectral overlap, we develop a bidirectional LSTM-based machine learning framework that infers channel characteristics by comparing solar cells' photovoltaic signals with pilot-based anchor data. Our commercial off-the-shelf (COTS) solar prototype achieves robust performance across varying distances and ambient lighting levels, significantly reducing bit error rates compared to conventional channel estimation methods. These findings mark a step toward sustainable, high-performance VLC systems powered by the multi-layer solar technologies.

Figures

Figures reproduced from arXiv: 2505.10226 by the authors.

Figure 1
Figure 1. A multi-material solar array decodes CSK sig [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. (a) CIE color space. (b) CIE XYZ standard ob [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗
Figure 3
Figure 3. IEEE 802.15.7 CSK constellations (blue dots). [PITH_FULL_IMAGE:figures/full_fig_p002_3.png] view at source ↗
Figures from the paper (17 more)
Figure 4
Figure 4. Figure 4: Spectral response of the true color sensor AS73211. 400 450 500 550 600 650 700 750 800 Wavelength 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Absorption Coefficient [PITH_FULL_IMAGE:figures/full_fig_p003_4.png]
Figure 7
Figure 7. Figure 7: (a) Seven silicon solar cells; (b) Seven multi [PITH_FULL_IMAGE:figures/full_fig_p003_7.png]
Figure 9
Figure 9. Figure 9: t-SNE visualization (a) seven silicon solar cells, [PITH_FULL_IMAGE:figures/full_fig_p004_9.png]
Figure 8
Figure 8. Figure 8: The time-domain voltage measurements from [PITH_FULL_IMAGE:figures/full_fig_p004_8.png]
Figure 11
Figure 11. Figure 11: Circuit diagram of the transmitter. three distinct wavelength channels: red centered at 625 nm, green at 525 nm, and blue at 465 nm. In addition, RGB LEDs are also widely applied in smart lighting infrastructure, al￾lowing for straightforward retrofitting and compatib…
Figure 12
Figure 12. Figure 12: Optial power vs PWM values for different [PITH_FULL_IMAGE:figures/full_fig_p005_12.png]
Figure 13
Figure 13. Figure 13: Format of the data packet. 4.2 Receiver Tandem solar cells are designed to enhance spectral coverage and response by incorporating multiple junctions with well￾defined and complementary absorption characteristics [4, 32, 44]. By leveraging the spectral selectivity of …
Figure 14
Figure 14. Figure 14: The demodulation pipeline for 𝑁-CSK em￾ploys a bidirectional LSTM followed by a fully con￾nected layer to process photovoltaic signals. This ar￾chitecture outputs probability distributions across the 𝑁 symbols defined in the 𝑁-CSK modulation scheme. 4.3 Demodulation T…
Figure 15
Figure 15. Figure 15: Differential differences between the sample [PITH_FULL_IMAGE:figures/full_fig_p007_15.png]
Figure 16
Figure 16. Figure 16: Solar-CSK transmitter [PITH_FULL_IMAGE:figures/full_fig_p007_16.png]
Figure 18
Figure 18. Figure 18: Experimental setup of Solar-CSK [PITH_FULL_IMAGE:figures/full_fig_p008_18.png]
Figure 17
Figure 17. Figure 17: Solar-CSK receiver. integrated analog-to-digital converter (ADC) within the Ar￾duino Due micro controller platform, which offers a high￾precision 12-bit resolution for accurate signal quantization. To ensure synchronized data acquisition, the sampling rate is specific…
Figure 19
Figure 19. Figure 19: Baud rate vs. BER. experiments were performed under dark conditions (0 Lux). The transmitter’s baud rate was varied from 50 Hz to 1000 Hz, while all solar cells were sampled at 2 kHz using the Arduino ADC. For signal transmission, the ASCII-encoded word ‘hello’ was us…
Figure 20
Figure 20. Figure 20: Number of cells vs. BER. 6.2 Impact of the Number of Solar Cells Following the determination of the optimal baud rate at 500 Hz, we conducted a detailed investigation into the relation￾ship between the number of solar cells and system perfor￾mance. To systematically e…
Figure 21
Figure 21. Figure 21: BER vs. # of anchors [PITH_FULL_IMAGE:figures/full_fig_p011_21.png]
Figure 24
Figure 24. Figure 24: BER as a function of distance. 0 147 618 1154 Ambient Light Intensity (Lux) 10 −2 10 −1 10 0 BER BER = 0.01 CSK-4 0 147 618 1154 Ambient Light Intensity (Lux) CSK-8 0 147 618 1154 Ambient Light Intensity (Lux) CSK-16 Channel Estimation ML wo/anchor ML w/anchor [PITH_…
Figure 25
Figure 25. Figure 25: BER as a function of ambient light intensity. [PITH_FULL_IMAGE:figures/full_fig_p011_25.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

51 extracted references · 51 canonical work pages

  1. [1]

    Adafruit. 2024. 3W RGB LED - Common Anode. https://www.adafruit. com/product/2530

    work page 2024

  2. [2]

    ams OSRAM. 2024. AS73211. https://ams.com/as73211

    work page 2024

  3. [3]

    IEEE Standard Association et al . 2011. IEEE standard for local and metropolitan area networks-Part 15.7: short-range wireless optical communication using visible light.IEEE: Piscataway, NZ, USA(2011), 1–309

    work page 2011

  4. [4]

    Andreas W Bett, Frank Dimroth, G Stollwerck, and OV Sulima. 1999. III-V compounds for solar cell applications.Applied Physics A69 (1999), 119–129

    work page 1999

  5. [5]

    Sovan Das, Enrique Poves, John Fakidis, Adrian Sparks, Stefan Videv, and Harald Haas. 2019. Towards energy neutral wireless communi- cations: Photovoltaic cells to connect remote areas.Energies12, 19 (2019), 3772

    work page 2019

  6. [6]

    José Ilton De Oliveira Filho, Abderrahmen Trichili, Boon S Ooi, Mohamed-Slim Alouini, and Khaled Nabil Salama. 2022. Self-powered weather station for remote areas and difficult-access locations.Optics Express30, 2 (2022), 2668–2679

    work page 2022

  7. [7]

    Digikey. 2024. AO3400A. https://www.digikey.com.au/en/products/ detail/alpha-omega-semiconductor-inc/AO3400A/1855772

    work page arXiv 2024

  8. [8]

    Stephanie Essig, Myles A Steiner, Christophe Allebe, John F Geisz, Bertrand Paviet-Salomon, Scott Ward, Antoine Descoeudres, Vincenzo LaSalvia, Loris Barraud, Nicolas Badel, et al . 2016. Realization of GaInP/Si dual-junction solar cells with 29.8% 1-sun efficiency.IEEE Journal of Photovoltaics6, 4 (2016), 1012–1019

    work page 2016

Show all 51 references
  1. [9]

    Stephanie Essig, Scott Ward, Myles A Steiner, Daniel J Friedman, John F Geisz, Paul Stradins, and David L Young. 2015. Progress towards a 30% efficient GaInP/Si tandem solar cell.Energy Procedia77 (2015), 464–469

    2015

  2. [10]

    John Fakidis, Henning Helmers, and Harald Haas. 2020. Simultaneous wireless data and power transfer for a 1-Gb/s GaAs VCSEL and photo- voltaic link.IEEE photonics technology letters32, 19 (2020), 1277–1280

    2020

  3. [11]

    Seyed Keyarash Ghiasi and Marco Zuniga. 2024. Exploiting Polariza- tion and Color to Enable MIMO Backscattering with Light. InPro- ceedings of the 22nd ACM Conference on Embedded Networked Sensor Systems. 771–783

    2024

  4. [12]

    Martin A Green and Mark J Keevers. 1995. Optical properties of intrin- sic silicon at 300 K.Progress in Photovoltaics: Research and applications 3, 3 (1995), 189–192

    1995

  5. [13]

    Liane Grobe, Anagnostis Paraskevopoulos, Jonas Hilt, Dominic Schulz, Friedrich Lassak, Florian Hartlieb, Christoph Kottke, Volker Jung- nickel, and Klaus-Dieter Langer. 2013. High-speed visible light com- munication systems.IEEE communications magazine51, 12 (2013), 60–66

    2013

  6. [14]

    Pengfei Hu, Parth H Pathak, Xiaotao Feng, Hao Fu, and Prasant Mo- hapatra. 2015. Colorbars: Increasing data rate of led-to-camera com- munication using color shift keying. Inproceedings of the 11th ACM conference on Emerging Networking experiments and technologies. 1–13

    2015

  7. [15]

    Konthoujam James Singh, Yu-Ming Huang, Tanveer Ahmed, An-Chen Liu, Sung-Wen Huang Chen, Fang-Jyun Liou, Tingzhu Wu, Chien- Chung Lin, Chi-Wai Chow, Gong-Ru Lin, et al. 2020. Micro-LED as a promising candidate for high-speed visible light communication. Applied Sciences10, 20 (...

    2020

  8. [16]

    Sindhubala Kadirvelu, Walter D Leon-Salas, Xiaozhe Fan, Jongseok Kim, Borja Peleato, Saeed Mohammadi, and B Vijayalakshmi. 2021. A circuit for simultaneous reception of data and power using a solar cell.IEEE Transactions on Green Communications and Networking5, 4 (2021), 2065–2075

    2021

  9. [17]

    Meiwei Kong, Jiaming Lin, Chun Hong Kang, Chao Shen, Yujian Guo, Xiaobin Sun, Mohammed Sait, Yang Weng, Huafan Zhang, Tien Khee Ng, et al. 2019. Toward self-powered and reliable visible light commu- nication using amorphous silicon thin-film solar cells.Optics express 27, 24 (...

    2019

  10. [18]

    Chenning Li, Hanqing Guo, Shuai Tong, Xiao Zeng, Zhichao Cao, Mi Zhang, Qiben Yan, Li Xiao, Jiliang Wang, and Yunhao Liu. 2021. NELoRa: Towards ultra-low SNR LoRa communication with neural- enhanced demodulation. InProceedings of the 19th ACM Conference on Embedded Networked S...

    2021

  11. [19]

    Chenning Li, Yidong Ren, Shuai Tong, Shakhrul Iman Siam, Mi Zhang, Jiliang Wang, Yunhao Liu, and Zhichao Cao. 2024. ChirpTransformer: Versatile LoRa encoding for low-power wide-area IoT. InProceed- ings of the 22nd Annual International Conference on Mobile Systems, Application...

    2024

  12. [20]

    Honglei Li, Xiongbin Chen, Junqing Guo, and Hongda Chen. 2014. A 550 Mbit/s real-time visible light communication system based on phosphorescent white light LED for practical high-speed low- complexity application.Optics express22, 22 (2014), 27203–27213

    2014

  13. [21]

    Tianxing Li, Qiang Liu, and Xia Zhou. 2016. Practical human sensing in the light. InProceedings of the 14th Annual International Conference on Mobile Systems, Applications, and Services. 71–84

    2016

  14. [22]

    Dong Ma, Guohao Lan, Mahbub Hassan, Wen Hu, Mushfika B Upama, Ashraf Uddin, and Moustafa Youssef. 2019. Solargest: Ubiquitous and battery-free gesture recognition using solar cells. InThe 25th annual international conference on mobile computing and networking. 1–15

    2019

  15. [23]

    Roger Martinez, Francisco Eugenio Lopez Giraldo, Jose Martin Luna Rivera, Juan David Navarro Restrepo, and Juan David Rojas Usuga

  16. [24]

    Roger Alexander Martínez-Ciro, Francisco Eugenio López-Giraldo, Andrés Felipe Betancur-Perez, and Jose Martín Luna-Rivera. 2019. Design and implementation of a multi-colour visible light communica- tion system based on a light-to-frequency receiver. InPhotonics, Vol. 6. MDPI, 42

    2019

  17. [25]

    Eric Monteiro and Steve Hranilovic. 2014. Design and implementation of color-shift keying for visible light communications.Journal of Lightwave Technology32, 10 (2014), 2053–2060

    2014

  18. [26]

    Sandro Nižetić, Petar Šolić, Diego Lopez-de-Ipiña Gonzalez-De, Luigi Patrono, et al. 2020. Internet of Things (IoT): Opportunities, issues and challenges towards a smart and sustainable future.Journal of cleaner production274 (2020), 122877

    2020

  19. [27]

    Omar M Saif, Abdel Halim Zekry, Mohamed Abouelatta, and Ahmed Shaker. 2023. A comprehensive review of tandem solar cells integrated on silicon substrate: III/V vs perovskite.Silicon15, 15 (2023), 6329– 6347. Wang et al

    2023

  20. [28]

    Kunal Sankhe, Mauro Belgiovine, Fan Zhou, Shamnaz Riyaz, Stratis Ioannidis, and Kaushik Chowdhury. 2019. ORACLE: Optimized ra- dio classification through convolutional neural networks. InIEEE INFOCOM 2019-IEEE conference on computer communications. IEEE, 370–378

    2019

  21. [29]

    Rohail Sarwar, Bin Sun, Meiwei Kong, Tariq Ali, Chuying Yu, Bo Cong, and Jing Xu. 2017. Visible light communication using a solar-panel receiver. In2017 16th International Conference on Optical Communica- tions and Networks (ICOCN). IEEE, 1–3

    2017

  22. [30]

    Won-Ho Shin, Se-Hoon Yang, Do-Hoon Kwon, and Sang-Kook Han

  23. [31]

    Thomas Smith and John Guild. 1931. The CIE colorimetric standards and their use.Transactions of the optical society33, 3 (1931), 73

    1931

  24. [32]

    Henry J Snaith. 2013. Perovskites: the emergence of a new era for low-cost, high-efficiency solar cells.The journal of physical chemistry letters4, 21 (2013), 3623–3630

    2013

  25. [33]

    Engineering ToolBox. 2004. Illuminance - recommended light level. https://www.engineeringtoolbox.com/

    2004

  26. [34]

    M Umeno, H Shimizu, T Egawa, T Soga, and T Jimbo. 1991. First results of AlGaAs/Si monolithic 2-terminal tandem solar cell grown by MOCVD. InThe Conference Record of the Twenty-Second IEEE Pho- tovoltaic Specialists Conference-1991. IEEE, 361–364

    1991

  27. [35]

    Yoshinori Umetsu, Yugo Nakamura, Yutaka Arakawa, Manato Fuji- moto, and Hirohiko Suwa. 2019. Ehaas: Energy harvesters as a sensor for place recognition on wearables. In2019 IEEE International Con- ference on Pervasive Computing and Communications (PerCom. IEEE, 1–10

    2019

  28. [36]

    Hao-Yu Wang, Jhao-Ting Wu, Chi-Wai Chow, Yang Liu, Chien-Hung Yeh, Xin-Lan Liao, Kun-Hsien Lin, Wei-Liang Wu, and Yi-Yuan Chen

  29. [37]

    Yiguang Wang, Nan Chi, Yuanquan Wang, Li Tao, and Jianyang Shi

  30. [38]

    Yanxiang Wang, Jiawei Hu, Hong Jia, Wen Hu, Mahbub Hassan, Ashraf Uddin, Brano Kusy, and Moustafa Youssef. 2023. Spectral-Loc: Indoor localization using light spectral information.Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies7, 1 (2023), 1–26

    2023

  31. [39]

    Xiao Wu, Inhee Lee, Qing Dong, Kaiyuan Yang, Dongkwun Kim, Jingcheng Wang, Yimai Peng, Yiqun Zhang, Mehdi Saligane, Makoto Yasuda, et al. 2018. A 0.04 MM 3 16NW wireless and batteryless sensor system with integrated cortex-m0+ processor and optical communica- tion for cellular...

    2018

  32. [40]

    2000.Color science: concepts and methods, quantitative data and formulae

    Günther Wyszecki and Walter Stanley Stiles. 2000.Color science: concepts and methods, quantitative data and formulae. Vol. 40. John wiley & sons

    2000

  33. [41]

    Kenuo Xu, Chen Gong, Bo Liang, Yue Wu, Boya Di, Lingyang Song, and Chenren Xu. 2022. Low-latency visible light backscatter networking with retromumimo. InProceedings of the 20th ACM Conference on Embedded Networked Sensor Systems. 448–461

    2022

  34. [42]

    Kenuo Xu, Kexing Zhou, Chengxuan Zhu, Shanghang Zhang, Boxin Shi, Xiaoqiang Li, Tiejun Huang, and Chenren Xu. 2023. When Visible Light (Backscatter) Communication Meets Neuromorphic Cameras in V2X. InProceedings of the 24th International Workshop on Mobile Computing Systems an...

    2023

  35. [43]

    Talia Xu, Miguel Chávez Tapia, and Marco Zúñiga. 2022. Exploiting digital{Micro-Mirror} devices for ambient light communication. In 19th USENIX Symposium on Networked Systems Design and Implemen- tation (NSDI 22). 387–400

    2022

  36. [44]

    Masafumi Yamaguchi, Kan-Hua Lee, Kenji Araki, and Nobuaki Kojima

  37. [45]

    Hanting Ye and Qing Wang. 2021. SpiderWeb: Enabling Through- Screen Visible Light Communication. InProceedings of the 19th ACM Conference on Embedded Networked Sensor Systems. 316–328

    2021

  38. [46]

    Hanting Ye, Jie Xiong, and Qing Wang. 2023. When vlc meets under- screen camera. InProceedings of the 21st Annual International Confer- ence on Mobile Systems, Applications and Services. 343–355

    2023

  39. [49]

    A review of recent progress in heterogeneous silicon tandem solar cells.Journal of Physics D: Applied Physics51, 13 (2018), 133002

    2018

  40. [2014]

    In2014 Sixth International Conference on Wireless Communications and Signal Processing (WCSP)

    High speed LED based visible light communication networks for beyond 10Gb/s wireless access. In2014 Sixth International Conference on Wireless Communications and Signal Processing (WCSP). IEEE, 1–6

  41. [2016]

    Self-reverse-biased solar panel optical receiver for simultaneous visible light communication and energy harvesting.Optics express24, 22 (2016), A1300–A1305

    2016

  42. [2018]

    Optics Communications407 (2018), 245–249

    Using pre-distorted PAM-4 signal and parallel resistance circuit to enhance the passive solar cell based visible light communication. Optics Communications407 (2018), 245–249

    2018

  43. [2024]

    Evaluation of Solar Panel Bandwidth for RGB Channels in Visible Light Communication.IEEE Latin America Transactions22, 3 (2024), 240–248

    2024

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.