Round 1

Reviewer 1 Report

The authors presents the first SDR-based VLC prototypes with LabVIEW. Main contribution of the paper is the extensive evaluation of the practical implementation of VLC systems with ceiling light and car taillight. The experimenral results are highly valuable for VLC-based positioning systems and V2X systems. However, the authors should revise the manuscript regarding following comments. 

- Introduction part: Lack of theory background and reference models.The literature review in the introduction part should be rewritten with more related works. 

- Introduction part: the research problem was not clearly mentioned. Also, the authors did not mention the motivation for practical implementation of VLC prototypes.

- Introduction part: Avoid using "we wanted", "we decided", the authors should add more text to convince using Skoda Octavia III taillight, Phillips Fortimo DLM 300, Thorlabs PDA36A-EC PIN.

- The authors should give more comments on results presented in Figure13 and Figure 14.

- The authors should give more comments on results presented in Figure 10; especially in range of 11-15 Hz.

- Legends of Figure 6, Figure 13, Figure 14 should be larger in size to be well recognized in print-version.

- Figure 13, 14, 17, 18, 22, 23, 27, 28, 34, 35: Various lines of EbN0 - distance relationship look similarly in print-version. The authors should revise these figures. 

Author Response

Dear Reviewer 1,

Thank you very much for all your constructive comments. Let us present our notes with explanations and adjustments. Every change/addition is marked by red color.

Introduction part: Lack of theory background and reference models. The literature review in the introduction part should be rewritten with more related works. 
We have greatly expanded introduction part. New figures and many new references were added. We have also revamped every reference and reorganized a number of paragraphs.

Introduction part: the research problem was not clearly mentioned. Also, the authors did not mention the motivation for practical implementation of VLC prototypes.
This section was also greatly expanded and adjusted. Appropriate references were added to back everything up.

Introduction part: Avoid using "we wanted", "we decided", the authors should add more text to convince using Skoda Octavia III taillight, Phillips Fortimo DLM 300, Thorlabs PDA36A-EC PIN.
Tables for each component were added. A small paragraph about each component was added as well, explaining choice of light sources or photodetectors.

The authors should give more comments on results presented in Figure13 and Figure 14. The authors should give more comments on results presented in Figure 10; especially in range of 11-15 Hz.
Small paragraph was added to present our results better. Please doublecheck it and if there are any other concerns, let us know.

Legends of Figure 6, Figure 13, Figure 14 should be larger in size to be well recognized in print-version.
Legend size was increased. Please doublecheck if everything is alright now.

Figure 13, 14, 17, 18, 22, 23, 27, 28, 34, 35: Various lines of EbN0 - distance relationship look similarly in print-version. The authors should revise these figures. 
Lines were adjusted to look a bit different. Every figure was revised according to suggestions.

Best regards,

Mr. Rene Jaros, Mr. Lukas Danys, and Dr. Radek Martinek

 

Reviewer 2 Report

The claim made in the abstract is ambiguous: is it claiming to be the first VLC system based on SDR or the first VLC system based on Labview implementation of SDR?  Either claim would be incorrect (e.g., "Demonstration of a Software Defined Visible Light Communication System" by Rahaim, et al. in 2011 or "Software defined VLC system: Implementation and performance evaluation" by Hussain et al. in 2015).

Introduction

"Current communication bands tend to be overfilled" implies that the communication bands are carrying more information than their capacity, which wouldn't be possible.

The claim that "VLC also offers higher security of transmission because this technology works 34 only when there is a direct visibility between transmitter and receiver" is incorrect; reflections may also leak information to eavesdroppers, or allow adversaries to insert signals.

Experimental Setup

The paper does not define "RSL".

Feasibility on Indoor VLC

The paper does not define "MER".

What does "Message symbols: 10 000" mean?  Only 10,000 symbols were sent?  That wouldn't be sufficient to measure BER = 10e-6 as the paper claims.

The paper does not describe how the changes in modulation are negotiated or communicated (i.e., how the adaptation happens) for this "adaptive modulation".

The paper fails to describe how the "red line" that "represents FEC limit" was calculated, which makes it difficult to check for correctness.

Figure 6 does not clearly describe what the values shown next to the arrows are (i.e., what do those values represent?).

Figure 7 (and the paper in general) fails to describe whether E_b is the optical energy (e.g., at the input of the photodetector) or the electrical energy (e.g., at the output of the photodetector), and does not describe how it was measured.

Feasibility on Outdoor VLC

How do you measure for a 550 cm range in a 500cm long box?

The paper does not define "EVM".

The paper's setup seems to have a partition that can get fogged up or rain on it; as such, the effects described aren't just from the atmospheric conditions described, but also (probably predominantly) from the wet/foggy partition that obstructs the light.  The author does not mention the temperature of the "rain" water or the humidity level, which would affect the fogging of the partition. The paper also does not mention whether measurements were taken after the partition has reached a steady-state temperature (with the "rain"), which would also affect fogging.  I would strongly recommend redoing the experiments without these partitions because the partitions negatively impact the reproducibility of the experiments.

The transparent partitions used may also be acting as a light-guide, an unintended effect.

How do the authors know that the laser is not having other effects on the experiment? E.g., increasing the shot noise?

Discussion

Figure 36 does not make sense to me: why is attenuation in fog a function of laser power?

Author Response

Dear Reviewer 2,

Thank you very much for all your constructive comments. Let us present our notes with explanations and adjustments. Every change/addition is marked by red color.

The claim made in the abstract is ambiguous: is it claiming to be the first VLC system based on SDR or the first VLC system based on LabVIEW implementation of SDR?  Either claim would be incorrect (e.g., "Demonstration of a Software Defined Visible Light Communication System" by Rahaim, et al. in 2011 or "Software defined VLC system: Implementation and performance evaluation" by Hussain et al. in 2015).

We have rephrased first sentence. Also, a whole paragraph was added about similar implementations based on LabVIEW and SDR.

Introduction

"Current communication bands tend to be overfilled" implies that the communication bands are carrying more information than their capacity, which wouldn't be possible.
Sentence was rephrased to avoid misunderstandings.

The claim that "VLC also offers higher security of transmission because this technology works 34 only when there is a direct visibility between transmitter and receiver" is incorrect; reflections may also leak information to eavesdroppers or allow adversaries to insert signals.
Sentence containing this statement was deleted. Instead a short paragraph mentioning VLC security was added to discussion.

Experimental Setup

The paper does not define "RSL".
RSL was defined in short paragraph.

Feasibility on Indoor VLC

The paper does not define "MER".
MER was defined in short paragraph.

What does "Message symbols: 10 000" mean?  Only 10,000 symbols were sent?  That wouldn't be sufficient to measure BER = 10e-6 as the paper claims.
We have adjusted our scaling in figures, since whole measurement was only up to BER values of 10e-5.

The paper does not describe how the changes in modulation are negotiated or communicated (i.e., how the adaptation happens) for this "adaptive modulation".
Measurements were carried out using static modulation schemes. Adaptive modulation will be a topic of further research. However, we have included a paragraph explaining how it will work.

The paper fails to describe how the "red line" that "represents FEC limit" was calculated, which makes it difficult to check for correctness.
FEC limit was explained in a short paragraph. We have added a few references as well.

Figure 6 does not clearly describe what the values shown next to the arrows are (i.e., what do those values represent?).
A sentence was added to clarify values next to arrows.

Figure 7 (and the paper in general) fails to describe whether E_b is the optical energy (e.g., at the input of the photodetector) or the electrical energy (e.g., at the output of the photodetector), and does not describe how it was measured.
Eb/N0 was defined in a short paragraph. A few references to its calculation were added to text.

Feasibility on Outdoor VLC

How do you measure for a 550 cm range in a 12500px long box?
Transmitter and receiver are situated outside of box which is open from both sides. A paragraph was added to clarify it.

The paper does not define "EVM".
EVM was defined in short paragraph.

The paper's setup seems to have a partition that can get fogged up or rain on it; as such, the effects described aren't just from the atmospheric conditions described, but also (probably predominantly) from the wet/foggy partition that obstructs the light.  The author does not mention the temperature of the "rain" water or the humidity level, which would affect the fogging of the partition. The paper also does not mention whether measurements were taken after the partition has reached a steady-state temperature (with the "rain"), which would also affect fogging.  I would strongly recommend redoing the experiments without these partitions because the partitions negatively impact the reproducibility of the experiments.
Thank you for this comment. We have carried out a number of experiments and difference with or without this partition were negligible. We have added a short paragraph about water and ambient temperatures. Another thing is, that the first iteration of our system was already dismantled, and we are currently building a new version, which is based on OFDM. Nevertheless, we will definitely adjust these experiments in future iterations.

The transparent partitions used may also be acting as a light-guide, an unintended effect.
According to earlier experiments, the transparent partitions had negligible effect on communication. Since the light was focused using planoconvex lens, only the main „beam“ was received by photodetector.

How do the authors know that the laser is not having other effects on the experiment? E.g., increasing the shot noise?
As mentioned earlier, we are using planoconvex lens, so laser had no effect on experiment.

Discussion

Figure 36 does not make sense to me: why is attenuation in fog a function of laser power?
We had to monitor fog dissipation somehow, and since we don’t have fog monitor, received laser power could be used to monitor it instead.

The whole paper was significantly expanded and improved, mainly thanks to your comments. We are thankful for your help and will keep everything in mind during measurements on new version of our system

Best regards,

Mr. Rene Jaros, Mr. Lukas Danys, and Dr. Radek Martinek

 

Reviewer 3 Report

This paper introduces the implementation of a VLC system with USRP and afterward the fruitful measurements under different settings.

Although scientifically speaking, the novelty is limited; the measurements could give other researchers information about the real performance of VLC under different working conditions.

I do not buy the claim that the authors implement the first prototype of VLC..SDR…LabVIEW. I can easily find many other implementations. See below:

https://www.researchgate.net/publication/261076943_An_SDR_implementation_of_a_visible_light_communication_system_based_on_the_IEEE_802157_standard

http://wnl.ku.edu.tr/uploads/1/0/5/9/10590997/vlc_ieeevtc_phy.pdf

https://cdn.intechopen.com/pdfs/55586.pdf

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7342278

http://jultika.oulu.fi/files/nbnfioulu-201502141097.pdf

https://arxiv.org/ftp/arxiv/papers/1701/1701.01156.pdf

https://hal.archives-ouvertes.fr/hal-01590267/document

https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7887870

Please double check.

In Figure 1 and Figure 9, I do not understand on the bottom why you connect USRP1 and USRP2 with the LabVIEW. Don’t you run LabVIEW at each USRP? 

Are the results shown in Figure 4 from analysis or experiments? If from experiments, why so perfect (seems that the BER is 0). Please clarify them.

In Figure 5, should the results of 5b and 5c swap? Please check.

In Figure 3, why the Attenuation first drops and then increases when the frequency is below 1MHz? Please explain the reason. Same for Figure 10.

In Figure 10, when the frequency is about 5MHz, why the Attenuation increases?

Is the photodiode able to detect signals modulated with 1024QAM/2048QAM?

In Figures 24, 29, and 36, x-axis, Frequency (db)?

Minor:

P1: Testing methodic -> Testing method

P1: radio(SDR) -> radio (SDR)

P5: figure 4 -> Figure 4;  figure 6 -> Figure 6.

Please check other typos?

Author Response

Dear Reviewer 3,

Thank you very much for all your constructive comments. Let us present our notes with explanations and adjustments. Every change/addition is marked by red color.

I do not buy the claim that the authors implement the first prototype of VLC..SDR…LabVIEW. I can easily find many other implementations. Please double check.

A whole paragraph about different LabVIEW SDR implementations were added. Introduction and discussion were greatly expanded, and multiple new references were used as well.

 

In Figure 1 and Figure 9, I do not understand on the bottom why you connect USRP1 and USRP2 with the LabVIEW. Don’t you run LabVIEW at each USRP?
Both figures were adjusted, since, as you mentioned, we run LabVIEW on both.

 

Are the results shown in Figure 4 from analysis or experiments? If from experiments, why so perfect (seems that the BER is 0). Please clarify them.

A whole paragraph about this problematic was added to better explain it. If needed, we can expand it even more.

 

In Figure 5, should the results of 5b and 5c swap? Please check.

We have doublechecked figure 5 and adjusted it.

 

In Figure 3, why the Attenuation first drops and then increases when the frequency is below 1MHz? Please explain the reason. Same for Figure 10.
A paragraph was added, discussing why it happened.

 

In Figure 10, when the frequency is about 5MHz, why the Attenuation increases?
Another paragraph was added, discussing why it happened.

 

Is the photodiode able to detect signals modulated with 1024QAM/2048QAM?
Photodiode is fast enough to detect even higher state modulations. However, it is needed to turn off TIA inside photodetector since it severely limits its functionality. We have already ordered APD photodiode from Hamamatsu Photonics with better TIA.

In Figures 24, 29, and 36, x-axis, Frequency (db)?
We have adjusted all figures. Thank you for this comment.

 

 

 

Minor:

P1: Testing methodic -> Testing method

P1: radio(SDR) -> radio (SDR)

P5: figure 4 -> Figure 4;  figure 6 -> Figure 6.

Please check other typos?
Multiple typos in text were fixed.

 

Best regards,

Mr. Rene Jaros, Mr. Lukas Danys, and Dr. Radek Martinek

 

Reviewer 4 Report

The paper presents experimental results on VLC applied to outdoor and indoor scenarios.

The paper presents a consistent work and appears very interesting.

However I have some concerns:

- the literature review and related works are almost absent. Instead, they are necessary to better understand the contest and motivation of the work and, above all, the novel contribution. Just to give few examples, 

Vehicular Visible Light Networks for Urban Mobile Crowd Sensing, 2018

Experimental verification of visible light communications based on multi-band CAP modulation, 2015

An experimental evaluation and prototyping for visible light communication

etc.

-  Figure 4  represent possible maximal reachable transmission speeds for different M-QAM and bandwidths. The authors should explain why the curves are made of steps, hence how they have been obtained.

- In addition, which was the light interference for the different measurements? How many lux for example?

- In outdoor the distance reached was around the same as in indoor, around 350cm. Isn't it strange in different ambient light conditions?

- Why in Fig. 12, BER is higher for lower bandwidth? I expected higher data rate but also higher BER. Different behaviors appear in Fig. 16 and 21.

Author Response

Dear Reviewer 4,

Thank you very much for all your constructive comments. Let us present our notes with explanations and adjustments. Every change/addition is marked by red color.

However I have some concerns:

- the literature review and related works are almost absent. Instead, they are necessary to better understand the contest and motivation of the work and, above all, the novel contribution. Just to give few examples, 
We have greatly expanded introduction and discussion. A number of references were added and multiple parameters were defined.

-  Figure 4  represent possible maximal reachable transmission speeds for different M-QAM and bandwidths. The authors should explain why the curves are made of steps, hence how they have been obtained.
A paragraph about this problematic was added. We can clarify it even more in next iteration if needed.

- In addition, which was the light interference for the different measurements? How many lux for example?
We have carried out a number of tests to test interference from ambient sources. However, we have concluded that it was negligible. Results in dark room were almost the same as from experiments carried in lab during bright sunny day.

- In outdoor the distance reached was around the same as in indoor, around 8750px. Isn't it strange in different ambient light conditions?
Outdoor distance was mainly limited by box length. Also, characteristics of each light source is a lot different. Ceiling light is composed of one powerful matrix, which is illumination conical area, whereas taillight is illuminating much narrower area and is using only a few LEDs.

- Why in Fig. 12, BER is higher for lower bandwidth? I expected higher data rate but also higher BER. Different behaviors appear in Fig. 16 and 21.
It is caused by characteristics of parts. We have already looked into it and ordered a different and more linear amplifier and new set of APDs with linear trans-impedance amplifier from Hamamatsu Photonics Japan.

 

Best regards,

Mr. Rene Jaros, Mr. Lukas Danys, and Dr. Radek Martinek

 

 

Reviewer 5 Report

The authors deploy a visible light communication system based on Software Defined Radio (SDR) in LabVIEW and they examine in detail two different types of LED lights, i.e., ceiling lights and LED car lamps/taillights. The authors have tested the proposed prototype on LabView and they have provided detailed numerical results regarding the transmit speed, the bit-error ratio, the error vector magnitude, the energy per bit to noise power spectral density ratio, etc.

+ The combination of the visible light communication with the software define radio concept is novel and has not already been addressed in the literature.

+ The development of the prototype in LabView and its detailed testing show its potential to be adopted in a real networking environment.

The authors should address the following comments to improve the presentation of their manuscript.

--- The provided literature review is poor (not even a page!) and the authors do not enable the reader to understand the recent advances in the field of visible light communications, e.g., "Resource allocation in visible light communication networks: NOMA vs OFDMA transmission techniques." In International Conference on Ad-Hoc Networks and Wireless, pp. 32-46. Springer, Cham, 2016, Resource allocation in next-generation broadband wireless access networks. IGI Global, 2017. Thus, the authors should include an additional section (Section II) and improve the discussion of the related work.

--- The authors have implemented the proposed visible light communication system based on Software Defined Radio (SDR) in LabVIEW, thus it is in an emulation-based environment. It would be very interesting for the reader to have an understanding of the implementation cost and the complexity cost of the proposed framework if it was implemented in a real communication environment. Also, the authors should discuss the signaling overhead that is imposed to the devices within the proposed framework and how this overhead has implications to the devices’ power consumption.

--- Based on the previous comment, the authors should provide some additional numerical results to show the power consumption and the energy-efficiency of the end-devices that are connected in the proposed system.

Author Response

Dear Reviewer 5,

Thank you very much for all your constructive comments. Let us present our notes with explanations and adjustments. Every change/addition is marked by red color.

The authors should address the following comments to improve the presentation of their manuscript.

--- The provided literature review is poor (not even a page!) and the authors do not enable the reader to understand the recent advances in the field of visible light communications, e.g., "Resource allocation in visible light communication networks: NOMA vs OFDMA transmission techniques." In International Conference on Ad-Hoc Networks and Wireless, pp. 32-46. Springer, Cham, 2016, Resource allocation in next-generation broadband wireless access networks. IGI Global, 2017. Thus, the authors should include an additional section (Section II) and improve the discussion of the related work.

We have greatly expanded whole article. Multiple new references were added, and a number of paragraphs were added to introduction and discussion.

--- The authors have implemented the proposed visible light communication system based on Software Defined Radio (SDR) in LabVIEW, thus it is in an emulation-based environment. It would be very interesting for the reader to have an understanding of the implementation cost and the complexity cost of the proposed framework if it was implemented in a real communication environment. Also, the authors should discuss the signaling overhead that is imposed to the devices within the proposed framework and how this overhead has implications to the devices’ power consumption.

In this article, we have presented our first iteration of testing system. We are currently switching to second iteration, which is built upon much more modern OFDM. Also, a number of new parts were ordered, such as new linear amplifier and new APD photodetectors from Hamamatsu Japan. Presented system is still not final and is more suitable for testing of parts. Final product will be minimalized and tailored especially for its purpose, thus lowering manufacturing costs dramatically.

--- Based on the previous comment, the authors should provide some additional numerical results to show the power consumption and the energy-efficiency of the end-devices that are connected in the proposed system.

Based on our previous reply, we have unfortunately dismantled our first iteration in favor of more advanced OFDM system. Also, the consumption of final version will be vastly different and much lower, so it’s not necessary to monitor it right away. However, we will include power consumption in future papers, based on OFDM.

Best regards,

Mr. Rene Jaros, Mr. Lukas Danys, and Dr. Radek Martinek

 

Round 2

Reviewer 3 Report

The authors had answered all my questions. I do not have any further comments.

Reviewer 4 Report

The authors modified the work according to the reviewer's comments. The paper is more clear and results are well explained.

Hence, the paper can be considered for publication.