A Comprehensive and Effective Framework for Traffic Congestion Problem Based on the Integration of IoT and Data Analytics

Round 1

Reviewer 1 Report

The authors propose a framework to avoid traffic congestion based on data obtained with IoT and data analytics.

In chapter 3 the authors are suggesting an architecture to support their idea, however, there is a lack of information about how each layer of the proposed architecture should be implemented, for example, communication protocol, how data will be acquired, and which data should be acquired. There is also no discussion about time to acquire data, decision taken, and traffic information returned to users. No discussion about the technologies used on the cloud or the fog. The deepest discussion presented is the algorithm to filter data from social networks, but there is no discussion on how trustable that information can be.

In section 4, the most of time the authors discuss ways to collect data, but if this section is about a functional model, is expected precise information of how the data was acquired to this functional model.

In experiment and results section, there is no clear information about the simulator used to simulate the traffic behaviour in order to make a fair analysis.

Was not clear on the experiments what the numbers 1-20 of axis x on the graphics represented.

The authors should focus in showing how things were done instead of discussing how things could be done to achieve the goal.

 

Author Response

The authors propose a framework to avoid traffic congestion based on data obtained with IoT and data analytics.

In chapter 3 the authors are suggesting an architecture to support their idea, however, there is a lack of information about how each layer of the proposed architecture should be implemented, for example, communication protocol, how data will be acquired, and which data should be acquired. There is also no discussion about time to acquire data, decision taken, and traffic information returned to users. No discussion about the technologies used on the cloud or the fog. The deepest discussion presented is the algorithm to filter data from social networks, but there is no discussion on how trustable that information can be.

 

==> This paragraph has been answered and the relevant paragraphs (section 4.5) have been updated in detail

In section 4, the most of time the authors discuss ways to collect data, but if this section is about a functional model, is expected precise information of how the data was acquired to this functional model.

==> This paragraph was answered in the same previous section (section 4.5) and it was clarified how to accurately feed the functional model with data and its sources. Also, within the results and experiments section, the methods of obtaining data and the applications that we made to obtain and participate in decision-making were elaborated.

In experiment and results section, there is no clear information about the simulator used to simulate the traffic behaviour in order to make a fair analysis.

==> This paragraph has been answered in Section 5.1 in detail, where the use of MATLAB in the simulation process has been determined. The simulation method was also described in some detail

==> The scheme has been modified and the deficiencies are clearly shown.

Authors should focus on showing how to get things done rather than discussing how to do things to achieve the goal.

==> Many paragrapghs have been reformulated in order to clearly answer this paragraph.

Reviewer 2 Report

In the present manuscript, the authors propose a comprehensive framework for a reliable and flexible solution for traffic congestion problem based on Fog computing. However, I will comment on some aspects to improve the quality of the article:
-The authors write the acronyms incorrectly. The correct form to write them is as in line 23. This error must be corrected throughout the document with all the acronyms used.
-There are acronyms that the authors have not written their meaning such as "RFID".
-When writing the article's own words such as "Figure", "Section", "Algorithm", "Equation", "Table", these must be with the first capital letter.
-Authors must be added in the related works, studies about use Induction Loops.
-I suggest adding these items to improve the quality of the item:
- Zambrano-Martinez, J. L., Calafate, C. T., Soler, D., Lemus-Zúñiga, L. G., Cano, J. C., Manzoni, P., & Gayraud, T. (2019). A centralized route-management solution for autonomous vehicles in urban areas. Electronics, 8 (7), 722.
- Nafi, N.S .; Khan, R.H .; Khan, J.Y .; Gregory, M. A predictive road traffic management system based on vehicular ad-hoc network. In Proceedings of the 2014 Australasian Telecommunication Networks and Applications Conference (ATNAC), Southbank, Australia, 26–28 November 2017; pp. 135-140.
-The "Listing" must be replaced by Algorithms. These algorithms must have the same format as the manuscript.
-There are equations that are not cited or numbered.
-Why from Table 2 to 7 do these Tables contain no data?
-The authors do not present which scenario has been chosen for the experiments. Some map of the scene must be added.
-Where have the data been obtained for the experiments?
-What are the Origin-Destination matrices?
-What is the reason to use Google Maps API and not to use OpenStreetMaps?
-Has the proposal presented by the authors been verified in any simulator?
-Who feeds information about the street status? If it is the User, what happens if the information is missing from the system?
-The conclusions must be improved.

Author Response

In the present manuscript, the authors propose a comprehensive framework for a reliable and flexible solution for traffic congestion problems based on Fog computing. However, I will comment on some aspects to improve the quality of the article:
-The authors write the acronyms incorrectly. The correct form to write them is as in line 23. This error must be corrected throughout the document with all the acronyms used.
-There are acronyms that the authors have not written their meaning such as "RFID".

==> article checked and all the acronyms are defined
-When writing the article's own words such as "Figure", "Section", "Algorithm", "Equation", "Table", these must be with the first capital letter.
==> All those words are corrected and written probably 
-Authors must be added in the related works, studies about use Induction Loops.
-I suggest adding these items to improve the quality of the item:
- Zambrano-Martinez, J. L., Calafate, C. T., Soler, D., Lemus-Zúñiga, L. G., Cano, J. C., Manzoni, P., & Gayraud, T. (2019). A centralized route-management solution for autonomous vehicles in urban areas. Electronics, 8 (7), 722.
- Nafi, N.S .; Khan, R.H .; Khan, J.Y .; Gregory, M. A predictive road traffic management system based on vehicular ad-hoc network. In Proceedings of the 2014 Australasian Telecommunication Networks and Applications Conference (ATNAC), Southbank, Australia, 26–28 November 2017; pp. 135-140.

===> the mentioned references are taken into account and cited inside the article under numbers {9 and 10 }

-The "Listing" must be replaced by Algorithms. These algorithms must have the same format as the manuscript.

==> the remarks are achieved and the replacement is done.

-There are equations that are not cited or numbered.

==> all the equations have numbered 

-Why from Table 2 to 7 do these Tables contain no data?

==>the idea is clarified in figure number 6 (Schema for the data received by cloud from different sources).

-The authors do not present which scenario has been chosen for the experiments. Some map of the scene must be added.

==> this is answered starting from line 231.

-Where have the data been obtained for the experiments?
-What are the Origin-Destination matrices?
-What is the reason to use Google Maps API and not to use OpenStreetMaps?
-Has the proposal presented by the authors been verified in any simulator?
-Who feeds information about the street status? If it is the User, what happens if the information is missing from the system?

==> this is answered in 4.5.2. Second level – main streets. where we create our Data set using the techniques and tools (like Cameras, drones, Supported apps. and from another hand, we used the same data set of ITMS (line 652).

-The conclusions must be improved.
==> the conclusion is changed.

Reviewer 3 Report

This manuscript claims to propose a comprehensive framework for the traffic congestion problem using cloud, fog and IoT layers. The proposal is valid and has good potential but I didn’t find it comprehensive as claimed. In fact, the contributions of the manuscript are rather limited compared to previous related-works (some of which are mentioned in section 2). Following are my comments on the this manuscript:

• Figures 1 and 2 can be somehow combined
• Listing 2 is introduced before listing 1
• Figures should be explained in the caption. More details are needed especially for the figures in the ‘Results’ section.
• Section 2 is very good. It provides sufficient background on the state-of-the-art. In Section 2, several previous related-works were discussed through pros and cons. Some drawbacks were mentioned, e.g. inaccuracy du to lighting, random movement of vehicles, etc, the cost of infrastructure, they do not work well in extreme weather conditions, privacy issues, and it was argued that none of the previous works had solved these issues using one framework. Thus, the authors would solve this in their comprehensive proposal. However, the presented experiments evaluated the relation between the number of cars on an intersection with the number of cars served (E1), average waiting time (E2 and E3), while the point of E4 is not clear to me but it obviously does not cover any of the mentioned drawbacks.
• First paragraph of section 5 >>rephrase. More technical details on the methods used for evaluation, and the parameterization of the tools, need to be presented.
• The first paragraph of section 5.1 is very important but it is not clear to me. More details are needed regarding the system authors are comparing with. Why is it different? Why is it suitable for comparison and evaluation? Why isn’t it discussed in Section 2?
• Pp8: more than one options??? is be sent???...Pp15: the proposed algorithm services a much larger number of ..>> serves more ..
• After long and sufficient description of the available options for level-2 (section4.1) the authors have not define their utilized method and its parameterization!
• Legends and axes titles are needed for figure 6. More discussion is needed regarding this figure too.
• Although I am not specializing in ML, it is clear that section 5.2 does not provide any technical details regarding the two algorithms claimed to be proposed. Only mentioning that two algorithms are developed and used is not sufficient. The algorithms need to detailed, discussed, evaluated against similar ones, and the limitations need to be highlighted. I haven’t found any of this. Be careful, image processing is a whole standing field in Computer Science where you need to be much accurate and scientific in your statements.
• Generally, I didn’t find the described material and methods sufficient.

 

Generally, the language of the manuscript is weak and poor and includes a lot of grammatical mistakes. Professional editing is highly recommended.

Author Response

This manuscript claims to propose a comprehensive framework for the traffic congestion problem using cloud, fog and IoT layers. The proposal is valid and has good potential but I didn’t find it comprehensive as claimed. In fact, the contributions of the manuscript are rather limited compared to previous related-works (some of which are mentioned in section 2). Following are my comments on the this manuscript:

Figures 1 and 2 can be somehow combined

==> All thanks to the reviewer for this note, but, if the two figures are combined together, as the reviewer said, the resulting Figure will be very crowded and cannot be easily understood.

Listing 2 is introduced before listing 1
==> the mistake is fixed

Figures should be explained in the caption. More details are needed especially for the figures in the ‘Results’ section.

==>all the figures are explained and captions are redefined

Section 2 is very good. It provides sufficient background on the state-of-the-art. In Section 2, several previous related-works were discussed through pros and cons. Some drawbacks were mentioned, e.g. inaccuracy du to lighting, random movement of vehicles, etc, the cost of infrastructure, they do not work well in extreme weather conditions, privacy issues, and it was argued that none of the previous works had solved these issues using one framework. Thus, the authors would solve this in their comprehensive proposal. However, the presented experiments evaluated the relation between the number of cars on an intersection with the number of cars served (E1), average waiting time (E2 and E3), while the point of E4 is not clear to me but it obviously does not cover any of the mentioned drawbacks.
===>
The difference was clarified in the text by paragraph 5, and it can be said in brief that the proposed solution takes into account the dynamics in distributing times between crowded streets and does not say to solve the problem of street congestion by causing congestion in another street. This is explained in more detail in paragraph 5. Experiment and results

First paragraph of section 5 >>rephrase. More technical details on the methods used for evaluation, and the parameterization of the tools, need to be presented.

The first paragraph of section 5.1 is very important but it is not clear to me. More details are needed regarding the system authors are comparing with. Why is it different? Why is it suitable for comparison and evaluation? Why isn’t it discussed in Section 2?
Pp8: more than one options??? is be sent???...Pp15: the proposed algorithm services a much larger number of ..>> serves more ..
After long and sufficient description of the available options for level-2 (section4.1) the authors have not define their utilized method and its parameterization!

==>The Experiment and Results section has been completely rewritten so that the work is better placed. All techniques used in the project were clearly demonstrated

Legends and axes titles are needed for figure 6. More discussion is needed regarding this figure too.
==> this is taken into account in all the figures 

Although I am not specializing in ML, it is clear that section 5.2 does not provide any technical details regarding the two algorithms claimed to be proposed. Only mentioning that two algorithms are developed and used is not sufficient. The algorithms need to detailed, discussed, evaluated against similar ones, and the limitations need to be highlighted. I haven’t found any of this. Be careful, image processing is a whole standing field in Computer Science where you need to be much accurate and scientific in your statements.
Generally, I didn’t find the described material and methods sufficient.
 
==> This point is made clear in paragraph 4.3. Level 3 - Analysis of historical data All algorithms used for machine learning have been reformulated and audited (see paragraphs 4.3.)

Generally, the language of the manuscript is weak and poor and includes a lot of grammatical mistakes. Professional editing is highly recommended.
==>The entire article has been revised and grammatically revised.

Round 2

Reviewer 2 Report

Thanks to the authors for performing the suggested changes. However, I have observed that they have not fixed the errors in the acronyms, so authors must correct them before publication. In addition, a minimal spell check of the language must be performed prior to publication. Another aspect for authors when submit an article in the future, authors must improve the quality of how to present a Cover Letter, since the quality of the research and the researchers is deduced.

Author Response

Thank you for your feedback and recommendation.

We have corrected all the language mistakes in the revised manuscript.

Reviewer 3 Report

I have completely re-reviewed this manuscript as some major issues appeared within the first submission with which I couldn’t comprehensively perform the review. Unfortunately, I still have the following comments about your second submission:

• The figures are still of a very low quality. The positioning, the captions, the resolution, the discussions, the harmonization with each other, etc. More professional approach for depicting the ideas of the figures need to be carried out. I also wonder if the coloring of figures components (e.g. Figs 1 and 2) has a meaning.
• I didn’t find the language of this version better than the previous one. I would, again, recommend a “professional editing” of the manuscript which will drastically enhance the understanding of what is being proposed/stated.
• I don’t get the relation between the proposed solution and solutions for energy harvesting in Section 2.
• According to Figure 8 and subsection 5.1.1, the proposed solution does not outperform ITMS in all cases. Furthermore, a part of the "very short" discussion of the results is clearly mistaken. That is, the performance of the experimented ITMS and the proposed solution DOES NOT tend to be close to the performance of the traditional system if there is no congestion! Frankly, it is the exact opposite!
• I wonder how the proposed framework is different from the framework proposed in [1]. It is also stated in this manuscript that “The accuracy of processing images captured by drones has been already established in our previous work” which is [1]. However, I checked this previous work and I haven’t found any accuracy evaluation!
• In your response, when you mentioned ‘paragraph 5’ did you mean ‘section 5’? I suppose this is the only explanation. So first, we need to agree that several previous works have been proposed to solve your manuscript’s research problem (as described in Section 2). Then, the reader needs to know WHY those works are not ‘good enough’, if you may say (which is also discussed in Section 2). After that, the reader needs to know in which terms, specifically, is the proposed solution better than the previous ones (stated in the last paragraph of Section 2 and first paragraph of Section 5 by mentioning that the proposed solution is comprehensive and solves ‘all’ of these issues). Finally, the research methods, experiments and results MUST prove that the proposed solution does indeed outperform the previous solutions and addresses the discussed challenges. The last point is very important for any scientific research to be ‘sound’.

I have briefly pointed this out in my previous review, yet it was not solved in the second version, thus I had to clarify my point here again. Now let’s compare if the challenges and outperformances, stated in Section 2 were proven in Section 5:

1. Inaccuracy: evaluated for the classification algorithms of the tweets but not for image processing.
2. Number of serviced vehicles and avg waiting time: evaluated but was not mentioned as a challenge in section 2. In fact, I have referred to the manuscript of the solution with which the comparison is conducted (i.e. ITMS [2]) and I found that the used values in Figure 10 do not comply with the values presented in the ITMS paper (i.e. Figure 14.b in pp: 17 of the ITMS paper).
3. Cost of new infrastructure: not evaluated
4. Working ‘well’ in extreme weather conditions: not evaluated (P.S: ‘well’ is too fuzzy and unprofessional)
5. Safety: not evaluated
6. Security and Privacy: not evaluated (P.S: note that relying on Twitter and Tweets imposes privacy and intellectual property issues!)

 

[1]: Alharbi, A., Halikias, G., Sen, A. A. A., & Yamin, M. (2021). A framework for dynamic smart traffic light management system. International Journal of Information Technology, 13(5), 1769-1776.

[2]: https://downloads.hindawi.com/journals/jat/2021/4037533.pdf

Author Response

Thank you for your detailed insight and feedback. 

Please see the attached PDF file for the detailed response.

Author Response File: Author Response.pdf

Round 3

Reviewer 3 Report

Thank you for your clarifications. I think the manuscript is now much better and focused.

Author Response

Response to Reviewer-3

Thank you for your recommendation.

 

Response to Editor

Thank you for your feedback. Please find below detailed response to your queries.

 

1. Add in the abstract the explicit aim of the proposed study and its improvements.

We have added the following in the Abstract to explicitly state the aim of the study.

 

“Therefore, this paper proposes a comprehensive framework to achieve a reliable, flexible and efficient solution for traffic congestion problem.”

 

2. Where is the novelty of the work compared to the related works?

We have added the following in the Abstract to state the novelty of the study.

 

“Unfortunately, no such framework that addresses the reliability, flexibility, and efficiency issues of smart-traffic management exists.”

 

3. What are you improving or doing better/different?.

We have added the following in the Abstract to state the contribution of the study.

 

“We propose a smart-traffic light algorithm at Level-1 for efficient management of congestion at intersections, tweet-classification and image-processing algorithms at Level-2 for reliable and accurate decision making, and support services at Level-4 of the functional model.”

 

4. According to Figure 8 and subsection 5.1.1, the proposed solution does not outperform ITMS in all cases.

Absolutely correct, and the manuscript never claims that the proposed solution outperforms ITMS in context of the number of vehicles serviced as shown in Figure 8.

 

5. The following papers should be considered.

The suggested works have been included.

 

6. Minor comment: Analysis about scalability features of the approach could be added to further improve the strength of the paper.

We agree that such an addition will immensely strengthen the paper. However, scalability analysis of the framework is a complete study in itself, and therefore, this article would not be able to justice with that in short discussion.