Autonomous Driving Assistance with Dynamic Objects Using Traffic Surveillance Cameras
Round 1
Reviewer 1 Report
Dear Authors,
After reviewing your manuscript I found it interesting. However, minor to major revision is recommended. The "core" of the study is there, however, the supporting text (especially the lack of a Discussion section) is not adequate. Please see the attached word document for more details.
Kind regards,
Reviewer
Comments for author File: 
Comments.doc
Author Response
We do appreciate of your comments. Please find out attachment.
Author Response File: Author Response.doc
Reviewer 2 Report
Dear Authors,
in the current state, this article has significant deficiencies:
(1) The introduction is a general ITS-overview with incorrect use of standard terms (e.g. RSU); the main intention of your activities is not clear
(2) What are the important ideas and results? I do not understand the main idea, setup and results. What are the basic questions to be answered?
(3) The article is lacking detail, e.g. what is the main idea for the transformation of information/data/coordinates. The basics of the transformation of geocoordinates is state of the art.
(4) The article contains a significant number of errors (spelling, incomplete sentences, ...)
Author Response
We do appreciate of your comments. Please find out attachment.
Author Response File: Author Response.doc
Reviewer 3 Report
Dear Authors,
I enjoyed reading this paper.
Still, there are some issues to deal with.
For instance:
- English language and style issues - Grammarly (https://app.grammarly.com) on default settings detected only for the text block resulting from the concatenation of Title+Abstract+Keywords 3 critical alerts (correctness issues) and 8 more advanced ones, namely: Word choice (3), Unclear sentences (2), Punctuation in compound/complex sentences (1), Passive voice misuse (1), and Faulty tense sequence (1). This meant an overall score of 84 out of 100 for this sample above. Moreover, since none of you appear to be a native English speaker, I suggest a total revision of the English language and style for the entire article using Grammarly or another specialized tool;
- The paper must follow the specific structure of the journal, namely:
Author Information, Abstract, Keywords, Introduction, Materials & Methods, Results, Discussion, Conclusions, etc., as indicated at: https://www.mdpi.com/journal/applsci/instructions. And this by changing some title names such as Overview, Evaluate the degree of coherence, etc.
- You must ensure that all figures have the required resolution (minimum 1000 pixels width/height, or a resolution of 300 dpi or higher according to the Journal’s instructions: https://www.mdpi.com/journal/applsci/instructions );
- You should provide the full specs. of the CPU and GPU used (lines 352, 353) including the model names (at least in a footnote);
- You should consider removing the error formula from the bottom of Figure 11 (only specify Err) and introduce an additional reference to a corresponding equation/formula in the main text, when referring to Figure 11 (line 388); You should apply the same to Figure 10 (e.g., H x PT);
- I think the references (in the main text) to tables and figures must be in the close proximity to them;
- There are so many figures (13) in the paper. Some of them (not essential for understanding the main content) should be moved to the Appendix section. If not existing, this section must be created;
- It seems that you provided more than a single conclusion in this paper. Therefore, the singular form (Conclusion instead of Conclusions) is not justified;
- I think more contributions in journal papers must be cited in this research both in the Introduction and Discussion sections. I believe that only 19 references, from which most are proceedings or other papers than journals (I counted only 6), are not enough;
- You should mention in the Acknowledgments section the contribution of the COCO and KODAS project partners as providers of the datasets;
Thank you for your contribution and for trying to make the world a better place!
Sincerely,
D.H.
Author Response
We do appreciate of your comments. Please find out attachment.
Author Response File: 
Author Response.pdf
Round 2
Reviewer 1 Report
Dear authors,
The paper has been appropriately revised. However, please consider adding a Discussion section where you one more address and comment on your study, previous studies, the advantages of your approach, the social and practical implications of the study, and please add additional references in this new section.
Kind regards,
Reviewer
Author Response
We have added a discussion section and explained what you mentioned. And we revised the entire manuscript for easier reading. Thank you.
Author Response File: Author Response.pdf
Reviewer 2 Report
Dear Authors,
pls. find my comments in the following text:
l.42 The use of road infrastructure for autonomous driving is for sure an advantage. However, it is not required.
l.62 Your definition of an RSU is simply wrong. Please make yourselves familiar with the terms used in the industry. The content of table 1 needs significant improvement, e.g. "Etc. ..." - I cannot imagine any relevance with respect of your research.
l.66 "ITS is a center-centered ... one-way ..." is wrong. Please read the norms on ITS and V2x-communication and refer to them correctly.
l.123 The author's have added a description of the organization of the paper (which is appreciated). However, the content does fit to the structure presented. E.g. "An overview and related works are introduced Section 2", with "2. Identification and tracking of dynamic objects" as title of sec. 2.
l. 137 "coordinate system transformation such as a Kafka server" - it is unclear, what the authors intend to communicate. A Kafka server is not a coordinate system transformation.
l. 154 Figure 1 is not consistent with the description in the text.
Section 2: the authors miss to explain the coordinate measurement / localization method. YOLO delivers a bounding box with variable size....
l.213 Figure 2 - Why is this figure included? What is the relevant information needed to understand this research? The authors include a figure of a system architecture without relevance.
l.245 How do the authors deal with (3D-) ambiguities. Identical points on the image sensor one are related to different possible 3D coordinates.
l. 308 The figure is missing detail. Where is "Err"? How is it measured?
l. 310 Figure 8 - The figure is missing Information - what are road lines? where are the critical paths?
l. 341 incomplete sentence, event data6) ???
l. 372ff - It is not clear, why the authors generally describe a Kafka-based system. Kafka is used in this study, but Kafka itself is neither developed nor improved. This part is lacking relation / relevance for the paper.
section 4.3 The authors fail to explain, why 30 FPS were needed and why 18 FPS would be enough. No relation is presented to "Autonomous Driving Assistance".
Appendix: What is the value add of this table?
Author Response
l.42 The use of road infrastructure for autonomous driving is for sure an advantage. However, it is not required.
=> We removed and revised the part of road infrastructure. We though that a road facility is helpful to safe driving.
l.62 Your definition of an RSU is simply wrong. Please make yourselves familiar with the terms used in the industry. The content of table 1 needs significant improvement,
e.g. "Etc. ..." - I cannot imagine any relevance with respect of your research.
=> We had confused in making Table 1 and made some typos as well as translation. So we removed the Table 1 and some words which are not pretty important to explain introduction.
l.66 "ITS is a center-centered ... one-way ..." is wrong.
Please read the norms on ITS and V2x-communication and refer to them correctly.
=> We revised the unclear and miss explanation about ITS and V2x with referring a previous research (new reference [1])
1.123 The author's have added a description of the organization of the paper (which is appreciated). However, the content does fit to the structure presented.
E.g. "An overview and related works are introduced Section 2", with "2. Identification and tracking of dynamic objects" as title of sec. 2.
=> We revised the description toward helping to understand readers.
- 137 "coordinate system transformation such as a Kafka server"
- it is unclear, what the authors intend to communicate.
A Kafka server is not a coordinate system transformation.
è We revised all of the description about the Kafka in overall paper. As your point, Kakfa is not a system. it was not only one of the distributed communication software, but also implemented a system on our experiment.
- 154 Figure 1 is not consistent with the description in the text.
=> We changed the Figure 1 to Figure 2 which shows an entire architecture of our system and revised it with extra explanation.
Section 2: the authors miss to explain the coordinate measurement / localization method. YOLO delivers a bounding box with variable size....
=> As your explanation, YOLO provide a 2D bounding box, we transferred 2D image(u,v) into 3D(x,y) with an assumption on the plane(z=0). We added the explanation about data processing and derived equations.
l.213 Figure 2 - Why is this figure included? What is the relevant information needed to understand this research?
The authors include a figure of a system architecture without relevance.
=> It is not pretty helpful to discussion and relevant information. So we removed it.
l.245 How do the authors deal with (3D-) ambiguities.
Identical points on the image sensor one are related to different possible 3D coordinates.
=> We used a pin-hole camera model and then described our assumption which is generally used. We added and revised them on section 3.
- 308 The figure is missing detail. Where is "Err"? How is it measured?
- 310 Figure 8 - The figure is missing Information - what are road lines?
where are the critical paths?
=> Traffic surveillance camera is not filmed on both same areas. It is not duplicated the region. So, we have an idea that it uses a virtual primary trajectory on the HD map, which is the center line of the road or the lane. The Err is measured between a virtual primary trajectory from HD map and the sample points of estimated trajectory from the image. The virtual primary trajectory is a reference line as the critical path. (Line 309 – 322)
. 341 incomplete sentence, event data6) ???
- We fixed the typos and revised it.
- 372 - It is not clear, why the authors generally describe a Kafka-based system. Kafka is used in this study, but Kafka itself is neither developed nor improved. This part is lacking relation / relevance for the paper.
=> We revised all of the description about the Kafka in overall paper. Kafka is for synchronized communication of traffic surveillance camera. We revised and removed which is not relevant information.
- section 4.3 The authors fail to explain, why 30 FPS were needed and why 18 FPS would be enough. No relation is presented to "Autonomous Driving Assistance".
=> 30 FPS is input frame rate of traffic camera. It is considered as a real time. The output detection feedback of the traffic surveillance camera is required fast response time. The quicker responded information is much helpful to a vehicle safe driving even for blind spots.
Appendix: What is the value add of this table?
=> We have the results of comparing the transformed and vertical reference points, and the distance between the two points which is an error in appendix 1.
Author Response File: Author Response.pdf
Reviewer 3 Report
Dear Authors,
You solved most of the issues found in the 1st round of review.
I think the manuscript is now ready for being published.
Congratulations!
Sincerely,
D.H.
Author Response
We revised the entire manuscript for better reading and understanding. Thank you.
Author Response File: Author Response.pdf
Round 3
Reviewer 2 Report
Dear authors,
the following items need your attentention:
general item: Language needs to be improved
L57-63: This part is supposed to describe the intention of ITSs. However, it falls short on explaining what ITSs are relevant for.
L98: The layers A, B, C should be described since they deviate from the typical description of a local dynamic map. NGII, 2019 should be properly cited.
L157: What is the meaning of this sentence?
L213: projective instead of "proejctive"? Why is lambda explained and the other parameters are not?
L279: Meaning of the sentence unclear.
L404: The "Spring framework" - citation missing
Comments for author File: Comments.pdf
Author Response
general item: Language needs to be improved
- We got a proof-reading process of this paper.
L57-63: This part is supposed to describe the intention of ITSs. However, it falls short on explaining what ITSs are relevant for.
- Generally, the map and the traffic surveillance camera are ITS enabling technologies as reference [2].
L98: The layers A, B, C should be described since they deviate from the typical description of a local dynamic map. NGII, 2019 should be properly cited.
- We add references in Table 1. A, B and C are categories of the Korea HD map from NGII.
L157: What is the meaning of this sentence?
- We revised on the introduction part of AI(image processing).
L213: projective instead of "proejctive"? Why is lambda explained and the other parameters are not?
- We revised it
L279: Meaning of the sentence unclear.
- We revised the unclear sentence.
L404: The "Spring framework" - citation missing
We add a reference. It is a famous frame for data communication.
Author Response File: Author Response.pdf
