Hardware Suitability of Complex Natural Resonances Extraction Algorithms in Backscattered Radar Signals
Round 1
Reviewer 1 Report
This paper studies the suitability of complex natural resonances extraction algorithms in the viewpoint of hardware implementation. The following concerns need to be addressed.
1. In Section 1, the organization of the paper is better to be in a separate paragraph.
2. Since the algorithms are known, the presented results and the drawn conclusion can be rather easily inferred from the preceding works in the literature. The novelty needs to be highlighted further.
3. There are only few recent works in the list of references. In addition, most of the compared works are too old. The literature review must be enhanced.
4. The exact values associated with the Figs. 18–20 can be given.
5. In Table 1, only one value was considered for the sphere unlike the others.
Extensive editing of English language required
Author Response
Dear reviewer,
Thank you very much for providing insight to our work. We have addressed your concerns as follows:
- The paragraph in the introduction has been separated as indicated.
- The exact values have been associated with figures 18 - 20.
- There was one object of each kind in table 1, and the backscattered response was measured from different angles and distances.
- English language has been reviewed.
- Modifications to the abstract, introduction, and all the sections including the conclusions have been made. These modifications highlight the utilization of these metrics as a tool to discern and choose a hardware platform that is more aligned with the specific requirements of the ultimate application. Also, the term “computational cost” has been changed to execution time as a performance metric.
- The literature has been reviewed as well.
- The novelty of the paper was enhanced in the introduction section.
Reviewer 2 Report
An interesting paper with promising research findings.
The authors have presented an appreciable qualitative research findings on the abstract and the conclusion. However, the quantitative performance metrics are lacking. The values of the "accuracy, number of samples used and running-time computational cost" should be stated in the abstract as quantitative performance metrics / research findings.
A good mathematical modelling. Define all acronyms at first mention e.g., SVD. Correct the typos / grammatical errors.
The literature review is compelling.
What is the estimating relationship of a hardware suitability factor (HSF)?
What is the efficiency of each method respecting their respective HSF?
Good English Language.
Author Response
Dear reviewer,
Thank you very much for providing insight to our work. We have addressed the corrections
as follows:
- “Quantitative performance metrics” has been included in the abstract.
- Acronyms have been defined and the list of abbreviations improved.
- A Hardware Suitability Factor has been described at the end of section 3 according to the results presented.
- Modifications to the abstract, introduction, and all the sections including the conclusions have been made. These modifications highlight the utilization of these metrics as a tool to discern and choose a hardware platform that is more aligned with the specific requirements of the ultimate application. Also, the term “computational cost” has been changed to execution time as a performance metric.
Reviewer 3 Report
Globally, the manuscript is very well written and organized. However, there are some issues and questions that must be addressed.
English needs a complete revision; please refer to the attached commented PDF document where some of the needed corrections are highlighted.
Define every acronym/abbreviation; although there is a list of abbreviations at the end of the manuscript, not all the used ones are included here.
Be consistent when referring to figures and tables; always use “Fig. x” or “fig. x” and do not use sometimes one and other times the other (and even other combinations; once again, refer to the attached commented document). Correct the reference to figures in line 130.
You should use a consistent terminology; e.g., you should always use “Late-Time Response”, and not sometimes “Late-Time response”.
Throughout the manuscript you refer to “computational cost”, but only in section 3 it becomes clear that you are referring only to it “in terms of execution time”, but it could be in terms of other parameters (i.e., complexity, memory, ...), as you acknowledge in the conclusions section. I recommend that you should make clear this fact throughout the whole document (and rewrite the needed sentences).
In the conclusions section you state “(…) with the intention of having a better insight when selecting the most appropriate method depending on the hardware, time and accuracy desired for a particular study.”, but you have done all the tests using the same hardware! Unless you present test results performed with different hardware configurations, this statement has to be corrected, as well as the title, the abstract and all references throughout the manuscript.
Comments for author File: 
Comments.pdf
Please refer to the comments above.
Author Response
Dear reviewer,
Thank you very much for providing reading our work. We have addressed the corrections as follows:
- We have used your attached document to correct English grammar, fixed the abbreviations, and used a consistent terminology.
- Modifications to the abstract, introduction, and all the sections including the conclusions have been made. These modifications highlight the utilization of these metrics as a tool to discern and choose a hardware platform that is more aligned with the specific requirements of the ultimate application. Also, the term “computational cost” has been changed to execution time as a performance metric.
- A Hardware Suitability Factor has been described at the end of section 3 according to the results presented.
- The literature has been reviewed as well.
- The novelty of the paper was enhanced in the introduction section.
Round 2
Reviewer 1 Report
Several concerns have been addressed except the following. Even after a further literature review, still only PMP, Prony, Cauchy, and VCM are compared in the relevant figures.
Extensive editing of English language required
Author Response
Dear referee,
Thank you again for your valuable input on our paper. We have performed an extensive English check.
The reason for comparing only MPM, Prony, Cauchy, and VCM methods is that currently, those are the only methods used for approximating CNR calculations from a backscattered signal based on the Singularity Expansion Method. There are more complex numerical solutions like the ones described by Dr. Giri and Dr. Tesche. However, the relative difficulty of obtaining this solution only for straight wires makes the short-time implementation in experimental scenarios untractable.
Best regards,
Reviewer 3 Report
All my previous concerns and questions have been addressed.
Author Response
Dear referee,
Thank you again for your valuable input on our paper. We have performed an overall check of the article.
Best regards,
Round 3
Reviewer 1 Report
The remaining concern has been addressed.
As selected above
