A Ground Moving Target Detection Method for Seismic and Sound Sensor Based on Evolutionary Neural Networks

Round 1

Reviewer 1 Report

Authors propose a ground moving target detection methon based on evolutionary neural networks method by combining genetic algorithm and neural networks from the signals of seismic and sound sensors. They claim their target classification accurcacy can reach 98.21% in experiments. 

However, authors should endeavor to improve the paper organization and remove typos for readership. Such as:

1. The indent of "Keywords" below the the abstract.

2. In the last paragraph of Introduction, "Section" is commonly used instead of "Chapter"

3. The "SN" of Eq.(1) is identical to the "SN" in Eq.(3)???, since "SN is the seismic signal after normalization" in the last sentence of Section 2.

4. Then, In Eq.(4), "SN" is used again for different meaning?  The previous "SN" is not used amymore in Section 3 or 4? Authors should clarify the differences or change the naming.

5. In last 4th line before Table 3, "11 experiments" means repeating 11 tests with different classification accuracies and then their medians are recorded? If it's true, why don't choose to record their means and deviations to compare with?

6. Some misplaced comma in equation (e.g. eq.6) affect the readability.

Then, It's suggested that authors can summarize the different schemes applied in benchmark methods they compare with for readership.

#

Author Response

Dear Reviewer:

 

This letter accompanies our revised manuscript “A Ground Moving Target Detection Method for Seismic and Sound Sensor Based on Evolutionary Neural Networks ID: applsci-1836602.” We’re grateful for the reviewers’ helpful comments, and hope our revision addresses them all. Below we detail the changes in our revision. We have included the text of the comments in Roman script; our responses are in italics. References to line numbers are for the revised manuscript. The revised manuscript has been added in the attachment.

 

1.The indent of "Keywords" below the the abstract.

Response：Thank you for pointing this error out. This error has been corrected in the new version of the manuscript.

 

2.In the last paragraph of Introduction, "Section" is commonly used instead of "Chapter".

Response：Thank you for your advice, which is valuable in improving the quality of the manuscript. We have made correction according to the Reviewer’s comments.

 

3.The "SN" of Eq.(1) is identical to the "SN" in Eq.(3)???, since "SN is the seismic signal after normalization" in the last sentence of Section 2.

Response：Thank you very much for discovering this error. We apologize for this problem and have corrected (in line 97 and Eq.(3)). The "SN" in Equation 3 should be modified to "SS", indicating the normalized sound signal.

 

4.Then, In Eq.(4), "SN" is used again for different meaning?  The previous "SN" is not used amymore in Section 3 or 4? Authors should clarify the differences or change the naming.

Response：We apologize for "SN" is used again in Eq.(4). In the newly submitted manuscript, we have changed "SN" to "N" in Eq.(4).

 

5.In last 4th line before Table 3, "11 experiments" means repeating 11 tests with different classification accuracies and then their medians are recorded? If it's true, why don't choose to record their means and deviations to compare with?

Response：Since training of the three benchmark methods and the proposed method leads to very good or very poor performance with small probabilities, the mean value of the accuracy is easily influenced by the small probability extreme data. Using the median of the accuracy as the standard of comparison for each model avoids the effect of small probability extreme data.

 

6.Some misplaced comma in equation (e.g. eq.6) affect the readability.

Response：We sincerely appreciate the valuable comments. We have corrected the misplaced comma in the article.

 

7.Then, It's suggested that authors can summarize the different schemes applied in benchmark methods they compare with for readership.

Response：Thank you for your advice. We have summarized the previous studies in the Table 1(in line 71).

Author Response File: Author Response.doc

Reviewer 2 Report

I think this topic is good, but the paper has some limitations that should be enhanced.

1. I suggest authors should present the background of the work by a figure that is easy for readers to understand.

2. A summary of previous studies should be presented in a table.

3. Figure 1's caption should be revised to make it clear, what is (a), (b), and (c)?

4. Again, Figure 1 is hard to follow.

5. Please recheck and make the figure/table captions more clear

6. I suggest to public your dataset so that other researchers can use it to study and compare.

7. Citation for equations 10 and 11 should be included. In addition, the meaning of each metric should be also clear.

8. Accuracy of the proposed method in Table 3 is lower than GA-SVM, Please make it clear.

9. If it is possible, please public code.

10. Figure 5 has no new information,in my opinion.

 

 

Author Response

Dear Reviewer:

 

This letter accompanies our revised manuscript “A Ground Moving Target Detection Method for Seismic and Sound Sensor Based on Evolutionary Neural Networks ID: applsci-1836602.” We’re grateful for the reviewers’ helpful comments, and hope our revision addresses them all. Below we detail the changes in our revision. We have included the text of the comments in Roman script; our responses are in italics on a yellow background. References to line numbers are for the revised manuscript. The revised manuscript has been added in the attachment.

 

 

1.I suggest authors should present the background of the work by a figure that is easy for readers to understand.

Response：Thank you for your advice, which is valuable in improving the quality of the manuscript. In the introduction section, we have added an figure that makes it easy for the reader to understand the background of the work (in line 36).

 

2.A summary of previous studies should be presented in a table.

Response：Thank you for your advice. We have summarized the previous studies in the Table 1(in line 71).

 

3&4. Figure 1's caption should be revised to make it clear, what is (a), (b), and (c)? Again, Figure 1 is hard to follow.

Response：Thank you very much for discovering this error. We have revised the title of Figure 1 and added subheadings for (a), (b), and (c)(in line 133). We also carefully revised the text describing Figure 1 in detail.（in line 119 to 131).

 

5.Please recheck and make the figure/table captions more clear

Response：We apologize for not clearly describing the tables and figure in the manuscript. We have checked the descriptions of the figure and tables in the full text and made changes accordingly.

 

6.I suggest to public your dataset so that other researchers can use it to study and compare.

Response：Thank you for your valuable suggestions. According to your suggestions, we will upload the dataset to MDPI's database.

 

7.Citation for equations 10 and 11 should be included. In addition, the meaning of each metric should be also clear.

Response：Thank you for your valuable and thoughtful comments. We have made explicit the each metric in each formula.(in line 291 to 293)

 

8.Accuracy of the proposed method in Table 3 is lower than GA-SVM, Please make it clear.

Response：We apologize for not clearly describing the advancement of the proposed method compared to the GA-SVM method in this manuscript. We have appended notes to the revised manuscript.(in line 388 to 390)

GA-SVM is slightly higher than the proposed method in this paper in terms of classification accuracy, but the performance time of the GA-SVM method is five times higher than the proposed method in this paper. This shows that the proposed method has the advantage of low complexity compared to GA-SVM and is more suitable for deployment on low-cost hardware.

 

9.If it is possible, please public code.

Response：We sincerely appreciate the valuable comments. In order to let other scholars better refer to our work, we are finishing the code of this manuscript and adding comments, and we will upload the code to github after finishing the work.

 

 

10.Figure 5 has no new information,in my opinion.

Response：Thank you for your suggestion, we have removed Figure 5 in the revised manuscript.

Author Response File: Author Response.doc

Round 2

Reviewer 1 Report

Authors need to endeavour correcting the text editing in final version. Such as:

1. mismatched italic style in line 238 and 239 with eq.(8).

2.In Table 4,  different "s" and "S" to represent "second"?

3.The black numbers in confusion matrix of Figure.8 are not clear to see.

#

Author Response

Thank you for your careful review of the manuscript and for your professional comments. We have carefully revised the manuscript according to your comments, and hope our revision addresses them all. We have included the text of the comments in Roman script; our responses are in italics. 
1.mismatched italic style in line 238 and 239 with eq.(8).
Thank you for your valuable comments. Is it that the italic style of the symbols in lines 241 and 242 do not match equation (8)？We have revised the style of the symbols from line 233 to line 242.

2.In Table 4,  different "s" and "S" to represent "second"?
Thank you for your advice. We have replaced the "s" and "S"with "second" in Table 4

3.The black numbers in confusion matrix of Figure.8 are not clear to see.
Thank you for your advice. We have revised Figure 8.

Reviewer 2 Report

All my comments were addressed. 

Author Response

We feel great thanks for your professional review work on our manuscript.