An Advanced Artificial Fish School Algorithm to Update Decision Tree for NLOS Acoustic Localization Signal Identification with the Dual-Receiving Method

Round 1

Reviewer 1 Report

This paper proposes an NLOS mitigation method for improving indoor localization results. In addition, they proved it in a TDOA localization system that reduces its error through posterior signal processing and a supervised learning method to detect NLOS signal links.

It is a remarkable and complete paper that is well-written, well-structured, and presents novelty in the field.

Some minor issues are:

1- Please adjust the style of your equations.

2- Improve the quality of some figures such as Figure 1.

3- Please make a discussion section comparing your results with those of the literature.

With these changes, the manuscript can be accepted.

Author Response

Dear professors,

The revisions have been completed according to the opinions of all the reviewers. It is quite convincing that each opinion is of great value, which improves the quality of this manuscript a lot. We are very grateful for this. According to these opinions, all sections have been improved, some codes have been rewritten, and some experiments have been redone, so as to complete the modification of this manuscript. Compared with this manuscript's version1 in MDPI System, our revisions refer to (Version2 Manuscript has been uploaded)

Point 1: “Please adjust the style of your equations.”

Response 1: Many thanks for your valuable comments. In fact, it is our mistake that we use the wrong font sizes, and now we have adjusted them compared with the rules and templates of MDPI. Although in Word 2016, these "Track Changes" functions cannot be noticed in the Manuscript Version 2, we have both finished this part indeed.

Point 2:” Improve the quality of some figures such as Figure 1.”

Response 2: Many thanks for your valuable comments. We attach great importance to the quality and resolution of these figures in this manuscript. So, Figure 1, Figure 2, Figure 5, Figure 6, Figure 7, and Figure 8 have been rebuilt. All of their formats have been adjusted to TIFF format without compression. Before that, necessary modifications have been made in the “regedit” of the Windows system. The physical figures of some equipment can also be provided in the compressed package later when necessary. At the same time, the figures (.pptx) in editable status are also added to the compressed package and can be both uploaded to the MDPI system.

Point 3:” Please make a discussion section comparing your results with those of the literature.”

Response 3: Many thanks for your valuable comments. These parts are added both in the Abstract and in the Conclusion part. Both parts can be found in Manuscript Version 2. Meanwhile, our novel AFSA optimization method can obtain a better effect and take less time compared with the original version or other methods. There is no doubt that our method can balance the relationship between accuracy and running time in flexible NLOS situations. This part has been added from the line 581 to the line 616 in Manuscript Version 2. During this part, many recent AFSA optimization methods from different literature are used to be compared with our methods. The same data set and the same classifier SVM are used to illustrate the advantages of our methods.

  Furthermore, the authors' second institution name has also been modified due to the sudden change of its English name officially. Although this institution does not change its Chinese official name, it adjusts the English official name indeed [1]. The actual institutional relationship and the order of the authors have not been changed.

Meanwhile, there is something special happened with our teammate last month, so we have to add a new Funding Resource from the author Ruixiang Kan to make sure to have enough ability to pay for APC. So please allow us to add this new Funding Resource (Innovation Project of GUET Graduate Education No. 2023YCXB05) to this manuscript.

Some other words and sentences are adjusted and several new reference papers are added when necessary.

References

[1]URL: https://cise.glut.edu.cn/

Author Response File: Author Response.docx

Reviewer 2 Report

Authors propose a method to update decision tree for NLOS scenario for the dual-receiving signal processing method. For this purpose, they use advanced artificial fish school algorithm (AFSA). The paper is well written and there are interesting results. However, there are some important issues to be resolved.

- Authors should give more references about AFSA. Although the title of the paper includes AFSA, I could not see a literature review about this algorithm. In my humble opinion, the following papers are milestones in this area.

1. X. Li, Z. Shao and J. Qian, “An optimizing method based on autonomous animats: fish-swarm algorithm （in Chinese）,” Systems Engineering Theory and Practice. 22(11), pp.32-38, 2002.

 2. J. Hu, X. Zeng and J. Xiao, "Artificial Fish School Algorithm for Function Optimization," 2010 2nd International Conference on Information Engineering and Computer Science, Wuhan, China, 2010, pp. 1-4, doi: 10.1109/ICIECS.2010.5678350.

- In the Abstract, authors say "... the system is matched with the homologous NLOS identification method based on a novel artificial fish swarm algorithm (AFSA) and the Decision Tree model". Is it swarm or school?

- What are the advantages of advanced AFSA as compared to AFSA. Authors should emphasize that clearly in the Abstract and conclusion.

- They said that there are only three coordinates among 12 which have larger positioning errors. What are the reasons? 

- If possible, authors are suggested to compare their results with the similar works in the literature.

Author Response

Dear professors,

The revisions have been completed according to the opinions of all the reviewers. It is quite convincing that each opinion is of great value, which improves the quality of this manuscript a lot. We are very grateful for this. According to these opinions, all sections have been improved, some codes have been rewritten, and some experiments have been redone, so as to complete the modification of this manuscript. Compared with this manuscript's version1 in MDPI System, our revisions refer to (Version2 Manuscript has been uploaded)

Point 1: About two milestones paper sited in this manuscript（Authors should give more references about AFSA. Although the title of the paper includes AFSA, I could not see a literature review about this algorithm. In my humble opinion, the following papers are milestones in this area.）

Response 1: Many thanks for your useful comments. In fact, these two papers are really very useful and play an important role in AFSA optimization method in the history. It is not proper that only reference 34 is cited in the original manuscript as the milestones paper. Now the revision has been completed. These can be seen from the line 401 to the line 402. Then, the following part in Manuscript Version 2 will continue to illustrate the updating strategies on the methods according to some basic rules from these two papers.

Point 2: About “Is it swarm or school?”（ In the Abstract, authors say "... the system is matched with the homologous NLOS identification method based on a novel artificial fish swarm algorithm (AFSA) and the Decision Tree model". Is it swarm or school?）

Response 2: Many thanks for your useful comments. It should be "School" and this part has been revised in abstract part in Manuscript Version 2. In fact, this is an error that should have been avoided indeed. Meanwhile, some necessary adjustments on the serial numbers in references part have also been done.

In addition, as for the actual name of “AFSA”, it is tough to determine what to use. The author himself (Dr. Li Xiao-lei from Zhejiang University, 2003) has given two different ways of writing on different occasions. One is Artificial Fish School Algorithm, this combination is what we adopt. We believe that the word “School” can describe the concept of collectivization and the meaning of group actions. This comes from Reference 34 in Manuscript Version 2. But to our great surprise, the other one is the Artificial Fish Swarm Algorithm in Reference 36 in Manuscript Version 2. Both combinations exist indeed and they have the same Chinese name (written as “人工鱼群算法” in Chinese). Finally, according to our comprehension and methods, the first combination is used by us.

Point 3:About “the author should emphasize the advantages of Advanced AFSA ”（ What are the advantages of advanced AFSA as compared to AFSA. Authors should emphasize that clearly in the Abstract and conclusion.）

Response 3: Many thanks for your useful comments. These parts are added both in the Abstract and in the Conclusion part. Our novel AFSA optimization method can obtain a better effect and take less time compared with the original version. There is no doubt that our method can balance the relationship between accuracy and running time in flexible NLOS situations. This part has been added from the line 531 to the line 542 in Manuscript Version 2. During this part, many recent AFSA optimization methods are used to be compared with our methods. In Section 3.3 in Manuscript Version 2, the same data set and the same classifier SVM are used to illustrate the advantages of our methods.

Point 4:About “the reason why only three coordinates among 12 which have larger positioning errors”（ They said that there are only three coordinates among 12 which have larger positioning errors. What are the reasons?）

Response 4: Many thanks for your useful comments. In fact, there are few differences among these coordinates. At the same time, some accidental factors may also need to be considered during our experiments, such as indoor temperature, the unstable effect of piezoelectric tweeters, etc. They may bring us negative factors. It is quite certain that selecting a higher quality omnidirectional facility as an essential part of the acoustic anchor will greatly reduce the contingency and uncertainty of our experiments. With more repeated experiments, these results may also be slightly different later on. This part can be found out from the line 849 to the line 855 in Manuscript Version 2.

Point 5:About “compare their results with the similar works in the literature.”（ If possible, authors are suggested to compare their results with the similar works in the literature.）

Response 5: Many thanks for your valuable comments. This part is added in Manuscript Version 2 to show that our novel AFSA optimization method can obtain a better effect and take less time compared with the original version or other methods. There is no doubt that our method can balance the relationship between accuracy and running time in flexible NLOS situations. This part has been added from the line 581 to the line 616 in Manuscript Version 2. This part focuses more on the optimization effect on the classifier when the same data set and the same application situations are guaranteed. During this part, many recent AFSA optimization methods from different literature are used to be compared with our methods. Meanwhile, our optimization method is also supposed to be connected with the real-world situations when deploying. So, some necessary modifications are needed before they are widely used. At that moment, some essential works should be finished for the facilities in the system.

Furthermore, the authors' second institution name has also been modified due to the sudden change of its English name officially. Although this institution does not change its Chinese official name, it adjusts the English official name indeed [1]. The actual institutional relationship and the order of the authors have not been changed.

Meanwhile, there is something special happened with our teammate last month, so we have to add a new Funding Resource from the author Ruixiang Kan to make sure to have enough ability to pay for APC. So please allow us to add this new Funding Resource (Innovation Project of GUET Graduate Education No. 2023YCXB05) to this manuscript.

Some other words and sentences are adjusted and several new reference papers are added when necessary.

References

[1]URL: https://cise.glut.edu.cn/

Author Response File: Author Response.docx

Reviewer 3 Report

The manuscript entitled “An Advanced Artificial Fish School Algorithm to Update Decision Tree for NLOS Acoustic Localization Signals Identification of Dual-Receiving Method” is written in an organized manner and the authors need to revise a little bit for further proceeding. Minor revision of the manuscript is required. The following suggestions may be included in the revision process.

1.      Give a brief introduction for decision tree.

2.      Elaborate CTGSA

3.       It is difficult to read the reference every single time for readers, because the authors stated references without giving a clear outline (line nos. 123,125,246,251,253, etc)

Conclusively, I recommend this paper for further proceeding and for peer review.

Author Response

Dear professors,

The revisions have been completed according to the opinions of all the reviewers. It is quite convincing that each opinion is of great value, which improves the quality of this manuscript a lot. We are very grateful for this. According to these opinions, all sections have been improved, some codes have been rewritten, and some experiments have been redone, so as to complete the modification of this manuscript. Compared with this manuscript's version1 in MDPI System, our revisions refer to (Version2 Manuscript has been uploaded)

Point 1:” Give a brief introduction for decision tree.”

Response 1: Many thanks for your useful comments. Decision Tree model is of great importance in our system. Our updating strategies are used to boost its performance and benefit the indoor localization system. According the background and its standpoints, a clear and brief introduction is necessary. So we add this part on brief introduction for DT model in our research situations. The relationship between its process and our applications are vividly described. This part can be vividly seen from the line 310 to the line 316 in Manuscript Version 2.

Point 2:“Elaborate CTGSA”

Response 2: Many thanks for your useful comments. CTGSA is introduced to upgrade the effect of the original AFSA. There are two parts in CTGSA, one is CT part, and the other is GSA part. CT part is used to energize the initialization process when the samples are not enough and it will also improve the process later on. In the process of training, aiming at the weakness of AFSA itself, introducing GSA into the main process of AFSA will make the training process faster and more efficient. This method can be helpful not only in AFSA, but in other algorithms as well. Both parts are essential and helping us to boost the methods in real-world situations. This part can be found from the line 531 to the line 542 in Manuscript Version 2.

Point 3:“clear outline before the key part”（ It is difficult to read the reference every single time for readers, because the authors stated references without giving a clear outline (line nos. 123,125,246,251,253, etc)）

Response 3: Many thanks for your useful comments. In fact, this manuscript rehearses the key NLOS coping method or mechanism from the references mentioned before. So some key outlines truly need to be made. And it will make the manuscript more clearly and easily to read and these parts can be connected to the following methods easily. All these essential issues have been added several parts in Manuscript Version 2. They can be seen from the line 267 to the line 272 and from the line 127 to the line 137.

Furthermore, the authors' second institution name has also been modified due to the sudden change of its English name officially. Although this institution does not change its Chinese official name, it adjusts the English official name indeed [1]. The actual institutional relationship and the order of the authors have not been changed.

Meanwhile, there is something special happened with our teammate last month, so we have to add a new Funding Resource from the author Ruixiang Kan to make sure to have enough ability to pay for APC. So please allow us to add this new Funding Resource (Innovation Project of GUET Graduate Education No. 2023YCXB05) to this manuscript.

Some other words and sentences are adjusted and several new reference papers are added when necessary.

References

[1]URL: https://cise.glut.edu.cn/

Author Response File: Author Response.docx