Wits: An Efficient Wi-Fi Based Indoor Positioning and Tracking System

Round 1

Reviewer 1 Report

The authors presented Wits: An Efficient Wi-Fi Based Indoor Positioning and Tracking System. The work seems very promising and has good merits. However, I would like the author's response to the following queries:

1. Most of the references are outdated, a lot of work has been done in 2020-2021 which I cannot see to be referred or consulted in this work. 
2. I cannot see any reference to the comparison of parameter-based vs. fingerprinting-based approaches. Please refer to https://doi.org/10.3390/electronics8020195 where the fingerprinting shows quite promising results. 
3. The experimental setup seems to be too limited, and simple, which cannot be applied to a real-time scenario. Furthermore, Figure 6 shows only the outdoor environment with no hurdles i-e. line-of-sight communication only. what about the non-line-of-sight environment?
4. I cannot see any significant contribution from the authors in terms of mathematical modeling. Eq. 1-29 are all well known, it seems to be a better optimization of the existing models, not a novel approach. 
5. The results/conclusion section failed to provide any qualitative/quantitative improvement for the proposed algorithm, which is a big limitation in this work. I would like to see some numerical representation like X% improvement compared to existing methods.  The results section just presents a summary of better computational time, ignoring several other KPIs in indoor positioning 

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

First of all, it was interesting in that it tried to solve an important problem of existing interior localization, and the authors suggested a solution to estimate the location by recognizing the speed of an object based on the status information of the Wi-Fi signal.
And there were quite a few papers that suggested a model with the premise that location estimation is possible if the initial location is known, and in this paper, I thought it was a paper that considered the realistic environment well in that it derived the error rate when knowing and not knowing the initial location. 
However, it needs to be improved in the following aspects.

1. Lack of explanation.
Most of the suggestions are composed of formulas, but there is a lack of explanation on how the formula was derived and what it means.
ex 1) Equation (6) says it can remove random noise, but there is no explanation as to why. The content that random noise can be removed is a large part of the contribution of this paper, so it would be good to show how it differs from the existing method by using simple values as an example.
ex2) The noise obtained by the proposed equation (19) can be removed, but the reason is not clear.

2. Comparison and concreteness with conventional similar technologies
The results of the proposed model showing some error rate have been specified, but it is not clear how much the proposed model has improved compared to the conventional method.
It was explained that the model proposed by this paper manifested a high accuracy in that it manifested an error of 0.235m on average, so it would be good to specify how much the error rate improved compared to the conventional similar model in a multipath environment.

3. Need to maintain consistency of evaluation criteria.
According to the evaluation, the experiment was conducted in two environments (environment 1 and 2), and I thought the explanation of why the tracks tested in the two environments were different was insufficient. And the authors introduced several possible tracks, and need to explain the criteria for selecting U-shaped, vertical, and circular shapes.

4. Lack of explanation.
There is no explanation for Figure 4 in the text.

5. Experimental conditions.
All two experimental environments in this paper are cases where only one person exists, so if there are multiple people, it will be necessary to choose which person's location to track, but there is no explanation.

6. Lack of explanation.
There is a lack of evidence on why two of the user's movements were used in the experiment. The experiment was set up when the user moves quickly and stops slowly and when he moves slowly and stops quickly, but there is a lack of explanation as to why the authors chose only two cases in particular and what to check. I wonder why the authors excluded cases such as moving fast and stopping fast, moving slow and stopping slowly.

7. Comparison of the performance of Figure 8.
There are many overlapping parts between the method proposed in Figure 8 and the conventional method graph, and it looks similar. So it seems difficult to say that the proposal for error estimation clearly shows good performance.

8. Terms don't match.
I understood that the name of the algorithm proposed in the paper was Wits, but the results of the experiment or the part explaining the proposed algorithm were written as "The algorithm in this paper" instead of the name.

9. Picture index.
In terms of the content and context of page 8, the picture described in the second paragraph seems to be Figure 4, but I think there is a typo in Figure 5.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors have well-addressed all my concerns in the revised version of the manuscript. I have no hesitation in recommending the article to be published in its current form. 

Reviewer 2 Report

I think this paper is ready for publication.