Low-Complexity Three-Dimensional AOA-Cross Geometric Center Localization Methods via Multi-UAV Network
Round 1
Reviewer 1 Report
1. Why AOA-based localization approach is superior to other alternatives? More details should be highlighted in the Introduction section.
2. More recent works on 2D-AOA estimation should be stressed, e.g., Efficient Super-Resolution Two-Dimensional Harmonic Retrieval With Multiple Measurement Vectors, In IEEE TSP; Compressive sampling framework for 2D-DOA and polarization estimation in mmWave polarized massive MIMO systems, In IEEE TWC;
3. Subtitle 3 should be consist with others;
4. Why (39) established, any reference? Besides, the superscript -1 should be explained;
5. Title of Table 2 is incorrect, please checked;
6. Format of references should be unifined.
Please check the manuscript carefully and make sure ALL the symbols and formats are unifined.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The paper contains a very potential valuable contribution. The subject is within the scope of the journal but the paper has serious flaws. The scientific soundness and the finding are not clear. In the opinion of this Reviewer the manuscript deserves to be published once the Author takes into account the raised issues.
It is not clear at all if the final goal is to localize the source or the UAV. Figure 2 shows a source and UAVs. The authors talk about multi-UAV network and seem that they use the information of more UAVs to calculate the sphere. At the same time (in the conclusions) they say that the methos is for UAV localization. But they calculate the source location. It’s very confusing.
Minor
1. The authors should check that all the used acronyms are explained and not repeated (as UAV, AoA, ToA, MLE, etc)
2. Mainly the English is good and there are only a few typos. However, the paper should be carefully rechecked.
1. Mainly the English is good and there are only a few typos. However, the paper should be carefully rechecked.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
A rather sleek formulation allowing M UAVs to locate a radioemitter is presented. The inscribed sphere (CIS) formulation allows an easy definition of 2 new methods (one estimates the noise effect, too, another does not).
Detailed comments:
p.1: expand ISAC better in the text
p.2: noval --> novel
p.3: In practical, --> In practise,
p.4: integrate the idea of geometric center -->
formulate the problem using the center of inscribed sphere of a tetrahedra (I think this is what you mean ..(?)
the line u_i u_0 --> this is bit vague. Are points u_i at the vertices? Is the u_0 the center of the inscribed sphere?
Figure 2 seems to show inscribed sphere and not the geometric center (the latter is usually in the literature the mean of vertex points)
p.7: introduce \hat{u} (the meaning). Somehow bind this to u_i u_0 of the Eq. (5). (Either at p. 4 or at p. 5 Eq. (13) ....(14).
p. 7: CIS, This --> CIS, this
p.8: compute --> computed
A note about alternative way to solve the center p and radius r of the inscribed sphere, and its radius r (3 + 1 unknowns) as a set of 4 linear equations (one per tetrahedron face). This applies when M= 3. This is not to force you to adapt this to your work, but I just propose you have a look at it at a suitable occasion in the future.
One condition for a face 123 is:
(c1 + n123 r - p ) . n123 = 0 (a point p is on a plane lifted from vertex c1 by r along the normal 123 where
n123 = ((c2 - c1) x (c3 - c1))^0 is the unit normal of the plane through (c1,c2,c3)
All 4 planes have a common point p (the center of inscribed sphere) at a distance r. Thus, one 4x4 linear equation gives p and r, and to include angular noise, one can repeat the solution 3 more times by varying 4th vertex (the source) in x-y-z directions to get the angular noise transformed through a linear mapping created by variations.
When M= 2, the same applies with only 3 equations, and p can be projected to be of 2-dimensional (on plane 1-2-3). For M= 4 .... the linear system would be overdetermined, or 4-tuplets of the set 1..M were to be used as the summands, but the noise term transformation would build up as well. Maybe worth of checking out (also I may be wrong, too, this is a quick sketch).
A weenie bit of proofreading needed from the part of authors...
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
No futher comment.
Author Response
Thanks to you for your thoughtful and insightful comments.
Reviewer 2 Report
Dear author,
thanks for the clarification that did not allow me to understand the proposed method at all.
Nonetheless, however, now another aspect that is not at all clear is the possible implementation in a real scenario. How do you plan to do data fusion of the information that each UAV collects? First of all I think you need a real-time measurement system in HW and not in SW (e.g. https://doi.org/10.1109/TCSII.2022.3141247 and https://doi.org/10.23919/EuCAP53622.2022.9769059 , paper that can be cited in the text.).
Another point is the time synchronization among the UAV. There are also in this case a lot of algorithms, like FLSA, TPSN and RBS. Also in this case, please cite some of them.
Even if you are still in the simulation phase, I think that to make the paper complete, it is necessary to at least propose a possible implementation solution.
Minor
1. The authors should check that all the used acronyms are explained and not repeated. They are still duplicated in the abstract and in the introduction (as UAV, AoA, ToA, etc).
Mainly the English is good and there are only a few typos
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 3
Reviewer 2 Report
Authors have properly enriched their work, by addressing each comment in a suitable way. The paper turns out to be notably improved.
Minor editing of English language required
