Resource Allocation and Offloading Strategy for UAV-Assisted LEO Satellite Edge Computing
Round 1
Reviewer 1 Report (Previous Reviewer 1)
The authors answered all my doubts. The article is more clear now than it was before. The reviewer recommends it for publication.
Reviewer 2 Report (Previous Reviewer 3)
Authors answered my concerns.
English is OK
Reviewer 3 Report (Previous Reviewer 4)
I thank the authors for carefully addressing all my comments. I believe that the paper has improved significantly. There is only a minor typo on line 1 of algorithm 2 that I believe can be easily corrected in the final version.
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
The manuscript considers a UAV-assisted integrated network architecture to solve an interesting offloading problem. The article is generally well-organized and easy to understand. However, the medium access technique was not discussed or mentioned, nor was the Doppler effect of moving devices mentioned. Reconsider mentioning them. In addition, the reviewer believes that the following issues should be rectified:
1) Some variables used in the equations were not defined, for example, sigma in equation 3. The reviewer recommends reviewing the rest of the document.
2) The authors should include more information about the distribution of the channel used. Why do the authors not consider small-scale fading?
The article is well written.
Author Response
We would like to commence by thanking you for constructive comments. Your expert knowledge of the field has helped us significantly. According to the valuable suggestions provided by you, we have revised the manuscript. We endeavored to address all the comments and our reflections are now given below point by point in the attached document.
Thank you again for your valuable time and efforts in handling the review of our manuscript as well as your expert comments and suggestions.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The abstract is very short and simple, thus unable to relate the technicality of the methodology and relevant emergency scenarios considered in the study.
Line 50, Pg 2: Unable to relate/connect between acronym OADDPG and resource allocation and offloading algorithm
DRL = Deep Recurrent Learning? Please run a check and make sure all abbreviation is properly defined and established on the first use.
Section 2 with an over-simplified literature study that made cross references with plenty of citations, for example, ref [16-18] on computational triage problem while ref [20–26] to relate with 'UAV trajectories' only - please elaborate more on the details, such as limitations or gaps based on these past studies that have been cited
Figure 1 - please fix some typos in the figure.
Sect 3.3. can be improved by adding relevant references (such as standardization) to support the assumption and selection of parameters, such as the value of BW, transmission power, antenna gain, etc.
Sect 5.1 is poorly presented. Present the parameters in a table format and elaborate to justify the selection behind these parameters and relate them with NTN/air-to-ground communications. The text towards the end of page 11 appeared truncated - please fix this too.
Sect 5.2 & Fig 7 - try to trace what TD3 represents - couldn't find it even in the sect related to previous works. Found DDPG in the section related to previous works. However not clear the original reference/work is since not been cited by the authors. In addition to that, please improve the discussion and figures by adding relevant references to the comparisons.
Sect 6 - please remove 'In this thesis' - because, obviously, this is not a thesis. Did all the authors run a proofread on the paper before making a final submission?
Page 12, Line 373 - please explain what does it mean by batch size? Relate with IoT applications or deployment scenarios considered in the paper.
Author Response
We would like to commence by thanking you for constructive comments. Your expert knowledge of the field has helped us significantly. According to the valuable suggestions provided by you, we have revised the manuscript. We endeavored to address all the comments and our reflections are now given below point by point in the attached document.
Thank you again for your valuable time and efforts in handling the review of our manuscript as well as your expert comments and suggestions.
Author Response File: Author Response.pdf
Reviewer 3 Report
1. I think the main issue with this paper are the results, the figures are very hard to understand and the results do not dipict a lot of enhancement. The authors need to provide their results in a different manner that makes it easier to understand the enhancement.
2. Some major references are missing such as https://0-doi-org.brum.beds.ac.uk/10.3390/drones6110334 which provides insites into drones and satellites. Can the authors explain how their work provides contribution in comparison.
3. The related work section is too small, there is a lot of work on this topic, which makes it diificult to accept this paper since a proper literature review is not provided.
English needs work but is ok.
Author Response
We would like to commence by thanking you for constructive comments. Your expert knowledge of the field has helped us significantly. According to the valuable suggestions provided by you, we have revised the manuscript. We endeavored to address all the comments and our reflections are now given below point by point in the attached document.
Thank you again for your valuable time and efforts in handling the review of our manuscript as well as your expert comments and suggestions.
Author Response File: Author Response.pdf
Reviewer 4 Report
The paper proposes a method for task allocation of computational resources for computation offloading to LEO satellites of IoT devices involving UAVs as relays. The proposed method minimizes a linear combination of the computation time and expended energy and consists of a Monte Carlo optimization to select the satellites and a deep deterministic policy gradient method to select the tranferred task size and computing speed. The problem addressed in this paper is interesting and relevant, and the effectiveness of the proposed method is evidenced by the numerical results. The main shortcoming if this paper is that the problem description, particularly the notation, is unclear in some parts.
Specifically, I consider that the following issues should be addressed before an eventual publication:
- I believe that it is implicit from the text that the link between each UAV device and each drone is fixed and done a-priori, but this should be more clear.
- In line 148 the UAV altitude is represented as Hm and in equation 7 as hm.
-In equations 25, 26 and 27 it is not clear how Ttran, Tprop and Tj are computed.
- It is not clear wht the cost U in equation 38 stands for.
-In algorithm 1 and at the end of page 11 there are formatting issues.
The English language is adequate. However, one can find some unclear sentences as in lines 128-130, or in line 162 when referring to the speed of light as "Lightning speed".
Author Response
We would like to commence by thanking you for constructive comments. Your expert knowledge of the field has helped us significantly. According to the valuable suggestions provided by you, we have revised the manuscript. We endeavored to address all the comments and our reflections are now given below point by point in the attached document.
Thank you again for your valuable time and efforts in handling the review of our manuscript as well as your expert comments and suggestions.
Author Response File: Author Response.pdf
