A Deep Learning Approach for Wireless Network Performance Classification Based on UAV Mobility Features

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The contribution of the paper lies in investigating the relationship between the mobility index values and the network performance of UAV-based networks using a backpropagation neural network (BPNN). The authors are suggested to modify this paper according to the following comments:

1) The title of the paper is not proper. Thus, the authors should change the title of this paper to reflect its content and give emphasis on the use of UAVs.

2) Many abbreviations are not defined throughout the text.

3) As there are numerous papers that investigate UAV-based networks and propose optimization methods, the authors should more clearly highlight the main contributions of this paper to ascertain the main technical contributions and improvements of this paper compared with previous work. Thus, the authors may provide a table summarizing the main differences/similarities of their paper with respect to the state-of-the-art.

4) The authors should better describe why the proposed BPNN method is necessary. Which are the special characteristics of this method that discourage the application of other methods? A discussion about the complexity of the proposed method should be also included along with a discussion regarding the feasibility of this method in real-world scenarios by taking into account the resource-constrained UAV nodes. Although the selected method seems appropriate, UAV's limitations in terms of energy and computing resources are not considered. These limitations may restrict the application of the proposed method.

5) A discussion on practical use case scenarios of the considered network will be useful.

6) In order to verify the superiority of the performance results of the proposed technique, these results should be compared with the results that can be obtained using other techniques in the literature.

Comments on the Quality of English Language

The paper is generally readable. However, the organization of the paper can be improved and there are several grammatical and syntax mistakes.

Author Response

Attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

1. The definition and symbolic expression of variables in Mobility Indicator Design need to be reformatted. The notations of Equation (9) and Equation (12) are incorrect and need to be modified.

2. The article lacks research and explanation of related work on deep learning based wireless communication.

3. The BP neural network is too basic and has fewer overall input features. It is recommended to increase the number of features and choose a more advanced network architecture.

4. The five features are highly correlated, and the generated parameters overlap more. Whether this feature selection will affect the routing performance is determined.

5.The proposed method used BPNN to explore the relationship between the motion characteristics of mobile nodes and the performance of routing protocols and realized the classification.  However, the reviewer should consider the temporal variation of features and performances and aggregate them like [1]. A comparison would be grateful between the proposed BPNN method and the feature aggregation neural network method in [1]. Please cite it and make comparisons in the manuscript.

[1] "AI-Driven Blind Signature Classification for IoT Connectivity: A Deep Learning Approach," in IEEE Transactions on Wireless Communications, vol. 21, no.  8, pp. 6033-6047, Aug. 2.

6. It has been clearly pointed out in the previous section that GM and RWP are not suitable for simulating UAV trajectories, and the comparison of the three models in Section3.1 is not necessary. It is recommended to illustrate the environmental conditions for dataset generation and model training.

Comments on the Quality of English Language

A few sentences have grammatical problems.

Author Response

Attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Author discuss on the Wireless Network Performance Classification based on Mobility Model Connectivity using Back Propagation Neural Network. Three mobility model explain which are

Mobility Model, Random Waypoint Mobility and Reference Point Group Mobility as well as the

indicator of the performance. Results shows in figure 5 which is RPGM is the good result, finally 

accuracy test have been done between RWP and RPMG model as shown ini figure 9 and 10.

Recommend to publish with minor revision is most of figure need to enlarge to see clearly and

in the references started from 9 that should be number 1.

Author Response

Thank you very much for your suggestions to our work. We have made adjustments and modifications to the images and references according to your requirements.  Hoping to receive your support.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Τhe authors have addressed my major concerns in a reasonable manner. The paper has been significantly improved and the revised version can be accepted for publication.

Comments on the Quality of English Language

Moderate editing of English language required.

Reviewer 2 Report

Comments and Suggestions for Authors

No further comments.

Comments on the Quality of English Language

N/A