Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Title: Research on Mitigating Atmosphere Turbulence Fading by Relay Selections in for Free-Space Optical Communication Systems with Multi-Transceivers

This paper proposes relay selection in free-space optical communication systems with multi-transceivers based on reinforcement learning to reduce atmospheric turbulence fading in limited channel conditions. The topic of the paper is interesting and the results reveal good improvement, however, some revisions should be done as follows:

 

1-     Considering the availability of analytical equations related to relay selection based on FSO for database preparation and supervised learning, why did the paper choose reinforcement learning? This issue should be clarified and addressed.

2-     Recently, many papers have been published on optimizing channel capacity using deep learning methods, but only a few are discussed in the paper. The literature survey should be completed.

3-     In this study, only the Málaga distribution was used. Optimization with other distribution models, such as gamma-gamma and Log-normal logarithm, should also be performed and added to the paper's results.

4-     When implementing relay-based free space optical communication for long channels, what are the reasons for utilizing OOK modulation? The paper should also include the performance of other modulations like PPM and NRZ, and compare them in terms of optimization.

5-     If there are geometric losses between the transmitter and the relay, as well as the relay and the receiver, how will the proposed deep learning model perform? Please clarify.

6-     To assess the improvement in channel capacity, it needs to be plotted and reported based on the training steps.

7-     How can the proposed learning model's performance be optimized in conditions of channel attenuation, such as smoke and fog, and constant power in the transmitter?

 

8-     Finally, the paper needs to be carefully edited. For example, in the Abstract “free-space optical communication systems (FSOC)” should be read as “free-space optical communication (FSOC) systems”.

Author Response

The response to reviewers’ comments are in the Word.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This paper proposes a Reinforcement Learning-based Relay Selection (RLRS) method based on Deep Q-Network (DQN) in a FSOC system with multiple transceivers, whose aim is to enhance the average channel capacity of the system. The proposed method is promising. Followings are the comments from the reviewer to improve the quality of the paper.

1). The literature should be improved. For an example, there are different methods to mitigate the different types of turbulence effects in optical wireless communication. More details and references can be provided to improve the literature. One example is 

Kapila W. S. Palitharathna, Roshan I. Godaliyadda, Vijitha R. Herath, and Himal A. Suraweera. 2018. Relay-assisted optical wireless communications in turbid water. In Proceedings of the 13th International Conference on Underwater Networks & Systems (WUWNet '18). Association for Computing Machinery, New York, NY, USA, Article 40, 1–5.

2). Some motivation should be provided why to use reinforcement learning.

3). It seems the main novelty of the paper is to use Malaga turbulence and mitigate the capacity loss. Try to explain clearly the main novelty and contribution.

4). May be the complete algorithm can be presented as a Pseudo code to show improve the readability.

5). Fig. 4, the average capacity curve seems to have less resolution and low averaging. Try to improve.

6). To provide more insights about the paper, better provide average capacity and performance gains with different system parameters such as number of relays, transmitters, transmit power, number of time slots, or with different channel parameters.

7). Format and align equation properly. Check paragraph breakings. Check Eq. 2, there are typos. After an equation the term "where" should not be in a new paragraph. Check the font sizes in page 5, some are not consistant.

8). There are several typos can be noted. Please check carefully and correct them.

Comments on the Quality of English Language

In general, the use of English in this paper is good. However, check carefully since further improvements to the use of English is possible.

Author Response

The response to reviewers’ comments are in the Word.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

No additional comment! The revised manuscript can be accepted for publication. Congratulation.