Next Article in Journal
A Fully Controllable UAV Using Curriculum Learning and Goal-Conditioned Reinforcement Learning: From Straight Forward to Round Trip Missions
Next Article in Special Issue
Hierarchical Online Air Combat Maneuver Decision Making and Control Based on Surrogate-Assisted Differential Evolution Algorithm
Previous Article in Journal
Application of Magnetometer-Equipped Drone for Mineral Exploration in Mining Operations
Previous Article in Special Issue
Jamming and Spoofing Techniques for Drone Neutralization: An Experimental Study
 
 
Article
Peer-Review Record

A Framework of Recommendation System for Unmanned Aerial Vehicle Autonomous Maneuver Decision

by Qinzhi Hao 1, Tengyu Jing 2, Yao Sun 1,*, Zhuolin Yang 2, Jiali Zhang 2, Jiapeng Wang 2 and Wei Wang 2
Reviewer 1: Anonymous
Reviewer 2:
Reviewer 3: Anonymous
Reviewer 4:
Submission received: 6 November 2024 / Revised: 22 December 2024 / Accepted: 23 December 2024 / Published: 30 December 2024
(This article belongs to the Collection Drones for Security and Defense Applications)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

 

 The authors put forward  Autonomous maneuvering decision-making in Unmanned Aerial Vehicles (UAVs) inspired by recommendation systems methods. This is an interesting work which makes a significant contribution tom the literature. To improve the article, the authors must address the following suggestions..
1. In elaborating the different recommendation methods, the authors mention model-based recommendation methods yet no further elaboration  or reference was made to this method. Its recommended that the authors provide a sentence describing this recommendation method with some reference for its usage or applications

2. The authors make the following emphatic statement .. “ The best-performing recommendation method at present is the  reinforcement learning method” is there any evidence or reference to support this claim?

3. IN section 2.1 “in maneuver decision-making, it means the set when the UAV cooperates and confronts ..” this statement should be clarified. Does the authors mean “the UAV cooperation and confrontation states”?

4. In section 2.1, the authors should provide a table that summarizes the mapping of the recommendation system to the RL learning problem.

5. It is recommended that the authors go through the manuscript to fix some typos and ambiguous statements such as;

a. “Refer to the modeling method mentioned in reference [26], and construct a recommendation system simulation UAV cooperation and confrontation environment”, on page 6

b.  Add “Where” after figure 3 ….” x,y,z represents the three-dimensional…”

c. The authors should avoid beginning sentences or paragraphs with symbols, for instance in the second paragraph of section 2.3, line 246 and 247

6. It will be more illustrative if the authors use different colored icons for the two sides UAVs in figure 4 same goes for figure 5.

7.  The authors should ensure that abbreviation are defined  first time they are used eg TD etc
8. The authors should extend the conclusion section to highlight the advantage and performance of their approach and the limitation therein. They should also include some possible extensions

 

 

Comments on the Quality of English Language

I think the english needs minor revisions 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper presents an innovative framework for autonomous maneuver decision-making in Unmanned Aerial Vehicles (UAVs) using a recommendation system approach. It leverages collaborative filtering and deep reinforcement learning to simulate and optimize UAV maneuvers in cooperative and confrontational scenarios. The study includes the implementation of both offline and online algorithms, such as KNN-UserCF and PER-DDPG, and evaluates their performance using simulation-based experiments.

The rationale for choosing certain recommendation algorithms over others could be elaborated.

The real world applicability is not discussed. The simulations operate under ideal conditions without considering disturbances like wind, communication delays, or sensor inaccuracies. Adding these factors would strengthen the claim that the framework can be used in real-world applications.

The construction of the dense reward system is logical and enhances learning speed. Nonetheless, the paper could further discuss potential limitations of the chosen reward structure and whether alternative reward functions were considered.

Further discussion on the computational trade-offs between these approaches would be valuable for real-time applications.

The conclusion need to be rewritten, especially to reflect/discuss the strengths and limitations of the research. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This paper presents a novel framework of recommendation system for unmanned aerial vehicle autonomous maneuver decision. The topic is interesting and the paper is well written. It can be accepted after minor revision.

 

My main conern is marking an 80% reduction in training time”. It is about the efficiency of the proposed framework. The authors should provide a detailed table. Moreover, computational analysis can be added.

 

Some minor comments are

(Page 1) The abbreviations in the article should be consistent, including the capitalization of the first letter.

(Page 1) Please pay attention to the singular and plural, such as application in Line 35.

(Page 2) I suggest the author to adjust the  introduction to highlight the motivation and innovation of the article.

(Page 4) The matrix and vector symbols in the article should be consistent.

(Page 5) There are many small paragraphs. I suggest the author to merge and organize them.

(Page 7) Figure 4 can be enlarged, otherwise it is difficult to see the words in it. Of course, this also includes some other pictures.

(page 12) I found that some parameter details are not given. The authors can summarize them in a table, which will be clearer.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The results in the article do not make clear the contribution or the engineering relevance of the results. There are many fielded autonomous decision making systems that have guarantees of performance under various environmental conditions.

Moreover, the results here do not appear to be reproducible as they can be sensitive to small changes in the simulation settings. How do you prove this framework is better than some arbitrarily put together collection of heuristics and algorithms?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

Comments made for the previous version are not adequately addressed. Specifically the references do not address the long history of decision theory, heuristics, and metaheuristics going with it.

The comments on engineering relevance have been adequately addressed.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Back to TopTop