Target Trajectory Prediction-Based UAV Swarm Cooperative for Bird-Driving Strategy at Airport
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe main idea of this manuscript is to develop a novel bird-driving strategy using UAV swarms to protect airports from bird strikes. The proposed system combines trajectory prediction, task allocation, and path planning using advanced techniques such as LSTM networks and Kalman Filters to enhance the efficiency of the UAV swarm. The system ensures timely and accurate bird dispersion, improving operational safety at airports by driving birds away from protected areas. The combination of LSTM networks with Kalman Filters allows the system to handle nonlinear and non-Gaussian noise effectively. However, there are some minor suggestions:
1. The authors should explain how it can be implemented in real-time UAV operations, considering the limited onboard processing capabilities.
2. Comparative results with other existing methods need to be added, to convince the proposed method's robustness.
3. It is suggested to use the same colors in Fig. 3a and 3b for the same method.
4. Fig.8 needs to be explained. Why is one of the "goal pose" down far away from the other four?
With these refinements and additional comparative analysis, it may contribute to the field of UAV applications.
Comments on the Quality of English LanguageThe Quality of English is fine.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsDear Authors
1. Please explain in detail how you prepared the training data set for bird flight trajectories. First, what bird species were recorded. What was the method of recording the trajectories. Was the spatial trajectory (XYZ) archived? Where was the recording carried out (actual location near the airport)? In what periods of time was the recording carried out (different seasons?, different months?) - birds often appear seasonally in specific locations. This will present a picture of how representative and actually useful the training sets were.
2. Why are the trajectories considered only in two dimensions and not in 3D space? See figure 3.
3. What types of errors were described in chapter 3.2.3. Is this an error in some specific axis? An average value?
4. Were any consultations with ornithologists carried out? The assumption that the appearance of drones in the bird's environment will force it to change its flight direction may be incorrect. Some birds may treat this as an attempted attack and counterattack, damaging one of the drones or injuring themselves. A damaged drone may pose an additional threat to aircraft.
5. The article would gain much credibility if an experiments verifying the assumptions and simulations were at least partially conducted.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThank you for submitting your work to Electronics, the manuscript presents a cooperative bird-driving strategy utilizing UAV swarms designed to mitigate the risks posed by birds at airports. The proposed approach integrates Long Short-Term Memory (LSTM) networks with Kalman Filters for target trajectory prediction, the Hungarian algorithm for task assignment, and 3D Dubins path planning for UAV flight paths. The objective is to improve response speed, task optimization, and accuracy in bird-driving tasks using UAV swarms. Simulation results demonstrate the strategy's effectiveness in dynamically repelling birds from protected airspace, offering improvements over traditional bird deterrence methods.
The integration of LSTM-Kalman Filter for bird trajectory prediction is an interesting research direction, as it combines the strengths of both real-time state estimation and handling nonlinear time series data.
The literature review in the introduction could use an addition to similar studies on other applications than airports such as:
Bhusal, S., Karkee, M., Bhattarai, U., Majeed, Y., & Zhang, Q. (2022). Automated execution of a pest bird deterrence system using a programmable unmanned aerial vehicle (UAV). Computers and Electronics in Agriculture, 198, 106972.
As the topic is very niche, I would suggest even a table with the relevant studies and why further research is needed to fill in knowledge gaps or improvements on performance.
The process noise covariance matrix is pointed but unclear how it works, further explanation needed.
Figures 3a and b are unclear, where does the trajectory start? The prediction error shows high alignment in the beginning and end, however, the segment visual show only alignment at one end.
In “4. UAV swarm cooperative bird-driving strategy at airport” it is unclear how many bird attacks are assumed and what are the general simulation assumptions. This section of the manuscript needs significant reconsideration.
Algorithm performance, using alternative models such as Particle Filters or Reinforcement Learning models to converge the same problem would give the audience a sense on the better performance of the proposed method.
Comments on the Quality of English LanguageVery minor English revisions required for run-on sentences and tense consistency
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsDear editor and authors
My only comment has to do with the implementation, although the authors do comment on the uav cost, what other equipment is required for the voerall setup. I.e. cameras for noticing the birds. Also,
--are the computations performed on the uav or off on a base station?
--Is the base station feeding the position of the flock of birds to the uavs? I am assuming it is a setup with multiple cameras, a gps tracker etc? This brings me back to the above question
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 5 Report
Comments and Suggestions for AuthorsPlease see the attached report.
Comments for author File: Comments.pdf
It is better to omit the colon before the formulas considering them as a part of the sentence.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe authors have adequately addressed my inquiries.
Comments on the Quality of English LanguageFine.
Author Response
Thank you for taking the time to provide valuable feedback on the manuscript amidst your busy schedule. Your patient guidance and professional advice have greatly benefited us! Thank you!
Reviewer 2 Report
Comments and Suggestions for AuthorsDear Authors,
Thank you for taking my comments into account and adding appropriate explanations and content to the manuscript. I believe that the article in its current form can be published
Author Response
Thank you for taking the time to provide valuable feedback on the manuscript amidst your busy schedule. Your patient guidance and professional advice have greatly benefited us! Thank you!
Reviewer 3 Report
Comments and Suggestions for AuthorsNo further comments
Author Response
Thank you for taking the time to provide valuable feedback on the manuscript amidst your busy schedule. Your patient guidance and professional advice have greatly benefited us! Thank you!
Reviewer 4 Report
Comments and Suggestions for AuthorsDear editor and authors
The authors have succesfully replied to my comments.
Yet, for some reason, they decided not to include these information that they provided me inside their paper.
I strongly believe that their comments must be included inside the paper as well.
Thus, I suggest minor revisions, for the authors to perform so.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
