A Novel Decision Support Methodology for Autonomous Collision Avoidance Based on Deduction of Manoeuvring Process

Round 1

Reviewer 1 Report

Thank you adding the additional information about your evaluation.

I kindly propose to rework you chapter for the simulation.

Please recognize that a position is not given in nm but in degrees lat log, courses are presented in 3 digits (eg. 003) with the information of heading oder cog. Distance are not given in kn but in nm. In the other diagramms you state miles. Did you mean nautical miles. Pleas apply standard nautical language and units in its correct usage.

Please review the paper carefully and provide additional information of you setup. Simulators used etc.

The technical description of the simulation is scientifically faulty.  eg. the units are wrong. This is not acceptable for a scientific journal. This cant be published without significant update.

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Authors' response is satisfying, the present  status of the paper can be published. 

Author Response

Dear Reviewer:

It’s highly appreciated of your time involved in reviewing the manuscript and your approval of the article. Thank you again for your positive comments and valuable suggestions to improve the quality of our manuscript.

Your sincerely,

All authors

Reviewer 3 Report

The authors made a significant effort to match the reviewers' requests. Despite the authors did not provide a point-by-point response letter, I can recognize that my comments have been more or less followed within the chaos of "track changes" mode.

There is no reason to ask for further review cycles, since I lost a bunch of time checking an almost unreadable .pdf, and trying to figure out myself the answers to my own comments.

NOT PROVIDING A REBUTTAL LETTER IS POOR AND DISRESPECTFUL PRACTICE. PLEASE DO ALWAYS PROVIDE POINT-BY-POINT A REBUTTAL LETTER TO PAPER REVIEWS IN THE FUTURE.

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

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 paper covers a relevant topic and is written in an understandable way.

The scientific assessment / proof of the approach is missing.

Some minor comments:

Figure 1 and the argumentation is not very helpful. Because most nav task are done by humans the number of human errors is naturally high.

The overyes of related work is comprehensive.

The paper is quite long with a lot of site topics addressed. A streamlining would help the readers .

The main remark is about the evaluation of the method. In chapter 4 application examples are given. it is not clear how they are generated / implemented (simulation?). The theare useful to understand the approach but is not a suitable experimental proof of concept. This 12 page description is very long but provides no scientific evidend. A proper verifiation and validation applicable for a safety critcal system is needed.

Reviewer 2 Report

Remarks enclosed

Comments for author File: Comments.docx

Reviewer 3 Report

The paper presents a COLREG compliant collision avoidance approach, and validates it in a simulated scenario. The paper is very long and presents some interesting contents. I have the following specific comments:

1) The paper claims to consider "good seamanship" into the algorithm. It is not clear how this is integrated into the decision process. It seems that, in the paper, "good seamanship" does not add anything not included in the COLREG.

2) The paper is very long and some concepts are repeated many times. I think that it could be optimized a bit. For instance, section 5.2 summarizes what the contributions of the paper are, but the conclusions do as well. My suggestion is that reorganizing the contents might improve readability.

3) I would suggest adding a description of the simulation setup used for the validation. How the dynamical model is implemented? how the control system interacts with the model and the collision avoidance system? How does the collision avoidance system get the target ship data? Did the authors take into account any detection system or measurement error? For instance,  a block diagram representing the experiment setup might help.

4) The authors might want to elaborate more on the implications of translating in the real world the proposed approach. For instance, measurement error, detection systems, real-time, limitations of the validation setup.

5) To complete the literature review, I would suggest adding some relevant references about RRT-based methods including COLREGs

Chiang, H.T.L.; Tapia, L. COLREG-RRT: An RRT-Based COLREGS-Compliant Motion Planner for Surface Vehicle Navigation. IEEE Robot. Autom. Lett. 2018, 3, 2024–2031

Zaccone R. COLREG-Compliant Optimal Path Planning for Real-Time Guidance and Control of Autonomous Ships. Journal of Marine Science and Engineering. 2021; 9(4):405. https://doi.org/10.3390/jmse9040405

6) Figure 2: some words are broken in two lines.

7) Figure 3: It is unclear a) where the entry point of the algorithm is, and what the outcomes are. b) what relation the blue box on the right represents. 

8) Equation 5 is incomplete.

9) Section 5.1: I would not indicate the distances with four decimal digits. I suggest to report them in meters.

10) The writing mixes US (e.g. recognize, maneuver) and UK spell (e.g. manoeuvre). Choose one and be consistent.

11) Figure 17 in the appendix: turning maneuvers are ovalized, I suggest using equal scales on the axes.