Fast Path Planning for Autonomous Ships in Restricted Waters

Round 1

Reviewer 1 Report

Dear authors,

I found that this research work develops  a  very interesting and an  attractive topic  for  a large community of the readers and for the specialists in the field. I appreciate your considerable efforts to be  very well informed in the field revealed by an excellent literature review updated with the most recent scientific papers shown in the Reference section. Also this  research is of  high quality, very  well written, with a high significance of the content  and scientific soundness. 

I think an  extension of  the PGHAPF method to generate adaptive, collision-free and COLREGS-constrained trajectories in  restricted waters by probabilistic calculations can be an  interesting challenge. The model validation of the ship maneuvering motion could increase also the  value of this work. For the readers is interesting to precise the software package used in all your simulations. Precise also the selection criterion for the truncated Taylor series order of the X,Y and N functions.

Thanks,

Author Response

Author Response File: Author Response.pdf

Reviewer 2 Report

It is a fine, but long paper.

There is a high count of abbreviations, of which many are not commonly used, this makes it, at times, more difficult to read then necessary.

The method describes the benefit of combining an off-line path planning  with force field for static and dynamic obstacle collision avoidance.

The off line path planning segments the path into a number ow way-points, a way to avoid being trapped in obstacles with concavities.

Both methods are well described in earlier papers, but there is a series of improvements and addition of regulations and realistic ship dynamics – that is fine.

Specific issues:

Table 1, row Emergency CA, what is the difference between “no” and “N”?

Line 102: PGHAPF is explained again, but this is fine with the number of new abbreviations.

There is other abbreviations that should have been expanded to full text, e.g. OS = Own Ship (and not Operating System). This would make the paper more readable – and not much longer.

Formular (3): There may be a rendering error in my pdf-file, the right hand side look like “\frac{1}{1 + e^{-\gamma \square  c}}” where I would have expected something like “\frac{1}{1 + e^{-\gamma f(c)}}”.

Formular (4): The total force field is defines as product of segment lines for an obstacle, this works in most cases, where all but one most likely has a value of 1, but a concave obstacle with straight lines in the concavity will give problems as more than one line line will have a value close to zero. You avoid the problem in figure 4, where the concavity is a curve. This problem can be avoided, but then (4) is too simple.

The flowchart in figure 10: The combination of Psi_os and Psi_os from the fixed obstacle path is not clear. Will a fixed obstacle "(f(p2) > \lambda) = yes" inhibit any processing from the F_tottal branch, i.e. no used path or dynamic obstacle calculation? This seems a little dangerous.

Towards the end there is a yes/no block (d_k < r_k …), but this has a third exit (neither “yes” nor “no”!), probably there should be a “last segment or not”.

The situation in figure 16(b) could easily be a situation where  the assumption of constant speed could be wrong – if the TS4 had been delayed a bit. Further, if the Obs4 obstacle had been with a major cavity, then OS could be trapped?
All in all there could be a situation where a new set of way-point should be calculated, maybe the other way around obstacle Obs4 or Obs5.

In figure 19 and the text around it, you discuss why using a bisector line is a good idea, but that has already been established earlier (section 2.3.1).

It would have been much better to try a simulation where all the TS was taken from real traffic (e.g.: https://www.marinetraffic.com through one of the narrow straits close to Singapore) and then focus on some of the close encounters, or a similar extensive simulation.

p { margin-bottom: 0.1in; line-height: 115%; }a:link { }

The lack of exhaustive simulation is a problem for this paper, as it claims that “...the PGHAPE method is able to generate adaptive, collision-free and COLREGS-constrained trajectories in restricted waters ...”.

p { margin-bottom: 0.1in; line-height: 115%; }a:link { }

Author Response

Author Response File: Author Response.docx

Reviewer 3 Report

This paper deals with Fast path planning for autonomous ships in restricted  waters based on a Path-Guided Hybrid APF method. This paper is well organized through whole paper and well express the previous work. However, I would like to point out following as:

Typically, water waters based on a path planning case should consider communication status in real system, did you consider it for computer simulation.

 I hope authors add following reference.

(1) Gia Huy Dinh  and Nam-Kyun Im, A Study on the Construction of Stage Discrimination Model and Consecutive Waypoints Generation Method for Ship’s Automatic Avoiding Action, International Journal of Fuzzy Logic and Intelligent Systems, Vol. 17, No. 4, December 2017, pp. 294-306, http://dx.doi.org/10.5391/IJFIS.2017.17.4.294

 

 

 

Author Response

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Thank you for the detailed explanations. 

- It seems now like you have done a lot of testing, but that is not visible in the paper. I would have preferred that you had prioritized the testing, and had some metrics for the performance, e.g. close encounter distance compared to number of encounters that require navigation action, and an idea of how dense traffic or environment should be before the method fails.

- I am a bit dissapointed that you only expanded one abbreviation (OS), but I would probably have done the same.

- in my view the major direct errors are corrected, so that must suffice for my part.

Author Response

Response to Academic Editor and Reviewer 2

Point 1:   Your manuscript "Fast path planning for autonomous ships in restricted waters based on a Path-Guided Hybrid APF method" has been reviewed by three independent experts. It appears that the recommendations of the reviewers were successfully addressed.

Response 1: Thank you for the approval. The comments and suggestions raised by the reviewers and editors are of great significance. They help me find the problems and correct them finally.

Point 2:   Yet I am still concerned by the amount of acronyms used in the paper including the very title. I suggest to avoid acronyms (as APF) in the manuscript title. Also, I suggest to reduce their use in the paper.

(Reviewer 2: I am a bit disappointed that you only expanded one abbreviation (OS), but I would probably have done the same)

Response 2: Thank you very much, I have reduced the amount of acronyms used in the paper (see the turquoise shadow text in the manuscript), and the title has been changed to “Fast path planning for autonomous ships in restricted waters”.

 

Point 3:   On top of this, I would also include a table summarising the acronym meanings at the beginning of the paper.

Response 3: Thanks, this is a good suggestion. A table summarising the acronym meanings has been included at the beginning of the manuscript.

 

Point 4:   Please also have an additional check on English.

Response 4: this manuscript has been checked and edited by a professional English editing service (as American Journal Experts, AJE), please see the attached certificate. Besides this, an additional check on English has been done after the latest revising.

Point 5:   In the light of the last reviewers reports and of my own reading of the contribution, I recommend to accept the paper after minor revisions (English and acronyms issues in particular - including the paper title)

 Response 5:  Thanks, the minor revisions have been done.