Multi-Criteria Selection of Surface Units for SAR Operations at Sea Supported by AIS Data

Round 1

Reviewer 1 Report

The article presents an interesting concept of automating the process of selecting vessels to participate in SAR operation based on data received from AIS. MRCC operators already use data from AIS and fitting their workstations with appropriate software will improve their work and accelerate mentioned process. The too general description of the criteria for selecting ships for the planned operation proposed by the authors makes it impossible to assess the practical usefulness of the algorithm. In order to raise the substantive level of the article, I suggest taking into account the detailed comments.

Detailed comments:

1. The abbreviation SRU listed in keywords should be expanded.
2. Passenger ships are one of the two basic types of merchant ships and therefore cannot be listed separately.
3. The authors should write how and when they obtained data from the network of AIS coast stations of the countries mentioned in the article – through the monitoring system established in accordance with the requirements of the so-called Copenhagen Declaration?
4. Authors should describe the content of Figure 2 or remove this drawing from the article. Readers familiar with the principles of encoding data transmitted by AIS do not need this information. For readers who are not familiar with these rules the Figure itself is incomprehensible without explanation.
5. Authors should justify the selection for simulation of only the two preference functions described in the article.
6. The following sentence requires a stylistic correction: “If the coordinator is familiar with the area, e.g., if they 309 have established that the search will be carried out in an area of a depth of 10 m, and the 310 group of vessels under analysis includes ships with a big draught, such as, e.g., tankers 311 drawing 14 m, they can eliminate such vessels during the assessment”.
7. It is not true that vessel’s length overall (LOA) is typically correlated with the height of the superstructure.
8. Military vessels (warships, coast guard vessels) are more useful for SAR operation than merchant ships.
9. Crew qualification are determined by the STCW Convention not by SOLAS.
10. The authors should justify the selection of the values of the criterion weight, and indifference i preference thresholds presented in Figure 7.
11. One simulation presented in this article does not allow for a clear conclusion that described method is especially useful in areas of high density, where ships are of various types, with different equipped and carrying out different tasks. In my opinion, it only suggests such a possibility. Its practical usefulness requires further simulation studies, especially in the selection of appropriate preference function and thresholds and weight functions.

Author Response

Many thanks to the Reviewer for valuable remarks and comments. We have taken into account all suggestions and proposals.

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Multi-criteria Selection of Surface Units for SAR Operations at Sea Supported by AIS Data

1. Search and rescue (called SAR) is so interesting, from the synthetic aperture radar (SAR) community. At least, synthetic aperture radar can also provide some useful information for the ocean search and rescue.
2. The references are too old to find some useful information about ocean search and rescue. Please add more strongly related work.
3. AIS information is near ports. How to search and rescue ships among the vast ocean. Have you used the satellite AIS from some descriptions in ls-ssdd-v1.0: a deep learning dataset dedicated to small ship detection from large-scale sentinel-1 sar images.
4. AIS information seems to be later than the actual ship operation time. How to solve the time interval between AIS launch and the real-time ship locations. Have you conducted some time interpolation operations? Can you draw some supports from the Google Earth high-speed ship detection in sar images based on a grid convolutional neural network, depthwise separable convolution neural network for high-speed sar ship detection. Sometimes, some constant coastline information is helpful for ocean search and rescue.
5. When AIS is lost, how to perform search and rescue. In fact, there are many “dark” ships at sea. It is so common. You can draw some other tools, e.g., synthetic aperture radar. It can work all-day and all-weather. It is so powerful.
6. I strongly recommend it to you in the future. In addition, I share a rather interesting paper to you. Oh, I forget the paper title. Please search “dark fleets” in the Science Advances journal.
7. Line 110, the format is terrible.
8. Figure 4, there are many parameters from the AIS information, why do you plot the Latitude and longitude? Aren’t the ship size important in quad-fpn: a novel quad feature pyramid network for sar ship detection, hyperli-net: a hyper-light deep learning network for high-accurate and high-speed ship detection from synthetic aperture radar imagery?
9. In general, to be honest, the language of this paper is rather terrible. Many places cannot make me understand. Please take this point carefully.
10. In addition, the presentation of the proposed method is also rather terrible. Or, maybe I do not understand it very well.

Consider the above suggestions, and then it can be accepted for publishing. Please add more references about search and rescue (I do not recommend here, please add them by yourselves). More and more other like synthetic aperture radar ocean observation tools can also be considered to be added. Then, the paper will become more complete.

Author Response

Many thanks to the Reviewer for valuable remarks and comments. We have taken into account all suggestions and proposals.

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

accept as it is.