Intelligent Radar Jamming Recognition in Open Set Environment Based on Deep Learning Networks
Round 1
Reviewer 1 Report
I think it is important to strengthen the summary, which is the most important tool to hook the reader.
This section could contain the following structure;
1.- General and Specific Background; presentation of the scientific area to be treated, starting from the most general until reaching the specific topic that the document will deal with. If you use jargon it is necessary to explain it clearly, although I personally prefer full explanations.
2.- Knowledge gap:
What specific question do you want to solve in your work?
3.- What is shown: Indicate what is being sought in your experiments and through what methods, answering the question that was asked in the knowledge gap.
4.- Experimental approach and results; Provide a high-level description of your most important methods and results, in a way that explains how the initial claim was answered.
5.- Implications: Describe how your findings influence our understanding of the relevant field and/or its implications for future studies.
Also, in the Interference Signal Analysis section, the paragraph from line 103 to 116:
Eleven types of common interference signals are simulated to verify the validity of the 103 interference recognition models proposed in this work; "The three types of suppression interference include noise amplitude modulation (AM) interference, noise frequency modulation (FM) interference, and radio frequency (RF) noise interference. Meanwhile, deceptive interference includes Range Gate Pull-Off (RGPO) interference, Rate Gate Pull-Off (VGPO) interference, Spread Spectrum Interference (SMSP), Interrupted Sampling Repeater interference (ISRJ) and cut and interleave (C&I) interference. Among these misleading interferences, RGPO and VGPO interferences can produce a single false target with the wrong distance or speed . SMSP, ISRJ and C&I interference can produce multiple spurious targets by modulating and retransmitting the stored signal in the digital RF memory system . Composite interference includes RGPO interference and AM noise, RGPO interference and FM noise, and RGPO interference and RF noise, which are additive interference. These 11 types of interference signals are simulated according to the generation mechanism [1-5], and the simulation results are shown in Section 4.
It might be more understandable for readers in a table of the models proposed in his work.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This paper proposes a jamming recognition technique using deep learning approach. The paper is well-written with adequate result. However, the paper lacks to report some of the interesting applications of radar remote respiration sensing technology such as identity authentication, security and so on. This paper lacks to report the most recent literature of the potential application of this particular work. Please add those literature to make the paper stronger and readable to the general audience. Please also try to discuss a little bit about the potential application areas of jamming technique. The relevant literature is attached below:
1. ‘Identity Authentication in Two-Subject Environments Using Microwave Doppler Radar and Machine Learning Classifiers”-IEEE Transactions on Microwave Theory and Technique.
2. BreathID: Radar’s New Role in Biometrics, IEEE Aerospace and Electronic Systems Magazine.
3. Radar-Based Non-Contact Continuous Identity Authentication, MDPI Remote Sensing
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper is a very interesting experimental study on OSR considering Radar Jamming. Multiple experiments are carried out in two settings which provides a clear result. I have no major concern with this paper.
The method itself is not novel, but the whole OSR field is very new and new applications are always interesting.
A posterior probability estimator and a decision threshold can be problematic, as pointed out in the paper below, how did you mitigate the issues?
https://aclanthology.org/N16-1061/
A very interesting approach is described to the above problem is the paper below:
https://dl.acm.org/doi/10.5555/3327546.3327590
In the OSR based on confidence score, the author should split the training sample multiple times with a variable number of known and unknown classes each time, so that the impact of the number of training classes can be estimated.
The authors can reduce the explanation and text of known concepts, such as residual module (figure 14), batch normalization (figure 15), and so on.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 4 Report
This paper investigates intelligent radar jamming recognition in open set environment based on deep learning networks. . The topic of this paper is interesting and it is well written. I think that the quality of this paper can be improved if the authors address the following aspects:
1 The main contributions of this paper should be further summarized and clearly demonstrated. This reviewer suggests the authors exactly mention what is new compared with existing approaches and why the proposed approach is needed to be used instead of the existing methods.
2 How scalable is the proposed approach?
3 The computational cost of the proposed approach isn’t discussed. The approach should be computationally efficient to be used in practical applications.
4 What are the limitations of the presented method in practical applications?
5 Use of deep learning networks for pattern classification is an important item in the presented jamming recognition method. The following related works can be cited to improve the literature review: “Detection of false data injection attacks in smart grid: A secure federated deep learning approach”, “A deep-learning intelligent system incorporating data augmentation for short-term voltage stability assessment of power systems”.
6 Although the manuscript is well written in terms of English, there are some (very few, indeed) grammatical and expression errors. It is suggested to proofread the paper.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The paper can be accepted in its current form.
Reviewer 4 Report
Thanks to the careful revision and detailed response made by the authors. All my concerns have been well addressed, and the revised manuscript has been much improved. I think this paper deserves to be published in its current form.
