Cooperative Multi-Node Jamming Recognition Method Based on Deep Residual Network
Round 1
Reviewer 1 Report
The paper is well written and very interesting to read, however I see the following minor issues that should be resolved before publishing it:
Fig 1. Is very small and should be enlarged. The word “channel” has two “n”.
Line 186. Why using an 18-layer residual network and not any other value? The reason is explained further below but should be clarified here.
Line 212. Instead of writing this long sentence “namely the hard fusion-based multi-node cooperative jamming recognition method and the soft fusion-based multi-node cooperative jamming recognition method”, abbreviate with “namely the hard and soft fusion-based cooperative jamming recognition methods”.
Line 224. Please, write only Psi (without quotation marks).
Line 249 and others. Please, leave a space between the value and the unit (10 MHz)
Figure 12 and 13 are not representing "Performance comparison of soft-fusion methods under different number of nodes" but confusion matrix of both soft-fusion and hard-fusion based methods.
Figure 12, 13 and 14. Labels "hardecisiton" and "softdecisiton". Do you mean "hard-decision" and "soft-decision"?
Figure 13 label says "Confusion matrix for test data (hardecisiton)" but in figure title "soft-fusion" is cited.
Can't see clearly differences beetween figures 12, 13 and 14. This aspect should be clarified because it is the core of the paper and from which the conclusions are drawn.
Line 373-376. Can you provide a percentage improvement (numerical) of both methods compared with the single-node method?
Please, do not capitalize the surnames of the references.
Author Response
Please see the attachments.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for Authors (will be shown to authors)
· The paper describes the attacks by malicious individuals on the wireless communication signals as the major factor threatening the viability of wireless communication.
· For solving this various anti-jamming techniques have been reported in which jamming recognition is described to be the key part in the anti-jamming techniques.
Query 1. “Jamming recognition methods can be divided into two categories based on the presence or absence of feature extraction,” disrupts the continuity due to the unstated two categories.
· The features-based category is evaluation by the performance parameters are described to be calculation of C (carrier factor coefficient), FSE (average frequency domain flatness coefficient) and other feature parameters of jamming, extraction of signal’s time-frequency plots – all such method rely on setting up the threshold values that are greatly influenced by channel environment.
· Compared to threshold-based features analysis, the convolutional Neural Network (CNN) has a multilayer network structure, which is suitable for solving image feature extraction problems and has advantages in extracting features related to special (2-dimentional) features, and so CNN are described to be bearing the potential for jamming and modulation recognition.
· The authors describe their work to be the first paper to apply a multi-node approach to the field of jamming recognition compared the previous of solving the problem of single-node.
Query 2. The work of the authors is based on deriving support for starting their work from [13], [14], and [15] – all there very old references.
· The title is supported by Figure 1 that explains the framework of the multi-node cooperative jamming recognition network consists of 78 one central node and M cooperative cognitive nodes while Equation (1) relates the communication signal, S(n), with the Jamming signal J(n), ad noise signal, v(n).
Is the subject matter presented in a comprehensive manner?
· The 15-page paper is presented with a good level of flow and the title framework is shown in Figure 1 to genuinely support the title “Cooperative multi-node jamming recognition method based on deep residual network”. The theory support and related explanation is comprehensively covered, justifying the contribution to the body of knowledge in terms of defining jamming recognition is the precondition and foundation of cognitive anti-jamming.
· The abstract has described the CNN-based anti-jamming methods as single-node methods having low jamming to signal ration (JSR) while the multi-node is claimed to be showing the single-node performance improved by 10%-15%.
· The trained network is used to classify the six malicious signals as consisting of single-tone jamming, multi-tone jamming, narrow-band jamming, broad-band jamming, comb jamming and sweep jamming,
· The single-node CNN-based performance is compared with that of Residual Network (ReNet) under different levels settings in Figure 7
Query 3. Figure 8 showing the performance comparison of ResNet and CNN under single-node method is mot described in the paper text.
· Figure 9 gives the performance curves of the hard fusion methods with different 295 numbers of co-cognitive nodes under Experiment 2, while Performance comparison of soft-fusion methods under different number of nodes is given in Experiment 3 in Figure 10.
· Figure 11 shows the Performance comparison between ResNet and CNN under soft and hard fusion methods fusion methods together under Experiment 4.
Query 4. The Conclusion very descriptive in its current form, should be made quantified and reflective of the Abstract.
Query 5. However, the readers do not see results to justify the term ‘cooperative’ in the title.
Are the references provided applicable and sufficient?
· The authors take support from twenty six (26) mostly recently from diverse journals references and including none from MDPI algorithms. The results are related to the title of “Cooperative multi-node jamming recognition method based on deep residual network”.
Query 6. References 2 and 13 are very old and be replaced by their current versions or them removed and their materials be supported from other relevant references.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments for author File: Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
