Pith. sign in

REVIEW 4 major objections 6 minor 47 references

A Unified Anti-Jamming Design in Complex Environments Based on Cross-Modal Fusion and Intelligent Decision-Making

T0 review · 4 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper argues that radar jamming type recognition and anti-jamming waveform selection should be designed as one closed-loop framework, and reports that its fused recognizer reaches 95.45% overall accuracy while its DQN converges…

desk verdict A plausible recognition-to-countermeasure integration whose DQN half is over-claimed and under-tested. read the letter →

arxiv 2506.07532 v1 pith:5HZALQVK submitted 2025-06-09 eess.SP

Huake Wang , Xudong Han , Bairui Cai , Guisheng Liao , Yinghui Quan This is my paper
classification eess.SP
keywords radaranti-jammingjammingrecognitioncross-modalfusionDeepQ-Network(DQN)time-frequencyanalysisactive-passiveDRFMfrequency-agile
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that radar jamming type recognition and anti-jamming waveform selection are two halves of one problem and should be built as a single closed loop, not as separate modules. It proposes a framework in which STFT and SPWVD time-frequency images are fused by a hierarchical ResNet, combined with time-domain features through an attention-based cross-modal module, and classified into three jamming types: aiming jamming, range false target jamming, and range dense false target jamming. The recognition result then feeds a Deep Q-Network that picks suppression waveforms, with hand-crafted reward functions tailored to each jamming type. In the paper's simulations the unified recognizer reaches 95.45% overall accuracy, beating SVM, VGG-16, and 2D-CNN, and the DQN converges around episode 600 with smaller fluctuations than SARSA. A reader should care because this is the integration pattern a practical cognitive radar would need under digital radio frequency memory (DRFM) jamming.

What carries the argument

The two load-bearing mechanisms are the cross-modal fusion module and the DQN strategy network with its reward functions. The fusion module lets two time-frequency representations, STFT and SPWVD, processed by a hierarchical ResNet and concatenated, be reweighted against time-domain features through a self-attention block: query, key, and value vectors are computed, and a global-average-pooling plus Sigmoid step produces channel weights that rescale the original features. That reweighted representation is what carries the recognition accuracy. On the decision side, the DQN selects discrete frequency-agile waveforms, uses an online and target Q-network with experience replay, and is driven by the paper's three hand-coded rewards: an SINR-threshold reward with values 30 and -100 for AJ, a frequency-hopping reward with a hopping cost for RFTJ, and a cover-pulse timing reward for RDFTJ. Coupling happens because the jamming type from the recognizer and the jamming parameters from passive radar enter the DQN state vector.

What would settle it

Run the trained DQN against a hardware-in-the-loop or high-fidelity simulated jammer and measure post-suppression detection probability, SINR, and angle-estimation error across episodes; if the policies that earn high simulated reward do not systematically improve those measured outcomes over fixed or random frequency-hopping policies, the reported convergence is an artifact of the hand-coded rewards rather than a real anti-jamming capability.

Watch

Extended reading notes

Core claim

The paper's central claim is that combining multi-modal jamming recognition with deep-reinforcement-learning waveform selection in one framework improves both stages at once. The recognition branch extracts two time-frequency representations, concatenates deep features from a hierarchical ResNet, and fuses them with time-domain features using a multi-head attention mechanism before a three-class Softmax classifier. The decision branch models anti-jamming as a DQN over discrete waveform actions, with an online network, a target network, and experience replay; its state includes the recognition result plus passive-radar parameter estimates. According to the reported results, this integrated design raises overall recognition accuracy to 95.45% compared with 91.70%, 93.69%, and 93.13% for SVM, VGG-16, and 2D-CNN, and its reward curves converge near episode 600 with visibly less oscillation than SARSA across AJ, RFTJ, and RDFTJ scenarios.

Load-bearing premise

The load-bearing premise is that the hand-coded reward functions in Eqs. (12)-(14) correctly encode what it means for a real radar to defeat AJ, RFTJ, and RDFTJ, so that maximizing them in simulation transfers to physical suppression performance.

Editorial extensions

If this is right

  • A radar that adopts this framework can identify AJ, RFTJ, or RDFTJ and switch to the matching suppression waveform in the same operational loop, instead of treating recognition and countermeasure selection as separate steps.
  • The reported convergence around episode 600 with small reward fluctuation suggests the DQN policy can reach a stable anti-jamming behavior within a few hundred simulated jamming episodes.
  • Because passive-radar broadband estimates supply the carrier frequency and bandwidth that narrowband active radar measures poorly, the framework's decisions are conditioned on more precise jamming parameters than active-only systems.
  • Across training-set sizes, the multi-modal recognizer reportedly keeps higher overall accuracy, recall, precision, and F1 than single-modal baselines, which is the regime where real collected jamming data are scarce.
  • In the constructed application scenario with a target at 5 degrees and a jammer at 2 degrees, the selected suppression waveforms restore usable monopulse angle estimation, indicating the anti-jamming choices preserve tracking information.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • An extension the paper leaves implicit is feeding softmax probabilities rather than the hard jamming label into the DQN state; that would let the decision network gracefully handle recognition uncertainty instead of committing to a possibly wrong class.
  • The reward-calibration question is the natural next experiment: the hand-set thresholds, the hopping cost, and the cover-pulse window should be checked against detection probability or receiver-operating-characteristic curves in a hardware-in-the-loop test, since maximizing simulated rewards may not track physical detection performance.
  • Ablating the attention weighting in the cross-modal module would isolate how much of the 95.45% overall accuracy comes from cross-modal fusion rather than from the larger network capacity of the hierarchical ResNet.
  • Because all simulations use a fixed carrier frequency, bandwidth, SNR, and JNR range, testing with parameter drift across episodes would reveal whether the learned DQN policy generalizes or is tuned to the simulated grid.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The manuscript proposes a unified radar anti-jamming framework that combines jamming-type recognition with DQN-based anti-jamming waveform decision-making. Recognition uses STFT and SPWVD time-frequency representations fused through an H-ResNet and an attention-based cross-modal fusion module with time-domain features. The recognized jamming type, together with passive-radar parameter estimates, is fed into a DQN that selects frequency-hopping or cover-pulse waveforms. The authors evaluate recognition accuracy against SVM, VGG-16, and 2D-CNN baselines and compare the decision network against SARSA in terms of reward convergence and stability.

Significance. If the claims are established, the proposed closed-loop recognition-plus-decision framework is a useful step toward practical cognitive radar electronic warfare, and the combination of cross-modal feature fusion with active-passive radar information is a plausible architecture. The recognition experiment reports a large margin in overall accuracy (Table VII), and the paper addresses a relevant gap by linking recognition output to downstream waveform selection. However, the current evidence is incomplete: the recognition comparisons lack statistical robustness, the reported improvement numbers are internally inconsistent, and the decision-making comparison relies entirely on the hand-crafted reward that the DQN is trained to maximize, with no external anti-jamming performance metric. The paper does not provide code, trained models, or a reproducible experimental protocol, which limits verification.

major comments (4)
  1. [Section IV.B and Section IV.D] The dataset splitting is stated inconsistently: Section IV.B says the proportion of datasets used for training, testing, and validation is 0.3:0.6:0.1, while Section IV.D says Table VII reports results with 80% training samples. Because all recognition claims depend on the exact training/evaluation protocol, the authors must specify the precise split, the number of samples per class, and whether the 0.3:0.6:0.1 split or an 80% split was used for the reported metrics.
  2. [Abstract and Table VII] The abstract reports improvements of 1.41%, 2.5%, and 14.51% in overall accuracy compared to SVM, VGG-16, and 2D-CNN, but Table VII and Section IV.D report improvements of 3.75%, 1.76%, and 2.32%, respectively. The two sets of numbers are inconsistent and the abstract's ordering and magnitudes appear to be erroneous. This must be corrected, because the abstract is the primary statement of the paper's contribution.
  3. [Section III.B.3, Eqs. (12)-(14), and Fig. 12] The central decision-making claim of faster convergence and greater stability than SARSA is supported only by the average-reward curves in Fig. 12, where the reward is computed from the hand-crafted functions in Eqs. (12)-(14). Since the DQN is trained to maximize exactly these rewards, convergence on this metric is not an independent measure of anti-jamming success. The paper should report external metrics such as post-suppression SINR, detection probability, false-target rejection rate, or monopulse angle-estimation error for the learned policies, and should calibrate the thresholds and costs (30, -100, and c) against a physical or high-fidelity radar model. In addition, the comparison with SARSA does not state whether SARSA was given the same state vector (including the FDJR jamming-type output and passive-radar parameters) and the same reward function; without this parity, Fig. 12 may reflect a difference in privileged information rather than an algorithmic advantage.
  4. [Section IV.D and Fig. 10] The recognition results are reported as a single confusion matrix and single-point metrics with no error bars, no multiple random seeds, and no statistical significance testing. Given the class distribution visible in Fig. 10 (e.g., RDFT has far more test samples than AJ or RFT), the claimed margins of 1.76-3.75% over baselines could be within run-to-run variability. The authors should provide mean and standard deviation over repeated runs or k-fold cross-validation, and report per-class support counts.
minor comments (6)
  1. [Section IV.D] The text refers to 'Table 4' when it should refer to Table VII, and contains the typo 'butter than' instead of 'better than'.
  2. [Section IV.B and Tables IV-V] Tables IV and V are presented after the text says 'the backbone network structure ... is shown in Table II and Table III'; the table numbering should be aligned.
  3. [Section I, Contribution 2] The abbreviation 'DRFTJ' is used inconsistently; the paper elsewhere uses RDFTJ. Please standardize all jamming-type acronyms.
  4. [Fig. 12] The subfigure captions duplicate the labels (a)-(c) for both SARSA and the proposed algorithm, which makes it hard to identify which curve corresponds to which method; use distinct labels.
  5. [Section III.A, Eq. (9)] The displayed definition of the ReLU activation is garbled; please rewrite it in standard mathematical notation.
  6. [Section III.A and Reference [36]] The H-ResNet module is cited to DenseNet [36], but the text does not explain the relationship; either cite a direct source or clarify how the module is derived from DenseNet.

Circularity Check

2 steps flagged · score 4.0 of 10

Recognition results are self-contained, but the DQN-vs-SARSA anti-jamming claim is evaluated by the same hand-coded reward the DQN is trained to maximize.

  1. self definitional [Section III.B.3, Eqs. (12)-(14); Section IV.C; Fig. 12]
    "In decision-making network based on deep reinforcement learning to evaluate the performance we use average reward and convergence speed. ... As observed from the reward function curves, in SARSA algorithm, convergence under AJ is not reached until approximately episode 800 ... The proposed method also achieves convergence around episode 600 ... The result indicates that the proposed method demonstrates superior robustness and stability compared to the baseline."

    The only reported DQN-vs-SARSA evidence is Fig. 12, which plots average reward. Eqs. (12)-(14) hand-code the desired anti-jamming behavior: reward 30 if the radar is unjammed or if the cover pulse lasts beyond the jammer's observation window, a high reward for frequency hops that are integer multiples of Δf, and -100 for full jamming. DQN is trained to maximize exactly this reward via the Bellman target of Eq. (10). Therefore, 'faster reward convergence and more stable strategy generation' is a statement about the optimized objective itself, not an independent measure of jamming suppression.

  2. self definitional [Section III.B.3, Eq. (14); Fig. 13(c); Fig. 15]
    "For RDFTJ, the radar typically employs a cover-pulse anti-jamming waveform. The single-step game reward function is expressed as follows: ... where t_cheat represents the time at which the radar transmits the cover pulse, and t_observe represents the observation window duration of the jammer."

    The reward is defined so that the agent receives 30 exactly when the radar's cover pulse is transmitted after/longer than the jammer's observation window, i.e., the correct anti-RDFTJ action is already encoded in the reward. The subsequent policy visualization (Fig. 13) and the claim that the decision network 'successfully mitigat[es] the interference' (Fig. 15) therefore demonstrate that the learner followed the pre-coded reward rule, rather than providing independent evidence that the selected waveform suppresses dense false targets.

full rationale

The jamming-recognition half of the paper is not circular: FDJR is evaluated against SVM, VGG-16, and 2D-CNN on a generated three-class dataset with OA, recall, precision, and F1 (Table VII, Fig. 11), so those accuracy gains are empirically self-contained. No load-bearing self-citation chain is present; background citations to prior group work (e.g., [17]) are not the derivation. The circularity is confined to the anti-jamming decision claim. Section IV.C declares decision-making performance to be average reward and convergence speed, and Fig. 12 compares DQN with SARSA on exactly the hand-crafted reward functions of Eqs. (12)-(14). Since those rewards already encode the desired frequency-hopping and cover-pulse behaviors, and DQN is trained to maximize them, reporting convergence of the same curves as 'anti-jamming performance' is a self-definitional evaluation. The absence of an external anti-jamming metric (detection probability, post-suppression SINR, or false-target rejection rate) for the learned policies means the DQN-vs-SARSA superiority claim rests on the optimized proxy itself. This is a partial circularity (score 4), not a total one, because the recognition contribution remains independently benchmarked and the DQN/SARSA comparison is at least a genuine comparison of two learners on the same reward.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper contributes an engineering architecture and an evaluation on a private simulation. Its central claims rest on hand-selected simulation parameters, hand-designed DQN rewards, and standard deep learning assumptions, none of which are independently validated. The ledger lists the main free parameters and domain assumptions that the reported performance depends on.

free parameters (4)
  • DQN reward thresholds and costs = 30, -100, c, Delta-f; c and Delta-f unspecified
    Chosen by hand in Eqs. (12) to (14) to shape anti-jamming behavior; no sensitivity analysis is provided, and these values drive the DQN convergence claims.
  • DQN hyperparameters = not reported
    Learning rate, discount factor, replay buffer size, mini-batch size, and epsilon-greedy schedule are absent from Section IV, yet they are required to reproduce Fig. 12.
  • Recognition network hyperparameters = dropout 0.4, 0.5; repeats 3, 5; input 224x224
    Chosen for the architecture in Tables IV and V with no ablation showing that these choices are optimal or robust.
  • Simulation parameters = carrier 1 GHz, bandwidth 28 MHz, pulse 10 us, PRI 50 us, SNR 10 dB, JNR 10 to 35 dB, angles 5 and 2 degrees
    Hand-set in Section IV.A to generate the private dataset; the dataset is not released, so the numbers cannot be checked externally.
assumptions (5)
  • domain assumption Jamming signal models in Eqs. (2) to (4) accurately represent AJ, RFTJ, and RDFTJ in a DRFM environment.
    All training and evaluation data are simulated from these equations in Section IV.A; if real jammers differ, recognition results do not transfer.
  • domain assumption STFT and SPWVD time-frequency representations provide sufficient discriminative information for the three jamming classes.
    Used as inputs to the recognition network; no theoretical guarantee or comparison against alternative TF transforms is given.
  • standard math DQN with the Bellman target in Eq. (10) converges to a good policy when rewards are as in Eqs. (12) to (14).
    Standard RL background; convergence is empirical, not proven for this non-stationary jamming environment.
  • domain assumption Passive radar provides accurate jamming parameters, including carrier frequency, bandwidth, and angle, to the decision network.
    Invoked in Section III.B.4; estimation errors are not modeled, but the strategy's effectiveness depends on these parameters.
  • domain assumption Active-passive fusion is beneficial because the passive wideband channel characterizes jamming parameters more precisely.
    Stated in Section III.B.4 without quantitative support or comparison against an active-only configuration.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Unified Anti-Jamming Design in Complex Environments Based on Cross-Modal Fusion and Intelligent Decision-Making." pith.science (2026). https://pith.science/paper/5HZALQVK

@misc{pith2026250607532,
  author       = {Pith},
  title        = {Pith review of: A Unified Anti-Jamming Design in Complex Environments Based on Cross-Modal Fusion and Intelligent Decision-Making},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5HZALQVK}},
  note         = {Machine review of arXiv:2506.07532}
}
read the original abstract

With the rapid development of radar jamming systems, especially digital radio frequency memory (DRFM), the electromagnetic environment has become increasingly complex. In recent years, most existing studies have focused solely on either jamming recognition or anti-jamming strategy design. In this paper, we propose a unified framework that integrates interference recognition with intelligent anti-jamming strategy selection. Specifically, time-frequency (TF) features of radar echoes are first extracted using both Short-Time Fourier Transform (STFT) and Smoothed Pseudo Wigner-Ville Distribution (SPWVD). A feature fusion method is then designed to effectively combine these two types of time-frequency representations. The fused TF features are further combined with time-domain features of the radar echoes through a cross-modal fusion module based on an attention mechanism. Finally, the recognition results, together with information obtained from the passive radar, are fed into a Deep Q-Network (DQN)-based intelligent anti-jamming strategy network to select jamming suppression waveforms. The key jamming parameters obtained by the passive radar provide essential information for intelligent decision-making, enabling the generation of more effective strategies tailored to specific jamming types. The proposed method demonstrates improvements in both jamming type recognition accuracy and the stability of anti-jamming strategy selection under complex environments. Experimental results show that our method achieves superior performance compared to Support Vector Machines (SVM), VGG-16, and 2D-CNN methods, with respective improvements of 1.41%, 2.5%, and 14.51% in overall accuracy. Moreover, in comparison with the SARSA algorithm, the designed algorithm achieves faster reward convergence and more stable strategy generation.

Figures

Figures reproduced from arXiv: 2506.07532 by the authors.

Figure 1
Figure 1. Time domain waveforms and time-frequency spectrograms of the three interference types. (a) AJ. (b) AJ-STFT. (c) AJ-SPWVD. (d) RFT. (e) RFT-STFT. (f) RFT-SPWVD. (g) RDFT. (h) RDFT-STFT. (i) RDFT-SPWVD. 2) Smoothed Pseudo Wigner-Ville Distribution (SPWVD): Unlike the STFT, which suffering from fixed resolution limitations due to windowing. The SPWVD [35] employs independent smoothing operations in time and frequency. … view at source ↗
Figure 3
Figure 3. TF Feature extraction and fusion process [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. H-ResNet module 2) Cross-Modal Fusion Network: In our cross-modal fusion model, the features involved in the fusion process including time domain features and the aforementioned fused features. First the temporal features are processed through 77  convolutional layer and 33  max-pooling layer. Then they pass through a residual block composed of 11 and 33  convolutional layers. The 11 convolutional kernels perfo… view at source ↗
Figures from the paper (11 more)
Figure 5
Figure 5. Figure 5: Overall flowchart of the proposed jamming recognition algorithm. After processing through convolutional layers, the input data are decomposed into multiple feature vectors. Subsequent residual mapping operations convert these features into query vectors Qq = i , key …
Figure 8
Figure 8. Figure 8: Framework of the anti-jamming decision-making strategy This Q-network model employs dual network: an online Q￾network and target Q-network, following the fixed target network mechanism proposed in the original DQN framework [22] [25]. The two networks share the same ar…
Figure 6
Figure 6. Figure 6: Cross fusion module In this work, the classifier is implemented with a multilayer perception (MLP) architecture [40], employing ReLU activation functions and dropout layers for overfitting mitigation. The ReLU activation functions expressed as: / ( ) (0, ) t tf F x max…
Figure 7
Figure 7. Figure 7: Layout of jamming deployment Jamming type J1 represents AJ, J2 represents RFTJ, and J3 represents RDFTJ. Jamming strategy Ⅰ indicates sequential switching among the three jamming types, with one type switched per CPI. Strategy Ⅱ denotes palindromic-sequence switching o…
Figure 9
Figure 9. Figure 9: Framework of the Active–Passive radar cooperative framework IV. EXPERIMENTAL RESULTS In this section, first we provided a detailed overview of the jamming dataset, followed by a description of the experimental settings for the proposed Fusion-based Deep Jamming Recogni…
Figure 10
Figure 10. Figure 10: Confusion matrix of the proposed jamming recognition method (a) (b) (c) (d) [PITH_FULL_IMAGE:figures/full_fig_p008_10.png]
Figure 11
Figure 11. Figure 11: Evolution of the evaluation metrics according to the training set size. (a) OA. (b) Recall. (c) Precision. (d) F1. 235 0 0 0 185 59 0 5 746 AJ RFT RDFT RDFT RFT AJ [PITH_FULL_IMAGE:figures/full_fig_p008_11.png]
Figure 12
Figure 12. Figure 12: Average Rewards of the SARSA and proposed algorithm. (a) AJ-SARSA. (b) RFT-SARSA. (c) RDFT-SARSA. (a) AJ-Proposed algorithm. (b) RFT-Proposed algorithm. (c) RDFT-Proposed algorithm. (a) (b) (c) [PITH_FULL_IMAGE:figures/full_fig_p009_12.png]
Figure 13
Figure 13. Figure 13: Policy output of the anti-jamming strategy design network (a) AJ. (b) RFT. (c) RDFT. As observed from the reward function curves, in SARSA algorithm, convergence under AJ is not reached until approximately episode 800, and the training phase exhibits significant oscil…
Figure 14
Figure 14. Figure 14: Selected radar waveforms under different jamming scenarios. (a) Sequential switching jamming . (b) Palindromic-sequence switching jamming. (c) Random switching jamming. (a) (b) (c) [PITH_FULL_IMAGE:figures/full_fig_p010_14.png]
Figure 15
Figure 15. Figure 15: Monopulse Angle Estimation Results After Anti-Jamming. (a) AJ. (b) RFT. (c) RDFT. Furthermore, in a scenario where both the target angle is 5° and the jammer angle is 2°, the decision network outputs corresponding suppression measures, successfully mitigating the inte…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

47 extracted references · 46 canonical work pages

  1. [1]

    SAR -ECCM using phase -perturbed LFM chirp signals and DRFM repeat jammer penalization[C]//IEEE International Radar Conference, 2005

    Soumekh M. SAR -ECCM using phase -perturbed LFM chirp signals and DRFM repeat jammer penalization[C]//IEEE International Radar Conference, 2005. IEEE, 2005: 507-512

    work page 2005

  2. [2]

    An intelligent jamming strategy design method against frequency agility radar[C]//2023 IEEE International Radar Conference (RADAR)

    Yang B, Li K, Jiu B, et al. An intelligent jamming strategy design method against frequency agility radar[C]//2023 IEEE International Radar Conference (RADAR). IEEE, 2023: 1-6

    work page 2023

  3. [3]

    A novel deceptive jamming approach against frequency diverse array radar[J]

    Tan M, Wang C, Xue B, et al. A novel deceptive jamming approach against frequency diverse array radar[J]. IEEE Sensors Journal, 2020, 21(6): 8323-8332

    work page 2020

  4. [4]

    An antideceptive jamming method for multistatic synthetic aperture radar based on collaborative localization and spatial suppression[J]

    Wang W, Wu J, Pei J, et al. An antideceptive jamming method for multistatic synthetic aperture radar based on collaborative localization and spatial suppression[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2020, 13: 2757-2768

    work page 2020

  5. [5]

    A survey of radar ECM and ECCM[J]

    Neng-Jing L, Yi-Ting Z. A survey of radar ECM and ECCM[J]. IEEE Transactions on Aerospace and Electronic Systems, 1995, 31(3): 1110- 1120

    work page 1995

  6. [6]

    Frequency -agile radar signal processing[J]

    Waters W M, Linde G J. Frequency -agile radar signal processing[J]. IEEE Transactions on aerospace and electronic systems, 1979 (3): 459- 464

    work page 1979

  7. [7]

    Kogon S M, Holder E J, Williams D B. Mainbeam jammer suppression using multipath returns[C]//Conference Record of the Thirty -First 0 0.5 1 1.5 2 2.5 1 2 3 4 5 6 7 8 9 strategy and behavio time sequential switching jamming jamming stra jammer radar 0 0.5 1 1.5 2 2.5 1 2 3 4 5 6 7 8 9 strategy and behavio time palindromic-sequence switching jamming jamming...

    work page 1997

  8. [8]

    Target Detection and Interference Mitigation in Future AI -Based Radar Systems[C]//2021 IEEE Radar Conference (RadarConf21)

    Deng H, Himed B. Target Detection and Interference Mitigation in Future AI -Based Radar Systems[C]//2021 IEEE Radar Conference (RadarConf21). IEEE, 2021: 1-4

    work page 2021

Show all 47 references
  1. [9]

    Richter Y, Balal N, Gerasimov J, et al. Deep Learning -Based Radar Processing: Simultaneous Target Classification, Activity Detection, and Range Estimation with Micro -Doppler Radar[C]//2024 IEEE International Conference on Microwaves, Communications, Antennas, Biomedical Engi...

    2024

  2. [11]

    Recognition of radar compound jamming based on convolutional neural network[J]

    Zhou H, Wang L, Guo Z. Recognition of radar compound jamming based on convolutional neural network[J]. IEEE Transactions on Aerospace and Electronic Systems, 2023, 59(6): 7380-7394

    2023

  3. [12]

    Deep learning and recognition of radar jamming based on CNN[C]//2019 12th international symposium on computational intelligence and design (ISCID)

    Liu Q, Zhang W. Deep learning and recognition of radar jamming based on CNN[C]//2019 12th international symposium on computational intelligence and design (ISCID). IEEE, 2019, 1: 208 - 212

    2019

  4. [13]

    Radar Signal Recognition Based on Deep Feature Fusion of Multiple Time -frequency Images[C]//2023 International Conference on Wireless Communications and Signal Processing (WCSP)

    Hu Z, Li H, Tang Z, et al. Radar Signal Recognition Based on Deep Feature Fusion of Multiple Time -frequency Images[C]//2023 International Conference on Wireless Communications and Signal Processing (WCSP). IEEE, 2023: 377-383

    2023

  5. [14]

    A Cross -Modality Contrastive Learning Method for Radar Jamming Recognition[J]

    Dong G, Wang Z, Liu H. A Cross -Modality Contrastive Learning Method for Radar Jamming Recognition[J]. IEEE Transactions on Instrumentation and Measurement, 2025

    2025

  6. [15]

    Convolutional neural network -based radar jamming signal classification with sufficient and limited samples[J]

    Shao G, Chen Y, Wei Y. Convolutional neural network -based radar jamming signal classification with sufficient and limited samples[J]. IEEE Access, 2020, 8: 80588-80598

    2020

  7. [16]

    Rcfusion: Fusing 4-d radar and camera with bird’s-eye view features for 3-d object detection[J]

    Zheng L, Li S, Tan B, et al. Rcfusion: Fusing 4-d radar and camera with bird’s-eye view features for 3-d object detection[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 1-14

    2023

  8. [17]

    Radar deception jamming recognition based on weighted ensemble CNN with transfer learning[J]

    Lv Q, Quan Y, Feng W, et al. Radar deception jamming recognition based on weighted ensemble CNN with transfer learning[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 60: 1-11

    2021

  9. [18]

    Adversarial deep learning for cognitive radio security: Jamming attack and defense strategies[C]//2018 IEEE international conference on communications workshops (ICC Workshops)

    Shi Y, Sagduyu Y E, Erpek T, et al. Adversarial deep learning for cognitive radio security: Jamming attack and defense strategies[C]//2018 IEEE international conference on communications workshops (ICC Workshops). IEEE, 2018: 1-6

    2018

  10. [19]

    Intelligent radar jamming recognition in open set environment based on deep learning networks[J]

    Zhou Y, Shang S, Song X, et al. Intelligent radar jamming recognition in open set environment based on deep learning networks[J]. Remote Sensing, 2022, 14(24): 6220

    2022

  11. [20]

    Jamming recognition based on AC - VAEGAN[C]//2020 15th IEEE International Conference on Signal Processing (ICSP)

    Tang Y, Zhao Z, Ye X, et al. Jamming recognition based on AC - VAEGAN[C]//2020 15th IEEE International Conference on Signal Processing (ICSP). IEEE, 2020, 1: 312-315

    2020

  12. [21]

    Jamming pattern recognition using spectrum waterfall: A deep learning method[C]//2019 IEEE 5th international conference on computer and communications (ICCC)

    Cai Y, Shi K, Song F, et al. Jamming pattern recognition using spectrum waterfall: A deep learning method[C]//2019 IEEE 5th international conference on computer and communications (ICCC). IEEE, 2019: 2113-2117

    2019

  13. [22]

    Cooperative Frequency Scheduling for Netted Agile Radars via Decentralized Q -Learning[J]

    Li S, Liao M, Xiong K, et al. Cooperative Frequency Scheduling for Netted Agile Radars via Decentralized Q -Learning[J]. IEEE Transactions on Instrumentation and Measurement, 2025

    2025

  14. [23]

    Intelligent anti -jamming based on deep reinforcement learning and transfer learning[J]

    Janiar S B, Wang P. Intelligent anti -jamming based on deep reinforcement learning and transfer learning[J]. IEEE Transactions on Vehicular Technology, 2024, 73(6): 8825-8834

    2024

  15. [24]

    Radar anti -jamming decision -making method based on DDPG -MADDPG algorithm[J]

    Wei J, Wei Y, Yu L, et al. Radar anti -jamming decision -making method based on DDPG -MADDPG algorithm[J]. Remote Sensing, 2023, 15(16): 4046

    2023

  16. [25]

    Deep Reinforcement Learning -Based Anti- Jamming Approach for Fast Frequency Hopping Systems[J]

    Cheng S, Ling X, Zhu L. Deep Reinforcement Learning -Based Anti- Jamming Approach for Fast Frequency Hopping Systems[J]. IEEE Open Journal of the Communications Society, 2025

    2025

  17. [26]

    Jiang W, Wang Y, Li Y, et al. An intelligent anti -jamming decision- making method based on deep reinforcement learning for cognitive radar[C]//2023 26th International Conference on Computer Supported Cooperative Work in Design (CSCWD). IEEE, 2023: 1662-1666

    2023

  18. [27]

    Deep -Q reinforcement learning for fairness in multiple -access cognitive radio networks[C]//2022 IEEE Wireless Communications and Networking Conference (WCNC)

    Ali Z, Rezki Z, Sadjadpour H. Deep -Q reinforcement learning for fairness in multiple -access cognitive radio networks[C]//2022 IEEE Wireless Communications and Networking Conference (WCNC). IEEE, 2022: 2023-2028

    2022

  19. [28]

    Zhu J, Ruan H, Han L, et al. Adaptive Selection of Cognitive Radar Waveforms Based on Improved Deep Q -Learning Network[C]//2023 3rd International Conference on Computer Science, Electronic Information Engineering and Intelligent Control Technology (CEI). IEEE, 2023: 515-522

    2023

  20. [30]

    Research on cooperative detection technology of networked radar based on data fusion[C]//2021 2nd China International SAR Symposium (CISS)

    Jiang W, Qi Z, Ye Z, et al. Research on cooperative detection technology of networked radar based on data fusion[C]//2021 2nd China International SAR Symposium (CISS). IEEE, 2021: 1-5

    2021

  21. [31]

    dissertation, Xidian University, Xi’an, China,

    Liu Y, Research on Anti-Deceptive Jamming Methods for Multistatic Radar Systems, Ph.D. dissertation, Xidian University, Xi’an, China,

  22. [32]

    dissertation, Xidian University, Xi’an, China, 2016

    Zhao S, Research on Cooperative Anti -Deceptive Jamming Methods for Multistatic Radar Systems, Ph.D. dissertation, Xidian University, Xi’an, China, 2016. (In Chinese)

    2016

  23. [33]

    A novel signal processing technique for Ku -band automobile FMCW fully polarimetric SAR system using triangular LFM[J]

    Jung D H, Kim D H, Azim M T, et al. A novel signal processing technique for Ku -band automobile FMCW fully polarimetric SAR system using triangular LFM[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 70: 1-10

    2020

  24. [34]

    Deconvolutive short -time Fourier transform spectrogram[J]

    Lu W, Zhang Q. Deconvolutive short -time Fourier transform spectrogram[J]. IEEE Signal Processing Letters, 2009, 16(7): 576-579

    2009

  25. [35]

    Time -frequency DOA estimate algorithm based on SPWVD[C]//2005 IEEE International Symposium on Microwave, Antenna, Propagation and EMC Technologies for Wireless Communications

    Hang H. Time -frequency DOA estimate algorithm based on SPWVD[C]//2005 IEEE International Symposium on Microwave, Antenna, Propagation and EMC Technologies for Wireless Communications. IEEE, 2005, 2: 1253-1256

    2005

  26. [36]

    Densenet: Implementing efficient convnet descriptor pyramids[J]

    Iandola F, Moskewicz M, Karayev S, et al. Densenet: Implementing efficient convnet descriptor pyramids[J]. arXiv preprint arXiv:1404.1869, 2014

    2014 arXiv

  27. [37]

    Multilevel scattering center and deep feature fusion learning framework for SAR target recognition[J]

    Liu Z, Wang L, Wen Z, et al. Multilevel scattering center and deep feature fusion learning framework for SAR target recognition[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: 1-14

    2022

  28. [38]

    Music Emotion Classification Model Based on Multi Feature Image Fusion[C]//2024 First International Conference on Software, Systems and Information Technology (SSITCON)

    Wu M. Music Emotion Classification Model Based on Multi Feature Image Fusion[C]//2024 First International Conference on Software, Systems and Information Technology (SSITCON). IEEE, 2024: 1-6

    2024

  29. [39]

    PFAF -Net: Pyramid feature network for multimodal fusion[J]

    Raza A, Huo H, Fang T. PFAF -Net: Pyramid feature network for multimodal fusion[J]. IEEE Sensors Letters, 2020, 4(12): 1-4

    2020

  30. [40]

    Mlp-mixer: An all-mlp architecture for vision[J]

    Tolstikhin I O, Houlsby N, Kolesnikov A, et al. Mlp-mixer: An all-mlp architecture for vision[J]. Advances in neural information processing systems, 2021, 34: 24261-24272

    2021

  31. [41]

    Reinforcement learning for compressed - sensing based frequency agile radar in the presence of active interference[J]

    Wang S, Liu Z, Xie R, et al. Reinforcement learning for compressed - sensing based frequency agile radar in the presence of active interference[J]. Remote Sensing, 2022, 14(4): 968

    2022

  32. [42]

    Research on efficient reinforcement learning for adaptive frequency-agility radar[J]

    LI Xinzhi and DONG Shengbo. Research on efficient reinforcement learning for adaptive frequency-agility radar[J]. Sensors, 2021, 21(23):

    2021

  33. [43]

    Distorted SAR target recognition with virtual SVM and AP -HOG feature[C]//IET International Radar Conference (IET IRC 2020)

    Yao D, Liu Z, Li F, et al. Distorted SAR target recognition with virtual SVM and AP -HOG feature[C]//IET International Radar Conference (IET IRC 2020). IET, 2020, 2020: 734-739

    2020

  34. [44]

    Xu S, Qi H, Zhang Y, et al. Research on Advanced Detection Method of Target Spectrum of Radar Altimeter Based on VGG16-Net[C]//2023 IEEE International Conference on Signal Processing, Communications and Computing (ICSPCC). IEEE, 2023: 1-6

    2023

  35. [45]

    Deep fusion for radar jamming signal classification based on CNN[J]

    Shao G, Chen Y, Wei Y. Deep fusion for radar jamming signal classification based on CNN[J]. IEEE Access, 2020, 8: 117236-117244. Huake Wang received the B.S. degree in electronics engineering, and the Ph.D. degree in signal and information processing from Xidian University, Xi...

    2020

  36. [1963]

    degree from Guangxi University in mathematics, Guangxi, China, in 1985, and the M.S

    He received the B.S. degree from Guangxi University in mathematics, Guangxi, China, in 1985, and the M.S. degree in computer software and the Ph.D. degree in signal and information processing from Xidian University, Xi ’ an, China, in 1990, and 1992, respectively. He is curren...

    1985

  37. [2008]

    Yinghui Quan (Senior Member, IEEE) received the B.S

    His research interests include array signal p rocessing, space-time adaptive processing, radar waveform design, and airborne/space surveillance and warning radar systems. Yinghui Quan (Senior Member, IEEE) received the B.S. and Ph.D. degrees in electrical engineering from Xidi...

    2004

  38. [2021]

    He won the National Science Fund for Distinguished Young Scholar s in

    He has been a Senior Visiting Scholar with the Chinese University of Hong Kong from 1999 to 2000. He won the National Science Fund for Distinguished Young Scholar s in

    1999

  39. [7931]

    doi: 10.3390/s21237931

    doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.