Jie Huang
Research Subjects
Jie Huang has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Dexiu Hu
Signal Processing Laboratory, Information Engineering University, Zhengzhou, China
4 shared publications - Min Xie
National Digital Switching System Engineering and Technological R&D Center, Zhengzhou, China
4 shared publications - Chuang Zhao
National Digital Switching System Engineering and Technological R&D Center, Zhengzhou, China
3 shared publications Jiantao Wang
School of Data and Target Engineering PLA Strategic Support Force Information Engineering University Zhengzhou China
3 shared publications- Yifan Sun
National Digital Switching System Engineering and Technological R&D Center, Zhengzhou, China
2 shared publications - Zewen Wang
Information Engineering University, Zhengzhou, China
2 shared publications
Latest Publications
- Open Access
- 1 January 2026
- Published by Wiley in International Journal of Aerospace Engineering
- Vol. 2026 (1)
Aircraft Control Acceleration Identification Method Based on Chaotic Adaptive Bald Eagle Search
Measurement noise in flight trajectory data degrades the accuracy of aircraft control acceleration identified by estimation algorithms such as the Kalman filter. This increased identification error constrains the reliability of subsequent tasks that rely on control acceleration information, including target recognition and trajectory prediction. This paper proposes a Chaotic Adaptive Bald Eagle Search (CABES)–based identification method. Flight trajectories are segmented using climb and heading change rates to isolate deterministic dynamic parameters, mitigating dynamic uncertainty effects. A global parametric dynamic model represents aircraft‐controlled responses through velocity‐frame acceleration components. CABES enhances optimization via chaotic mapping for global exploration and adaptive parameter balancing search efficiency. The applicability of the proposed method is verified via CRLB‐based identifiability analysis, model‐mismatch analysis, Monte Carlo simulations, and real flight trajectory experiments. CABES achieves higher acceleration identification accuracy than EKF, PSO, and BES, with superior error performance under real trajectory disturbances.
- 1 January 2026
- Published by Wiley in International Journal of Aerospace Engineering
- Vol. 2026 (1)
Open AccessAircraft Control Acceleration Identification Method Based on Chaotic Adaptive Bald Eagle Search
Measurement noise in flight trajectory data degrades the accuracy of aircraft control acceleration identified by estimation algorithms such as the Kalman filter. This increased identification error constrains the reliability of subsequent tasks that rely on control acceleration information, including target recognition and trajectory prediction. This paper proposes a Chaotic Adaptive Bald Eagle Search (CABES)–based identification method. Flight trajectories are segmented using climb and heading change rates to isolate deterministic dynamic parameters, mitigating dynamic uncertainty effects. A global parametric dynamic model represents aircraft‐controlled responses through velocity‐frame acceleration components. CABES enhances optimization via chaotic mapping for global exploration and adaptive parameter balancing search efficiency. The applicability of the proposed method is verified via CRLB‐based identifiability analysis, model‐mismatch analysis, Monte Carlo simulations, and real flight trajectory experiments. CABES achieves higher acceleration identification accuracy than EKF, PSO, and BES, with superior error performance under real trajectory disturbances.
- Open Access
- 1 January 2026
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 20 (1)
Robust LPI Radar Signal Recognition via SNR‐Guided Generative Restoration and Active Recurrent Glance‐and‐Focus
Low probability of intercept (LPI) radar signal recognition under low signal‐to‐noise ratio (SNR) remains a formidable challenge, as intercepted signals are often submerged in severe noise, leading to modulation feature ambiguity. Although generative models offer potential for signal restoration, applying them blindly can result in ‘hallucinations’ or incorrect structural guesses. To address these issues, this paper proposes a unified framework that integrates SNR‐guided generative restoration with an active recurrent glance‐and‐focus classification strategy. First, we introduce an RA‐UnetJiT restoration network, where the generative process is dynamically modulated by the input SNR via adaptive layer normalisation (AdaLN). This mechanism leverages U‐Net structural priors to strictly constrain the generation, ensuring the recovered spectrograms maintain semantic fidelity even at extremely low SNRs. Second, to address the computational redundancy of sparse spectrograms, we formulate feature extraction as a sequential decision process (SDP). A deep reinforcement learning (PPO) agent actively scans the restored image, selecting a sequence of informative high‐resolution glimpses. Finally, a transformer encoder models the spatiotemporal dependencies of these glimpses for robust classification. Experimental results on 16 types of radar emitters demonstrate that the proposed method effectively mitigates noise interference and achieves state‐of‐the‐art accuracy (e.g., at −10 dB) with significantly reduced computational cost compared to traditional sliding‐window baselines.
- 1 January 2026
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 20 (1)
Open AccessRobust LPI Radar Signal Recognition via SNR‐Guided Generative Restoration and Active Recurrent Glance‐and‐Focus
Low probability of intercept (LPI) radar signal recognition under low signal‐to‐noise ratio (SNR) remains a formidable challenge, as intercepted signals are often submerged in severe noise, leading to modulation feature ambiguity. Although generative models offer potential for signal restoration, applying them blindly can result in ‘hallucinations’ or incorrect structural guesses. To address these issues, this paper proposes a unified framework that integrates SNR‐guided generative restoration with an active recurrent glance‐and‐focus classification strategy. First, we introduce an RA‐UnetJiT restoration network, where the generative process is dynamically modulated by the input SNR via adaptive layer normalisation (AdaLN). This mechanism leverages U‐Net structural priors to strictly constrain the generation, ensuring the recovered spectrograms maintain semantic fidelity even at extremely low SNRs. Second, to address the computational redundancy of sparse spectrograms, we formulate feature extraction as a sequential decision process (SDP). A deep reinforcement learning (PPO) agent actively scans the restored image, selecting a sequence of informative high‐resolution glimpses. Finally, a transformer encoder models the spatiotemporal dependencies of these glimpses for robust classification. Experimental results on 16 types of radar emitters demonstrate that the proposed method effectively mitigates noise interference and achieves state‐of‐the‐art accuracy (e.g., at −10 dB) with significantly reduced computational cost compared to traditional sliding‐window baselines.
- 6 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. PP (1530437X) , 1
CSE-HMM-RFN: A Method for Multi-functional Radar Pulse Deinterleaving
Despite available prior knowledge, existing pulse deinterleaving algorithms for multi-functional radars (MFRs) suffer from limited robustness and high computational complexity. To address these limitations, we propose the CSE-HMM-RFN, a novel hybrid pulse deinterleaving framework. To begin with, a two-layer hierarchical hidden Markov model (HHMM) is developed that maps operational tasks to pulse groups based on MFR radiation mechanisms. Then, the temporal dynamics within and between pulse groups are characterized by their temporal intervals, enabling a decomposed deinterleaving strategy, that is, pulse group extraction followed by temporal combination search. Accordingly, CSE-HMM-RFN implements two core modules. Each module constructs hidden Markov models (HMMs) where non-zero states indicate target pulses or pulse groups, and employs a Viterbi algorithm guided by residual fence network (RFN) path weights to backtrack optimal state transition paths. Experimental results demonstrate that CSE-HMM-RFN achieves enhanced robustness and reduced computational resources compared to baseline algorithms.
- 6 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. PP (1530437X) , 1
CSE-HMM-RFN: A Method for Multi-functional Radar Pulse Deinterleaving
Despite available prior knowledge, existing pulse deinterleaving algorithms for multi-functional radars (MFRs) suffer from limited robustness and high computational complexity. To address these limitations, we propose the CSE-HMM-RFN, a novel hybrid pulse deinterleaving framework. To begin with, a two-layer hierarchical hidden Markov model (HHMM) is developed that maps operational tasks to pulse groups based on MFR radiation mechanisms. Then, the temporal dynamics within and between pulse groups are characterized by their temporal intervals, enabling a decomposed deinterleaving strategy, that is, pulse group extraction followed by temporal combination search. Accordingly, CSE-HMM-RFN implements two core modules. Each module constructs hidden Markov models (HMMs) where non-zero states indicate target pulses or pulse groups, and employs a Viterbi algorithm guided by residual fence network (RFN) path weights to backtrack optimal state transition paths. Experimental results demonstrate that CSE-HMM-RFN achieves enhanced robustness and reduced computational resources compared to baseline algorithms.
- Open Access
- 1 January 2025
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 19 (1)
Radar Signal Deinterleaving in Complex Electromagnetic Environments by a Multi‐Class Classification Perspective
Radar signal deinterleaving is challenging due to dense pulse interleaving and diverse PRI modulations. This work reframes it as a multi‐class classification problem, treating each emitter as a distinct class. Existing methods suffer from error accumulation in sequential processing or fail to integrate parallel classifier outputs effectively. To address these flaws, we propose OvR‐C‐MC, a complete one‐vs.‐rest (OvR) decomposition framework. Key innovations include (1) true multi‐class decomposition: parallel binary classifiers maintain the OvR paradigm's theoretical guarantees, avoiding error propagation in sequential binary classifiers. We integrate classifier outputs with a prioritisation mechanism to resolve conflicts, ensuring a more robust and accurate classification process than existing methods. (2) HMC‐based OvR classifier: hidden Markov chains (HMCs) form the basis of each binary classifier, enabling support for any regularised PRI modulation types through state transition property and providing a more comprehensive solution. Experimental results demonstrate that the proposed method significantly outperformed existing approaches, particularly in dense interleaved scenarios, whereas maintaining compatibility with diverse PRI modulation types. Thus, the proposed systematic perspective for radar signal deinterleaving provides robust support for radar signal reconnaissance in complex electromagnetic environments.
- 1 January 2025
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 19 (1)
Open AccessRadar Signal Deinterleaving in Complex Electromagnetic Environments by a Multi‐Class Classification Perspective
Radar signal deinterleaving is challenging due to dense pulse interleaving and diverse PRI modulations. This work reframes it as a multi‐class classification problem, treating each emitter as a distinct class. Existing methods suffer from error accumulation in sequential processing or fail to integrate parallel classifier outputs effectively. To address these flaws, we propose OvR‐C‐MC, a complete one‐vs.‐rest (OvR) decomposition framework. Key innovations include (1) true multi‐class decomposition: parallel binary classifiers maintain the OvR paradigm's theoretical guarantees, avoiding error propagation in sequential binary classifiers. We integrate classifier outputs with a prioritisation mechanism to resolve conflicts, ensuring a more robust and accurate classification process than existing methods. (2) HMC‐based OvR classifier: hidden Markov chains (HMCs) form the basis of each binary classifier, enabling support for any regularised PRI modulation types through state transition property and providing a more comprehensive solution. Experimental results demonstrate that the proposed method significantly outperformed existing approaches, particularly in dense interleaved scenarios, whereas maintaining compatibility with diverse PRI modulation types. Thus, the proposed systematic perspective for radar signal deinterleaving provides robust support for radar signal reconnaissance in complex electromagnetic environments.
- 30 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 61 (3) , 6011-6025
Radar Signal Deinterleaving Based on Hidden Markov Chains and Residual Fence Networks
For prior-knowledge-informed scenarios, this paper proposes a radar signal deinterleaving method based on hidden Markov chains and residual fence networks (RFNs) with enhanced applicability for different pulse repetition interval (PRI) types and improved performance under PRI jitter. The proposed approach accommodates the five main PRI modulation types by modeling interleaved pulse streams as hidden Markov models (HMMs). Pulse deinterleaving is transformed into a state-sequence prediction problem using HMMs and further into a path optimization problem within RFNs. This method utilizes global information for reliable sequence separation. Experimental results indicate that the method effectively deinterleaves fixed, staggered, dwell-and-switch, sliding, and wobulated sequences under unintentional jitter, significantly improving performance under intentional jitter. The practical value and applicability of the method are demonstrated, providing a new perspective for deinterleaving. The code will be open at: https://github.com/xm980426/HMC-RFN.git .
- 30 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 61 (3) , 6011-6025
Radar Signal Deinterleaving Based on Hidden Markov Chains and Residual Fence Networks
For prior-knowledge-informed scenarios, this paper proposes a radar signal deinterleaving method based on hidden Markov chains and residual fence networks (RFNs) with enhanced applicability for different pulse repetition interval (PRI) types and improved performance under PRI jitter. The proposed approach accommodates the five main PRI modulation types by modeling interleaved pulse streams as hidden Markov models (HMMs). Pulse deinterleaving is transformed into a state-sequence prediction problem using HMMs and further into a path optimization problem within RFNs. This method utilizes global information for reliable sequence separation. Experimental results indicate that the method effectively deinterleaves fixed, staggered, dwell-and-switch, sliding, and wobulated sequences under unintentional jitter, significantly improving performance under intentional jitter. The practical value and applicability of the method are demonstrated, providing a new perspective for deinterleaving. The code will be open at: https://github.com/xm980426/HMC-RFN.git .
- 23 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 21 (1545598X) , 1-5
Radar Signal Deinterleaving Based on Markov Chains in Scenarios Known a Priori
Existing deinterleaving methods are limited to fixed and staggered pulse sequences, and cannot handle dwell & switch (D&S), sliding, and wobulated pulse sequences. This study proposes a new approach that models actual sequences as Markov chains using the pulse index as the time parameter and the pulse index of the ideal sequence as the state. The transformation turns radar pulse deinterleaving into a Markov chain-based prediction problem. Since the constructed Markov chain encompasses all the PRI modulation information and is applicable to different objects, the method has great potential in practicality and applicability, providing a fresh perspective and direction to deinterleaving.
- 23 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Geoscience and Remote Sensing Letters
- Vol. 21 (1545598X) , 1-5
Radar Signal Deinterleaving Based on Markov Chains in Scenarios Known a Priori
Existing deinterleaving methods are limited to fixed and staggered pulse sequences, and cannot handle dwell & switch (D&S), sliding, and wobulated pulse sequences. This study proposes a new approach that models actual sequences as Markov chains using the pulse index as the time parameter and the pulse index of the ideal sequence as the state. The transformation turns radar pulse deinterleaving into a Markov chain-based prediction problem. Since the constructed Markov chain encompasses all the PRI modulation information and is applicable to different objects, the method has great potential in practicality and applicability, providing a fresh perspective and direction to deinterleaving.
- 8 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 60 (6) , 9307-9323
Deinterleaving Method for Radar Signals With Dwell and Switch Pulse Repetition Intervals in a Prior Unknown Scenario
Dwell and switch (D&S) pulse repetition intervals (PRIs) are vital in modern radar systems. However, the existing deinterleaving methods that do not rely on prior information are generally ineffective at handling D&S signals. Considering the characteristics of D&S PRIs, spectrum analysis and maximal clique are employed in this study to deinterleave D&S signals in a prior unknown scenario. Sequence search technology is introduced for D&S subsequences to ensure complete D&S subsequences within the search results. Furthermore, D&S subsequence identification technology based on spectrum analysis is developed. Following vector processing of the search results, D&S subsequences are distinguished from fixed sequences based on the number of spectral peaks within the principal interval. Additionally, a subsequence consolidation technique based on the maximal clique is introduced, where the sequence search result is regarded as the vertex, and the existence of an edge between corresponding vertices is determined by evaluating the complementarity and mutual exclusion of the combination. Through the construction of an undirected graph, subsequences from the same D&S sequence are consolidated using the maximal clique extraction. The simulation results demonstrate that the detection threshold can effectively identify subsequences, whereas the complementarity and mutual exclusion constraints efficiently detect associations between subsequences, mitigating false association pairs from false alarms of fixed sequences. For a pulse missing rate of 30% or less, the analysis accuracy of D&S parameters reaches 98.23%, highlighting the outstanding performance and promising application prospects of the proposed method in deinterleaving D&S sequences and decoding their modulation patterns.
- 8 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 60 (6) , 9307-9323
Deinterleaving Method for Radar Signals With Dwell and Switch Pulse Repetition Intervals in a Prior Unknown Scenario
Dwell and switch (D&S) pulse repetition intervals (PRIs) are vital in modern radar systems. However, the existing deinterleaving methods that do not rely on prior information are generally ineffective at handling D&S signals. Considering the characteristics of D&S PRIs, spectrum analysis and maximal clique are employed in this study to deinterleave D&S signals in a prior unknown scenario. Sequence search technology is introduced for D&S subsequences to ensure complete D&S subsequences within the search results. Furthermore, D&S subsequence identification technology based on spectrum analysis is developed. Following vector processing of the search results, D&S subsequences are distinguished from fixed sequences based on the number of spectral peaks within the principal interval. Additionally, a subsequence consolidation technique based on the maximal clique is introduced, where the sequence search result is regarded as the vertex, and the existence of an edge between corresponding vertices is determined by evaluating the complementarity and mutual exclusion of the combination. Through the construction of an undirected graph, subsequences from the same D&S sequence are consolidated using the maximal clique extraction. The simulation results demonstrate that the detection threshold can effectively identify subsequences, whereas the complementarity and mutual exclusion constraints efficiently detect associations between subsequences, mitigating false association pairs from false alarms of fixed sequences. For a pulse missing rate of 30% or less, the analysis accuracy of D&S parameters reaches 98.23%, highlighting the outstanding performance and promising application prospects of the proposed method in deinterleaving D&S sequences and decoding their modulation patterns.
- Open Access
- 28 February 2024
- Published by Institution of Engineering and Technology (IET) in Electronics Letters
- Vol. 60 (5)
A method for localization TDOA estimation based on signal‐level fusion and analysis
Passive localization using time difference of arrival (TDOA) is one of the common methods for radiated source target localization, and its localization performance depends on the accuracy of TDOA estimation. In many localization scenarios, the signals of target are intricate and often contain multiple ones, so how to effectively extract the time difference parameters from them becomes a difficult problem. This letter proposes a signal‐level fusion method for TDOA estimation to address this problem. At the receiver terminal, multiple signals are fused to obtain their cross‐ambiguity function, so as to estimate the TDOA value for target localization. The effect of fused signals on the accuracy of TDOA estimation is analyzed and the concept of frequency depression is proposed. Simulation experiments verify that the estimation method with signal‐level fusion gets a higher accuracy of the TDOA value and a smaller position error in target localization.
- 28 February 2024
- Published by Institution of Engineering and Technology (IET) in Electronics Letters
- Vol. 60 (5)
Open AccessA method for localization TDOA estimation based on signal‐level fusion and analysis
Passive localization using time difference of arrival (TDOA) is one of the common methods for radiated source target localization, and its localization performance depends on the accuracy of TDOA estimation. In many localization scenarios, the signals of target are intricate and often contain multiple ones, so how to effectively extract the time difference parameters from them becomes a difficult problem. This letter proposes a signal‐level fusion method for TDOA estimation to address this problem. At the receiver terminal, multiple signals are fused to obtain their cross‐ambiguity function, so as to estimate the TDOA value for target localization. The effect of fused signals on the accuracy of TDOA estimation is analyzed and the concept of frequency depression is proposed. Simulation experiments verify that the estimation method with signal‐level fusion gets a higher accuracy of the TDOA value and a smaller position error in target localization.
- Open Access
- 3 February 2024
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 18 (7) , 1080-1093
Open space radar specific emitter identification using MSAK‐CNN‐LSTM network
To enhance the capability of identifying unknown emitters in open spaces, an open‐multiscale attention kernel (MSAK)‐convolutional neural network‐long short‐term memory (CNN‐LSTM) structure is proposed. To this end, first, a MSAK module and CNN‐LSTM structure are introduced, and then, the depth and complexity of the feature extraction network are improved to enhance its representation capability. To classify unknown emitters accurately, the MSAK‐CNN‐LSTM model is improved to obtain an open‐MSAK‐CNN‐LSTM model with open‐set recognition capability. Additionally, the two preprocessing procedures are summarised, and their strengths and weaknesses are compared. Experimental results show that the proposed open‐MSAK‐CNN‐LSTM model achieves satisfactory accuracy in identifying unknown emitters in open space. In addition, it has significant advantages in low signal‐to‐noise ratio (SNR) scenarios.
- 3 February 2024
- Published by Institution of Engineering and Technology (IET) in IET Radar, Sonar & Navigation
- Vol. 18 (7) , 1080-1093
Open AccessOpen space radar specific emitter identification using MSAK‐CNN‐LSTM network
To enhance the capability of identifying unknown emitters in open spaces, an open‐multiscale attention kernel (MSAK)‐convolutional neural network‐long short‐term memory (CNN‐LSTM) structure is proposed. To this end, first, a MSAK module and CNN‐LSTM structure are introduced, and then, the depth and complexity of the feature extraction network are improved to enhance its representation capability. To classify unknown emitters accurately, the MSAK‐CNN‐LSTM model is improved to obtain an open‐MSAK‐CNN‐LSTM model with open‐set recognition capability. Additionally, the two preprocessing procedures are summarised, and their strengths and weaknesses are compared. Experimental results show that the proposed open‐MSAK‐CNN‐LSTM model achieves satisfactory accuracy in identifying unknown emitters in open space. In addition, it has significant advantages in low signal‐to‐noise ratio (SNR) scenarios.