Machine learning-based radar waveform classification for cognitive EW

Orduyilmaz, Adnan; Yar, Ersin; Kocamis, Mehmet Burak; Serin, Mahmut; Efe, Murat

doi:10.1007/s11760-021-01901-w

Cited by 14 publications

(9 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…All the samples were simulated as in the environment of −6 dB SNR. The compared methods in Table 2 could be grouped into three branches, including traditional neural network models [25,30,40], meta-learning models based on optimization [8,9,17], and metric learning [10,12,13].…”

Section: Comparison With Different Methods Of Baselinementioning

confidence: 99%

“…Models based on metric learning generally fitted this task more and performed better than that of optimization. Likewise, ResNet, CNN, and LSTM represent the identification methods adopted in [25,30,40], respectively. Typically, the traditional methods trained their models with the base set, which can hardly recognise the unknown targets even if the support samples were used to fine-tune the model.…”

Section: General Parametersmentioning

confidence: 99%

See 1 more Smart Citation

Adaptive feature extraction and fine‐grained modulation recognition of multi‐function radar under small sample conditions

Zhai

Zhang

et al. 2022

IET Radar Sonar & Navi

View full text Add to dashboard Cite

Multi-function radars (MFRs) are sophisticated sensors with fine-grained modes, which modify their modulation types and parameters range generating various signals to fulfil different tasks, such as surveillance and tracking. In electromagnetic reconnaissance, recognition of MFR fine-grained modes can provide a basis for analysing strategies and reaction. With the limit of real applications, it is hard to obtain a large number of labelled samples for existing methods to learn the difference between categories. Therefore, it is essential to develop new methods to extract general knowledge of MFRs and identify modes with only a few samples. This paper proposes a few-shot learning (FSL) framework based on efficient neural architecture search (ENAS) with high robustness and portability, which designs a suitable network structure automated and quickly adapts to new environments. The experimental results show that the proposed method can still achieve excellent fine-grained modulation recognition performance (92.6%) under the condition of -6 dB signal-to-noise ratio (SNR), even when each class only provides one fixed-duration signal sample. The robustness is also verified under different conditions.

show abstract

Section: Comparison With Different Methods Of Baselinementioning

confidence: 99%

Section: General Parametersmentioning

confidence: 99%

Adaptive feature extraction and fine‐grained modulation recognition of multi‐function radar under small sample conditions

Zhai

Zhang

et al. 2022

IET Radar Sonar & Navi

View full text Add to dashboard Cite

show abstract

“…The convolutional neural networks (CNNs), which work phenomenally well on computer vision tasks like image classification and recognition, are the most widely used type of DNNs that have been employed to recognise the LPI waveforms transmitted from radar and have achieved great success [34][35][36][37][38][39][40]. As a pre-processing procedure, the intercepted radar signals are first transformed into the time-frequency images using various time-frequency techniques, which include the Choi-William distribution (CWD) [34,35], the Fourier-based Synchrosqueezing transform (FSST) [36], the Wigner Ville distribution (WVD) [37], and the short-time Fourier transform [38][39][40]. After that, the classic CNN models, for example, VGG16, ResNet50, DenseNet, GoogLeNet, and AlexNet, or other self-built typical CNN architectures, could be employed for radar signal recognition.…”

Section: Rank-4: Adaptive Radar Countermeasures (Arc)mentioning

confidence: 99%

“…With the recent development of deep neural networks (DNNs), a classification accuracy of higher than 90% at SNR = −4 dB has been reported in many research works [34,35]. The convolutional neural networks (CNNs), which work phenomenally well on computer vision tasks like image classification and recognition, are the most widely used type of DNNs that have been employed to recognise the LPI waveforms transmitted from radar and have achieved great success [34][35][36][37][38][39][40]. As a pre-processing procedure, the intercepted radar signals are first transformed into the time-frequency images using various time-frequency techniques, which include the Choi-William distribution (CWD) [34,35], the Fourier-based Synchrosqueezing transform (FSST) [36], the Wigner Ville distribution (WVD) [37], and the short-time Fourier transform [38][39][40].…”

Section: Rank-4: Adaptive Radar Countermeasures (Arc)mentioning

confidence: 99%

“…where λ is the Gaussian pulse length parameter and b ¼ Δ f =ð2T s Þ is the frequency modulation rate, with Δ f and T s denoting the frequency sweep and the effective pulse duration, respectively. For adaptive waveform selection, λ and Δ f are selected from λ∈ [20,30,40,50, 60] μs and Δ f ∈[0.1, 0.325, 0.55, 0.775, 1] MHz, respectively. The tracking results when using the fixed waveforms corresponding to the maximum time-bandwidth product and adaptively selecting the waveforms according to the policy of and the MSE are shown in Figure 4.…”

Section: Rank-3: Active Electronically Scanned Array (Aesa)mentioning

confidence: 99%

See 1 more Smart Citation

Moving target detection and tracking with multiplatform radar network (MRN)

Geng

Wang

et al. 2021

IET Radar Sonar & Navi

View full text Add to dashboard Cite

A multiplatform radar network (MRN) composed of multiple transmitting and receiving facilities is advantageous over monostatic radar in moving target detection (MTD) and tracking due to the diversity gain. Within the framework of the Internet of Battlefield Things (IoBT), the authors propose that the airborne platforms equipped with the cognitive multimodal radar (CMR), the subarray‐based MIMO radar (SMR), the active electronically scanned array (AESA) radar, and the adaptive radar countermeasures (ARC) should team up in target detection and tracking. To obtain better target detection and tracking performance, the probing signals from the CMR, the SMR, and the AESA radar are adaptively optimised based on the scenario under consideration. The ARC consists of two parts: 1) RF signal sensing and recognition with deep neural network (DNN), that is, to detect, identify, and localise multiple radio frequency emitters; 2) intelligent fake return creation and jamming to confuse the enemy radars in real time in the field. This work aims to serve as a good reference for future researchers interested in developing MRN with diverse nodes and the general teaming of radars on the battlefield.

show abstract

Object detection based deinterleaving of radar signals using deep learning for cognitive EW

Kocamış,

Orduyılmaz,

Taşcıoğlu

2024

SIViP

View full text Add to dashboard Cite

In a real-world environment, multifunction radars (MFRs) pose major challenges to the electronic support system that is a part of cognitive electronic warfare. One of the main problems is deinterleaving of signals belonging to MFRs. The ability of MFRs to change their carrier frequency, pulse width, and pulse repetition interval (PRI) from pulse to pulse makes the deinterleaving task challenging. In this paper, an object detection based deinterleaving approach exploiting amplitude patterns caused by radar beam motions, which are determined based on radar antenna scan types, is proposed to deinterleave MFR signals. Amplitude patterns are created in two-dimensional images using amplitude (AMP) and time of arrival (TOA) parameters obtained from radar pulses. Deinterleaving is performed using a deep learning algorithm applied to these images. To the best of the authors’ knowledge, object detection based deinterleaving of radar signals using AMP-TOA images has not been considered so far. Contrary to the common approaches, in which searching for PRI patterns is required, amplitude patterns formed by the radar beam motions on the electronic support system are used in the proposed method. This enables robust identification of signals of MFRs having PRI, carrier frequency, and pulse width agility. With the proposed method, the intention of the radar, i.e., search or tracking, can be acquired in the deinterleaving stage, which ensures earlier situational awareness. The performance of the method is evaluated for varying number of radar signals from one to five through simulations, in which scenarios with missing pulses at different rates are considered. Simulation results demonstrate that, on the average, more than 0.98 mean average precision (mAP50) is achieved at 30% missing pulse rate. The performance of the method for search radar signals is slightly lower compared to that achieved for tracking radar signals due to the lower number of pulses received by the system. However, more than 0.93 AP50 is achieved for search radar signals in the presence of five radar signals with different missing pulse rates. Besides, real-time performance can be achieved using the proposed method on the GPU platform.

show abstract

Machine learning-based radar waveform classification for cognitive EW

Cited by 14 publications

References 13 publications

Adaptive feature extraction and fine‐grained modulation recognition of multi‐function radar under small sample conditions

Adaptive feature extraction and fine‐grained modulation recognition of multi‐function radar under small sample conditions

Moving target detection and tracking with multiplatform radar network (MRN)

Object detection based deinterleaving of radar signals using deep learning for cognitive EW

Contact Info

Product

Resources

About