Noise‐robust HRRP target recognition method via sparse‐low‐rank representation

The multilayered feed-forward neural network was applied to automatic target recognition using the high range resolution (HRR) profiles in this paper. To extract effective features from the HRR profiles, the product spectrum originally proposed for the speech signal processing was introduced to the radar target recognition community. The product spectrum was defined as the product of the power spectrum and the group delay function, which could combine the information contained in the magnitude spectrum and phase spectrum of the HRR profiles and carry more details about the shape of the aircrafts. A multilayered feed-forward neural network was selected as classifier. The HRR profiles were obtained using the two-dimensional backscatters distribution data of four different scaled aircraft models. Simulations were presented to evaluate the classification performance with the product spectrum based features. The results demonstrate that the product spectrum based features outperform the original HRR profiles and the multilayered feed-forward neural network is effective for the application of automatic target recognition of aircrafts.

ABSTRACTThe range alignment technique is extended and applied to full polarization processing technique using the fully polarized data for a target. By the maximum likelihood-based span, the fused high-resolution range profile (HRRP) is achieved. Then the range alignment is carried out on the fused HRRPs and the offsets are used to align the original HRRPs of each polarimetric channel. Applying this technique, the signal-to-noise ratio (SNR) of the HRRP is increased and the performance of the range alignment is improved in comparison with that of the conventional techniques using singly polarized data, and therefore the inverse synthetic aperture radar images of the target can be clearly obtained even in low SNR condition. Experiments from measured data under field conditions are presented to validate the proposed technique.

In target recognition based on high range resolution profile (HRRP), the major challenge is to extract invariant and robust features from the measured high dimension range profile samples. Focused on the reduction of dimensionality, a novel algorithm for extracting HRRP features is proposed in this paper, and the Legendre moments with orthogonal and translation-invariant properties was adopted. The proposed algorithm utilizes the average HRRP in aspect-frame, and then preprocesses the HRRP with amplitude l1 normalization. The Legendre moments are calculated with mapping range values into orthogonal definition domain to provide invariant feature vectors. Meanwhile, spectrum analysis (SA), which can be regard as first-order approximation of relax algorithm, is adopted to reconstruct the real HRRP obtained by inverse fast fourier transform (IFFT) in order to desensitize moments in low SNR environment. Several experiments with four different targets measured data demonstrate that Legendre moments features with flexible maximum order to be chosen have significant advantages over central moments. SA can provide a more robust recognition performance in low signal noise ratio (SNR) environment as 5 dB compared with those of HRRP directly obtained by IFFT.

Radar high-resolution range profile (HRRP) has received intensive attention from the radar automatic recognition (RATR) community. Since the initial phase of a complex HRRP is strongly sensitive to target position variation, which is referred to as the initial phase sensitivity, only the amplitude information in the complex HRRP, what is called the real HRRP, is used for RATR. This paper proposes a novel feature extraction method for the complex HRRP. The extracted complex feature vector contains the difference phase information between range cells but no initial phase information in the complex HRRP. The recognition algorithms, frame-template-database establishment methods and preprocessing methods used in the real HRRP-based RATR can also be applied to the proposed complex feature vector-based RATR. The recognition experiments based on measured data show that the proposed complex feature vector can obtain better recognition performance than the real HRRP if only the cell interval parameters are proper.

This study focuses on the bistatic high range resolution profiles (HRRPs) of precessing targets and considers how to utilise bistatic HRRPs for feature extraction. The movement model of the slippery bistatic scattering centres of precessing targets and the wideband bistatic echo model are introduced firstly. The dynamic simulation method of the bistatic HRRPs of the precessing target is presented using the bistatic scattering coefficients obtained via electromagnetic calculations, and the simulation results validate the theoretical analysis. Furthermore, the characteristics of the bistatic HRRP are analysed and show that the bistatic HRRP generally reflects the projection of the target onto the bisector of the bistatic angle, except that the bistatic angle is so wide that the bistatic HRRP is quite different from the monostatic HRRP. This research is helpful for imaging, and identifying precessing targets in the bistatic radar.

In radar High Resolution Range Profile (HRRP) recognition field, an effective feature extraction process is the main point to improve whole recognition performance. As the targets' HRRP is the amplitude of coherent summations of complex return from scatters in each range cell, some data-driven compression methods, like PCA, LDA, etc. are widely used and achieve good results under experimental conditions, but still doesn't contribute much to solving the inherent problem of HRRP, such as small-sample problem. For these reasons, a series of SOM based feature extraction methods are introduced to explore the possibility in HRRP recognition. From machine learning perspective, these SOM based methods might more comprehensively describe the inherent structure relations within the original HRRP data distribution with its competitive learning process. In this paper, three category simulation experiments are presented to analysis the feature extraction, data storage, cluster result and classifier design. Of these experiments, SOM combined with multi-layer neural network can reduce the over fitting problem of directly apply the multi-layer neural network for HRRP recognition, which further prove the feasibility of SOM application in HRRP recognition.

This study considers how to utilize the high range resolution profiles (HRRPs) of wobbling targets for feature extraction. The movement model of wobbling targets and the bistatic HRRP model are introduced firstly. Based on the bistatic HRRPs of the wobbling target simulated via the bistatic scattering coefficients obtained via electromagnetic calculations, the polarization characteristics of the bistatic HRRP are analysed. This study is useful for bistatic radar identification of wobbling targets.

Acoustic bird species classification under low SNR and small-scale dataset conditions

A high range resolution profile (HRRP) is a good characteristic for radar target, and radar target recognition by using HRRP has been studied extensively. Target recognition by fusion the target features of multiple polarimetric directions is studied in this paper, the framework of multi-polarized HRRP classification by SVM ensemble is given, and a novel approach of multi-polarized range profile classification by SVM dynamic ensemble based on fuzzy integral is proposed. Experiments on the measured multi-polarized HRRP indicate the effectiveness of our conclusion.

Radar automatic target recognition (RATR) based on high resolution range profiles (HRRPs) has received intensive attention in recent years. The data-driven HRRP recognition methods based on neural networks have achieved outstanding performance with a complete target-aspect template library. However, such methods can barely achieve satisfactory performance under the condition of incomplete target-aspects or low signal-to-noise ratio. In order to solve this problem, a prior-knowledge guided neural network (PriorK-NN) is proposed for HRRP target recognition, which incorporates prior-knowledge about scattering centers and target-aspects into neural networks. Firstly, by combining the physical generative mechanism of HRRP with neural networks, we proposed an interpretable scattering center layer (SC-layer). The SC-layer can effectively reduce the adverse effects of noise and extract the location and intensity of target dominant scattering centers. In addition, we proposed a loss function named targetaspect supervised contrastive loss based on the prior-knowledge about target-aspects. By using the proposed loss function, clusters of HRRPs belonging to the same class but with different target-aspects are pulled together, while simultaneously pushing apart clusters of HRRPs from different classes. Therefore, the proposed loss function can improve the recognition performance under the condition of incomplete target-aspects. Experiments on the aircraft electromagnetic simulation dataset and the measured dataset validated the effectiveness of our proposed method on noise-corrupted HRRP recognition and demonstrated superior performance compared with other HRRP recognition methods under the condition of incomplete target-aspects.

Convolution neural networks (CNNs) represent one of the workhorses of artificial intelligence applications. As a typical artificial intelligence application, a high‐resolution range profile (HRRP) target recognition method based on CNNs has aroused a lot of research interest. Most CNNs use a relatively small and single‐scale convolution kernel size to control the number of parameters and computational complexity, but recent studies indicate that CNNs with a small kernel size cannot extract enough spatial information, which hurts the recognition performance. Aiming at this problem, this paper proposes a multi‐scale group‐fusion one‐dimensional convolution neural network (MSGF‐1D‐CNN) for HRRP target recognition. MSGF‐1D‐CNN utilises multi‐scale group one‐dimensional convolution (MSG 1D‐Conv) and point‐wise convolution (PW‐Conv) to replace the standard convolution. Multi‐scale group one‐dimensional convolution can significantly reduce complexity and capture the information of targets within HRRP in different levels of detail to enhance feature extraction, while PW‐Conv can realise the fusion of multi‐scale features to help boosting recognition performance. Experiments on five mid‐course ballistic targets in the HRRP dataset show that MSGF‐1D‐CNN has superior recognition performance, and the parameter number of the model is reduced by more than 2.4 times than standard 1D‐CNN. Furthermore, MSGF‐1D‐CNN shows better performance on fine‐grained HRRP target recognition and anti‐noise robustness in most cases.

Aiming at the Carrier-Free Ultra-Wide-Band Radar (CF-UWBR) under the near-range conditions, a novel target recognition method based on the Gram Angle Field (GAF) feature extraction and the Inception V3 network is proposed in this paper. The one-dimensional High-Resolution Range Profile (HRRP) of CF-UWBR is highly sensitive to observation angle so that the performance of direct recognition using the original time-domain HRRP data is not satisfying. In response to this problem, the GAF method is employed to convert the HRRP data into a two-dimensional image firstly, to maintain and highlight the correlation features between the adjacent time points in the polar coordinate domain. Furthermore, the Inception V3 network is selected for the target recognition based on the dual-mode GAF features, due to the excellent performance with the fewer computing resources. The experimental results on the actual measured CF-UWBR data set indicated that the proposed method is superior in performance to the methods based on the Support Vector Machine (SVM), the Gated Recurrent Unit (GRU) network, or the Convolutional Neural Network (CNN) with the original HRRP image. Especially when the dual-mode GAFs are used, the recognition accuracy of 97.63% can be achieved, about 1.95% higher than the STOA.

<p>The aspect sensitivity of high-resolution range profile (HRRP) leads to the anomalous change of the HRRP statistical characteristic, which is one of inextricable problems on the target recognition based on HRRP. Aiming at the HRRP statistical characteristic, an adaptive angular-sector segmentation method is proposed through based on the grey relational mode. Comparing to the equal interval angular-sector segmentation method, the new method improves the recognition performance. And these simulation results of five kinds of aircraft targets HRRPs prove the feasibility and validity.</p>

One-dimension hierarchical local receptive fields based extreme learning machine for radar target HRRP recognition

In this paper, a novel approach, namely kernel projection vector space (KPVS), is proposed for radar target recognition using high-resolution range profile (HRRP). First, the HRRP samples are mapped into a high-dimensional feature space using nonlinear mapping. Second, the kernel projection vectors, are obtained by kernel discriminant analysis. Then, for each class, the kernel projection vector space is formed using all the training kernel projection vectors of class. Finally, the minimum hyper plane distance classifier (MHDC) is used for classification. The aim of KPVS method is to represent the feature area of target using kernel projection vector space, and effectively measure the distance between the test HRRP and feature area via minimum hyper plane distance (MHD) metric. The experimental results of measured data show that the proposed method has better performance of recognition than KPCA and KFDA.