Current Issue

28 August 2023, Volume 34 Issue 4
2023, 34(4):  0. 
Abstract ( )   PDF (112KB) ( )  
Related Articles | Metrics
Recognition of dynamically varying PRI modulation via deep learning and recurrence plot
Pengcheng WANG, Weisong LIU, Zheng LIU
2023, 34(4):  815-826.  doi:10.23919/JSEE.2022.000071
Abstract ( )   HTML ( )   PDF (2584KB) ( )  
Figures and Tables | References | Related Articles | Metrics

Recognition of pulse repetition interval (PRI) modulation is a fundamental task in the interpretation of radar intentions. However, the existing PRI modulation recognition methods mainly focus on single-label classification of PRI sequences. The prerequisite for the effectiveness of these methods is that the PRI sequences are perfectly divided according to different modulation types before identification, while the actual situation is that radar pulses reach the receiver continuously, and there is no completely reliable method to achieve this division in the case of non-cooperative reception. Based on the above actual needs, this paper implements an algorithm based on the recurrence plot technique and the multi-target detection model, which does not need to divide the PRI sequence in advance. Compared with the sliding window method, it can more effectively realize the recognition of the dynamically varying PRI modulation.

Robust least squares projection twin SVM and its sparse solution
Shuisheng ZHOU, Wenmeng ZHANG, Li CHEN, Mingliang XU
2023, 34(4):  827-838.  doi:10.23919/JSEE.2023.000103
Abstract ( )   HTML ( )   PDF (6601KB) ( )  
Figures and Tables | References | Related Articles | Metrics

Least squares projection twin support vector machine (LSPTSVM) has faster computing speed than classical least squares support vector machine (LSSVM). However, LSPTSVM is sensitive to outliers and its solution lacks sparsity. Therefore, it is difficult for LSPTSVM to process large-scale datasets with outliers. In this paper, we propose a robust LSPTSVM model (called R-LSPTSVM) by applying truncated least squares loss function. The robustness of R-LSPTSVM is proved from a weighted perspective. Furthermore, we obtain the sparse solution of R-LSPTSVM by using the pivoting Cholesky factorization method in primal space. Finally, the sparse R-LSPTSVM algorithm (SR-LSPTSVM) is proposed. Experimental results show that SR-LSPTSVM is insensitive to outliers and can deal with large-scale datasets fastly.

Range estimation of few-shot underwater sound source in shallow water based on transfer learning and residual CNN
Qihai YAO, Yong WANG, Yixin YANG
2023, 34(4):  839-850.  doi:10.23919/JSEE.2023.000095
Abstract ( )   HTML ( )   PDF (3506KB) ( )  
Figures and Tables | References | Related Articles | Metrics

Taking the real part and the imaginary part of complex sound pressure of the sound field as features, a transfer learning model is constructed. Based on the pre-training of a large amount of underwater acoustic data in the preselected sea area using the convolutional neural network (CNN), the few-shot underwater acoustic data in the test sea area are retrained to study the underwater sound source ranging problem. The S5 voyage data of SWellEX-96 experiment is used to verify the proposed method, realize the range estimation for the shallow source in the experiment, and compare the range estimation performance of the underwater target sound source of four methods: matched field processing (MFP), generalized regression neural network (GRNN), traditional CNN, and transfer learning. Experimental data processing results show that the transfer learning model based on residual CNN can effectively realize range estimation in few-shot scenes, and the estimation performance is remarkably better than that of other methods.

Low-complexity iterative equalization for OTFS based on alternating minimization
Xin HE, Haoxiang JIA, Yutong SUN, Zijian ZHOU, Danfeng ZHAO
2023, 34(4):  851-860.  doi:10.23919/JSEE.2023.000089
Abstract ( )   HTML ( )   PDF (8015KB) ( )  
Figures and Tables | References | Related Articles | Metrics

To achieve robust communication in high mobility scenarios, an iterative equalization algorithm based on alternating minimization (AM) is proposed for the orthogonal time frequency space (OTFS) system. The algorithm approximates the equalization problem to a convex function optimization problem in the real-valued domain and solves the problem iteratively using the AM algorithm. In the iterative process, the complexity of the proposed algorithm is reduced further based on the study of the cyclic structure and sparse property of the OTFS channel matrix in the delay-Doppler (DD) domain. The new method for OTFS is simulated and verified in a high-speed mobile scenario and the results show that the proposed equalization algorithm has excellent bit error rate performance with low complexity.

Research on infrared dim and small target detection algorithm based on low-rank tensor recovery
Chuntong LIU, Hao WANG
2023, 34(4):  861-872.  doi:10.23919/JSEE.2023.000004
Abstract ( )   HTML ( )   PDF (3845KB) ( )  
Figures and Tables | References | Related Articles | Metrics

In order to rapidly and accurately detect infrared small and dim targets in the infrared image of complex scene collected by virtual prototyping of space-based downward-looking multiband detection, an improved detection algorithm of infrared small and dim target is proposed in this paper. Firstly, the original infrared images are changed into a new infrared patch tensor mode through data reconstruction. Then, the infrared small and dim target detection problems are converted to low-rank tensor recovery problems based on tensor nuclear norm in accordance with patch tensor characteristics, and inverse variance weighted entropy is defined for self-adaptive adjustment of sparseness. Finally, the low-rank tensor recovery problem with noise is solved by alternating the direction method to obtain the sparse target image, and the final small target is worked out by a simple partitioning algorithm. The test results in various space-based downward-looking complex scenes show that such method can restrain complex background well by virtue of rapid arithmetic speed with high detection probability and low false alarm rate. It is a kind of infrared small and dim target detection method with good performance.

A tunable adaptive detector for distributed targets when signal mismatch occurs
Yufeng CUI, Yongliang WANG, Weijian LIU, Qinglei DU, Xichuan ZHANG, Xuhui LI
2023, 34(4):  873-878.  doi:10.23919/JSEE.2023.000029
Abstract ( )   HTML ( )   PDF (6473KB) ( )  
Figures and Tables | References | Related Articles | Metrics

Aiming at the problem of detecting a distributed target when signal mismatch occurs, this paper proposes a tunable detector parameterized by an adjustable parameter. By adjusting the parameter, the tunable detector can achieve robust or selective detection of mismatched signals. Moreover, the proposed tunable detector, with a proper tunable parameter, can provide higher detection probability compared with existing detectors in the case of no signal mismatch. In addition, the proposed tunable detector possesses the constant false alarm rate property with the unknown noise covariance matrix.

A modified OMP method for multi-orbit three dimensional ISAR imaging of the space target
Libing JIANG, Shuyu ZHENG, Qingwei YANG, Peng YANG, Zhuang WANG
2023, 34(4):  879-893.  doi:10.23919/JSEE.2023.000066
Abstract ( )   HTML ( )   PDF (6990KB) ( )  
Figures and Tables | References | Related Articles | Metrics

The conventional two dimensional (2D) inverse synthetic aperture radar (ISAR) imaging fails to provide the targets’ three dimensional (3D) information. In this paper, a 3D ISAR imaging method for the space target is proposed based on mutli-orbit observation data and an improved orthogonal matching pursuit (OMP) algorithm. Firstly, the 3D scattered field data is converted into a set of 2D matrix by stacking slices of the 3D data along the elevation direction dimension. Then, an improved OMP algorithm is applied to recover the space target’s amplitude information via the 2D matrix data. Finally, scattering centers can be reconstructed with specific three dimensional locations. Numerical simulations are provided to demonstrate the effectiveness and superiority of the proposed 3D imaging method.

A high frequency vibration compensation approach for ultrahigh resolution SAR imaging based on sinusoidal frequency modulation Fourier-Bessel transform
Siyu CHEN, Yong WANG, Rui CAO
2023, 34(4):  894-905.  doi:10.23919/JSEE.2023.000059
Abstract ( )   HTML ( )   PDF (5204KB) ( )  
Figures and Tables | References | Related Articles | Metrics

Ultrahigh resolution synthetic aperture radar (SAR) imaging for ship targets is significant in SAR imaging, but it suffers from high frequency vibration of the platform, which will induce defocus into SAR imaging results. In this paper, a novel compensation method based on the sinusoidal frequency modulation Fourier-Bessel transform (SFMFBT) is proposed, it can estimate the vibration errors, and the phase shift ambiguity can be avoided via extracting the time frequency ridge consequently. By constructing the corresponding compensation function and combined with the inverse SAR (ISAR) technique, well-focused imaging results can be obtained. The simulation imaging results of ship targets demonstrate the validity of the proposed approach.

Mission scheduling of multi-sensor collaborative observation for space surveillance network
Xi LONG, Weiwei CAI, Leping YANG, Tianyu WANG
2023, 34(4):  906-923.  doi:10.23919/JSEE.2023.000104
Abstract ( )   HTML ( )   PDF (5741KB) ( )  
Figures and Tables | References | Related Articles | Metrics

With increased dependence on space assets, scheduling and tasking of the space surveillance network (SSN) are vitally important. The multi-sensor collaborative observation scheduling (MCOS) problem is a multi-constraint and high-conflict complex combinatorial optimization problem that is non-deterministic polynomial (NP)-hard. This research establishes a sub-time window constraint satisfaction problem (STWCSP) model with the objective of maximizing observation profit. Considering the significant effect of genetic algorithms (GA) on solving the problem of resource allocation, an evolution heuristic (EH) algorithm containing three strategies that focus on the MCOS problem is proposed. For each case, a task scheduling sequence is first obtained via an improved GA with penalty (GAPE) algorithm, and then a mission planning algorithm (heuristic rule) is used to determine the specific observation time. Compared to the model without sub-time windows and some other algorithms, a series of experiments illustrate the STWCSP model has better performance in terms of total profit. Experiments about strategy and parameter sensitivity validate its excellent performance in terms of EH algorithms.

Modular flexible “Tetris” microsatellite platform based on sandwich assembly mode
Jun ZHOU, Hao ZHANG, Guanghui LIU, Cheng CHENG, Jiaolong ZHANG
2023, 34(4):  924-938.  doi:10.23919/JSEE.2023.000094
Abstract ( )   HTML ( )   PDF (10682KB) ( )  
Figures and Tables | References | Related Articles | Metrics

In this paper, a flexible modular “Tetris” microsatellite platform is studied to implement the rapid integration and assembly of microsatellites. The proposed microsatellite platform is fulfilled based on a sandwich assembly mode which consists of the isomorphic module structure and the standard mechanical-electric-data-thermal interfaces. The advantages of the sandwich assembly mode include flexible reconfiguration and efficient assembly. The prototype of the sandwich assembly mode is built for verifying the performance and the feasibility of the proposed mechanical-electric-data-thermal interfaces. Finally, an assembly case is accomplished to demonstrate the validity and advantages of the proposed “Tetris” microsatellite platform.

Selecting suitable key supplier for core components during smart complex equipment central-private enterprises collaborative development process: from two different forms of evaluation information and matching perspective
Xin HUANG, Xiaoyan QI, Hongzhuan CHEN, Xiang CAI
2023, 34(4):  939-954.  doi:10.23919/JSEE.2023.000099
Abstract ( )   HTML ( )   PDF (5074KB) ( )  
Figures and Tables | References | Related Articles | Metrics

With the development of central-private enterprises integration, selecting suitable key suppliers are able to provide core components for smart complex equipment. We consider selecting suitable key suppliers from matching perspective, for it not only satisfies natural development of smart complex equipment, it is also a good implementation of equipment project in central-private enterprises integration context. In in this paper, we carry out two parts of research, one is evaluation attributes based on comprehensive analysis, and the other is matching process between key suppliers and core components based on the matching attribute. In practical analysis process, we employ comprehensive evaluated analysis methods to acquire relevant attributes for the matching process that follows. In the analysis process, we adopt entropy-maximum deviation method (MDM)-decision-making trial and evaluation laboratory (DEMATEL)-technique for order preference by similarity to an ideal solution (TOPSIS) to obtain a comprehensive analysis. The entropy-MDM is applied to get weight value, DEMATEL is utilized to obtain internal relations, and TOPSIS is adopted to get ideal evaluated solution. We consider aggregating two types of evaluation information according to similarities of smart complex equipment based on the combination between geometric mean and arithmetic mean. Moreover, based on the aforementioned attributes and generalized power Heronian mean operator, we aggregate preference information to acquire relevant satisfaction degree, then combine the constructed matching model to get suitable key supplier. Through comprehensive analysis of selecting suitable suppliers, we know that two-sided matching and information aggregation can provide more research perspectives for smart complex equipment. Through analysis for relevant factors, we find that leading role and service level are also significant for the smart complex equipment development process.

Hound: a parallel image distribution system for cluster based on Docker
Zijie LIU, Junjiang LI, Can CHEN, Dengyin ZHANG
2023, 34(4):  955-965.  doi:10.23919/JSEE.2023.000105
Abstract ( )   HTML ( )   PDF (736KB) ( )  
Figures and Tables | References | Related Articles | Metrics

Current applications, consisting of multiple replicas, are packaged into lightweight containers with their execution dependencies. Considering the dominant impact of distribution efficiency of gigantic images on container startup (e.g., distributed deep learning application), the image “warm-up” technique which prefetches images of these replicas to destination nodes in the cluster is proposed. However, the current image “warm-up” technique solely focuses on identical image distribution, which fails to take effect when distributing different images to destination nodes. To address this problem, this paper proposes Hound, a simple but efficient cluster image distribution system based on Docker. To support diverse image distribution requests of cluster nodes, Hound additionally adopts node-level parallelism (i.e., downloading images to destination nodes in parallel) to further improve the efficiency of image distribution. The experimental results demonstrate Hound outperforms Docker, kubernetes container runtime interface (CRI-O), and Docker-compose in terms of image distribution performance when cluster nodes request different images. Moreover, the high scalability of Hound is evaluated in the scenario of ten nodes.

A multi-enterprise quality function deployment paradigm with unstructured decision-making in linguistic contexts
Jian’gang PENG, Guang XIA, Baoqun SUN, Shaojie WANG
2023, 34(4):  966-980.  doi:10.23919/JSEE.2022.000130
Abstract ( )   HTML ( )   PDF (4895KB) ( )  
Figures and Tables | References | Related Articles | Metrics

This paper presents an operational framework of unstructured decision-making approach involving quality function deployment (QFD) in an uncertain linguistic context. Firstly, QFD is extended to the multi-enterprise paradigm in a real-world manufacturing environment. Secondly, hesitant fuzzy linguistic term sets (HFLTSs), which facilitate the management and handling of information equivocality, are designed to construct a house of quality (HoQ) in the product planning process. The technique of computing with words is applied to bridge the gap between mechanisms of the human brain and machine processes with fuzzy linguistic term sets. Thirdly, a multi-enterprise QFD pattern is formulated as an unstructured decision-making problem for alternative infrastructure project selection in a manufacturing organization. The inter-relationships of cooperative partners are directly matched with a back propagation neural network (BPNN) to construct the multi-enterprise manufacturing network. The resilience of the manufacturing organization is considered by formulating an outranking method on the basis of HFLTSs to decide on infrastructure project alternatives. Finally, a real-world example, namely, the prototype manufacturing of an automatic transmission for a vehicle, is provided to illustrate the effectiveness of the proposed decision-making approach.

An accurate detection algorithm for time backtracked projectile-induced water columns based on the improved YOLO network
Yasong LUO, Jianghu XU, Chengxu FENG, Kun ZHANG
2023, 34(4):  981-991.  doi:10.23919/JSEE.2023.000106
Abstract ( )   HTML ( )   PDF (1357KB) ( )  
Figures and Tables | References | Related Articles | Metrics

During a sea firing training, the intelligent detection of projectile-induced water column targets in a firing video is the prerequisite for and critical to the automatic calculation of miss distance, while the correct and precise calculation of miss distance is directly affected by the accuracy, false alarm rate and time delay of detection. After analyzing the characteristics of projectile-induced water columns, an accurate detection algorithm for time backtracked projectile-induced water columns based on the improved you only look once (YOLO) network is put forward. The capability and accuracy of detecting projectile-induced water column targets with the conventional YOLO network are improved by optimizing the anchor box through K-means clustering and embedding the squeeze and excitation (SE) attention module. The detection area is limited by adopting a sea-sky line detection algorithm based on gray level co-occurrence matrix (GLCM), so as to effectively eliminate such disturbances as ocean waves and ship wakes, and lower the false alarm rate of projectile-induced water column detection. The improved algorithm increases the mAP50 of water column detection by 30.3%. On the basis of correct detection, a time backtracking algorithm is designed with mean shift to track images containing projectile-induced water column in reverse time sequence. It accurately detects a projectile-induced water column at the time of its initial appearance as well as its pixel position in images, and considerably reduces detection delay, so as to provide the support for the automatic, accurate, and real-time calculation of miss distance.

Influencing factor analysis of interception probability and classification-regression neural network based estimation
Yi NAN, Guoxing YI, Lei HU, Changhong WANG, Zhenbiao TU
2023, 34(4):  992-1006.  doi:10.23919/JSEE.2023.000092
Abstract ( )   HTML ( )   PDF (5671KB) ( )  
Figures and Tables | References | Related Articles | Metrics

The interception probability of a single missile is the basis for combat plan design and weapon performance evaluation, while its influencing factors are complex and mutually coupled. Existing calculation methods have very limited analysis of the influence mechanism of influencing factors, and none of them has analyzed the influence of the guidance law. This paper considers the influencing factors of both the interceptor and the target more comprehensively. Interceptor parameters include speed, guidance law, guidance error, fuze error, and fragment killing ability, while target performance includes speed, maneuverability, and vulnerability. In this paper, an interception model is established, Monte Carlo simulation is carried out, and the influence mechanism of each factor is analyzed based on the model and simulation results. Finally, this paper proposes a classification-regression neural network to quickly estimate the interception probability based on the value of influencing factors. The proposed method reduces the interference of invalid interception data to valid data, so its prediction accuracy is significantly better than that of pure regression neural networks.

Shuffled frog leaping algorithm with non-dominated sorting for dynamic weapon-target assignment
Yang ZHAO, Jicheng LIU, Ju JIANG, Ziyang ZHEN
2023, 34(4):  1007-1019.  doi:10.23919/JSEE.2023.000102
Abstract ( )   HTML ( )   PDF (5482KB) ( )  
Figures and Tables | References | Related Articles | Metrics

The dynamic weapon target assignment (DWTA) problem is of great significance in modern air combat. However, DWTA is a highly complex constrained multi-objective combinatorial optimization problem. An improved elitist non-dominated sorting genetic algorithm-II (NSGA-II) called the non-dominated shuffled frog leaping algorithm (NSFLA) is proposed to maximize damage to enemy targets and minimize the self-threat in air combat constraints. In NSFLA, the shuffled frog leaping algorithm (SFLA) is introduced to NSGA-II to replace the inside evolutionary scheme of the genetic algorithm (GA), displaying low optimization speed and heterogeneous space search defects. Two improvements have also been raised to promote the internal optimization performance of SFLA. Firstly, the local evolution scheme, a novel crossover mechanism, ensures that each individual participates in updating instead of only the worst ones, which can expand the diversity of the population. Secondly, a discrete adaptive mutation algorithm based on the function change rate is applied to balance the global and local search. Finally, the scheme is verified in various air combat scenarios. The results show that the proposed NSFLA has apparent advantages in solution quality and efficiency, especially in many aircraft and the dynamic air combat environment.

Optimal maneuver strategy to improve the observability of angles-only rendezvous
Ronghua DU, Wenhe LIAO, Xiang ZHANG
2023, 34(4):  1020-1032.  doi:10.23919/JSEE.2023.000091
Abstract ( )   HTML ( )   PDF (4622KB) ( )  
Figures and Tables | References | Related Articles | Metrics

This paper proposes an optimal maneuver strategy to improve the observability of angles-only rendezvous from the perspective of relative navigation. A set of dimensionless relative orbital elements (ROEs) is used to parameterize the relative motion, and the objective function of the observability of angles-only navigation is established. An analytical solution of the optimal maneuver strategy to improve the observability of angles-only navigation is obtained by means of numerical analysis. A set of dedicated semi-physical simulation system is built to test the performances of the proposed optimal maneuver strategy. Finally, the effectiveness of the method proposed in this paper is verified through the comparative analysis of the objective function of the observability of angles-only navigation and the performances of the angles-only navigation filter under different maneuver schemes. Compared with the cases without orbital maneuver, it is concluded that the tangential filtering accuracy with the optimal orbital maneuver at the terminal time is increased by 35% on average, and the radial and normal filtering accuracy is increased by 30% on average.

Robust output regulation problem with prescribed performance for nonlinear strict feedback systems
Haichao ZHU, Weiyao LAN
2023, 34(4):  1033-1041.  doi:10.23919/JSEE.2023.000098
Abstract ( )   HTML ( )   PDF (2648KB) ( )  
Figures and Tables | References | Related Articles | Metrics

This paper investigates the problem of robust output regulation control with prospected transient property for strict feedback systems. By employing the internal model principle, the robust output regulation problem with a prospected property can be transformed to a robust stabilization problem with a new output constraint. Then, by constructing the speed function and adopting barrier Lyapunov function technique, the dynamic feedback controller can be designed not only to drive error output of the closed-loop system entering into a prescribed performance bound within a given finite time, but also to achieve that the error output converges to zero asymptotically. The effectiveness of the results is illustrated by a simulation example.

Research on FMC analytic algorithm of PBN track transition coding
Guangming ZHANG, Siqian REN, Weiwei ZHAO, Xiuyi LI
2023, 34(4):  1042-1052.  doi:10.23919/JSEE.2023.000090
Abstract ( )   HTML ( )   PDF (3579KB) ( )  
Figures and Tables | References | Related Articles | Metrics

Airborne navigation database (NavDB) coding directly affects the result of analysis on the instrument flight procedure by the modern aircraft flight management computer (FMC). A reasonable flight track transition mode can improve the track tracking accuracy and flight quality of the aircraft. According to the path terminator (PT) and track transition characteristics of the performance based navigation (PBN) instrument flight procedure and by use of the world geodetic system (WGS)-84 ellipsoidal coordinate system, the algorithms for “fly by” and “fly over” track transition connections are developed, together with the algorithms for coordinates of fix-to-altitude (FA) altitude termination point and heading-to-an-intercept (VI) track entry point and for track transition display of the navigation display (ND). According to the simulation carried out based on the PBN instrument approach procedure coding of a certain airport and the PBN route data at a high altitude, the algorithm results are consistent with the FMC-calculated results and the actual ND results.

Discussion on neighborhood optimal trajectory online correction algorithm and its application range
Wanli LI, Jiong LI, Humin LEI
2023, 34(4):  1053-1062.  doi:10.23919/JSEE.2023.000097
Abstract ( )   HTML ( )   PDF (4270KB) ( )  
Figures and Tables | References | Related Articles | Metrics

This paper presents a neighborhood optimal trajectory online correction algorithm considering terminal time variation, and investigates its application range. Firstly, the motion model of midcourse guidance is established, and the online trajectory correction-regenerating strategy is introduced. Secondly, based on the neighborhood optimal control theory, a neighborhood optimal trajectory online correction algorithm considering the terminal time variation is proposed by adding the consideration of terminal time variation to the traditional neighborhood optimal trajectory correction method. Thirdly, the Monte Carlo simulation method is used to analyze the application range of the algorithm, which provides a basis for the division of application domain of the online correction algorithm and the online regeneration algorithm of midcourse guidance trajectory. Finally, the simulation results show that the algorithm has high real-time performance, and the online correction trajectory can meet the requirements of terminal constraint change. The application range of the algorithm is obtained through Monte Carlo simulation.

Reliability modeling of mutual DCFP considering failure physical dependency
Ying CHEN, Tianyu YANG, Yanfang WANG
2023, 34(4):  1063-1073.  doi:10.23919/JSEE.2023.000108
Abstract ( )   HTML ( )   PDF (5073KB) ( )  
Figures and Tables | References | Related Articles | Metrics

Degradation and overstress failures occur in many electronic systems in which the operation load and environmental conditions are complex. The dependency of them called dependent competing failure process (DCFP), has been widely studied. Electronic system may experience mutual effects of degradation and shocks, they are considered to be interdependent. Both the degradation and the shock processes will decrease the limit of system and cause cumulative effect. Finally, the competition of hard and soft failure will cause the system failure. Based on the failure mechanism accumulation theory, this paper constructs the shock-degradation acceleration and the threshold descent model, and a system reliability model established by using these two models. The mutually DCFP effect of electronic system interaction has been decomposed into physical correlation of failure, including acceleration, accumulation and competition. As a case, a reliability of electronic system in aeronautical system has been analyzed with the proposed method. The method proposed is based on failure physical evaluation, and could provide important reference for quantitative evaluation and design improvement of the newly designed system in case of data deficiency.

Rolling bearing fault diagnostics based on improved data augmentation and ConvNet
Delanyo Kwame Bensah KULEVOME, Hong WANG, Xuegang WANG
2023, 34(4):  1074-1084.  doi:10.23919/JSEE.2023.000109
Abstract ( )   HTML ( )   PDF (5926KB) ( )  
Figures and Tables | References | Related Articles | Metrics

Convolutional neural networks (CNNs) are well suited to bearing fault classification due to their ability to learn discriminative spectro-temporal patterns. However, gathering sufficient cases of faulty conditions in real-world engineering scenarios to train an intelligent diagnosis system is challenging. This paper proposes a fault diagnosis method combining several augmentation schemes to alleviate the problem of limited fault data. We begin by identifying relevant parameters that influence the construction of a spectrogram. We leverage the uncertainty principle in processing time-frequency domain signals, making it impossible to simultaneously achieve good time and frequency resolutions. A key determinant of this phenomenon is the window function’s choice and length used in implementing the short-time Fourier transform. The Gaussian, Kaiser, and rectangular windows are selected in the experimentation due to their diverse characteristics. The overlap parameter ’s size also influences the outcome and resolution of the spectrogram. A 50% overlap is used in the original data transformation, and ±25% is used in implementing an effective augmentation policy to which two-stage regular CNN can be applied to achieve improved performance. The best model reaches an accuracy of 99.98% and a cross-domain accuracy of 92.54%. When combined with data augmentation, the proposed model yields cutting-edge results.