Abstract:Traditional and neural network-based methods being in existence of rolling bearing fault prediction, a dual adaptive sliding time window fault prediction model is proposed. Firstly, the rolling bearing vibration signal is mapped into fault features characterized as its degradation state by setting up a state estimation non-linear operator capable of removing correlations. Secondly, taking a loss function as a criterion, an adaptive update mechanism for the model parameters is set up, and a sliding time window capable of adaptively selecting the data length is constructed. Finally, the validity of the proposed failure prediction model is verified by simulating the occurrence of failures under the combined sudden and gradual failures in practice using the whole life cycle data of rolling bearings released by Xi’an Jiaotong University. The experimental results show that the prediction model proposed can accurately identify at the beginning moment and at the failure moment for the rolling bearing at the degradation stage, and truly reflect the trend of equipment performance degradation. The prediction error is only 0.068% and the prediction time is only 1.385% of the interval between failures, meeting the needs of rolling bearing failure prediction under condition of complex operation.

Abstract:Calibration test is the key segment of flight load measurement by strain bridge method. In order to simulate the characteristics of aerodynamic load pressure center continuously changing with aircraft maneuvering in real flight, a load calibration test method with variable pressure center is proposed. The force balance and constraint loads of the whole aircraft are calculated and analyzed, and the multi-point coordinated loading system is applied to the wing load calibration test of an aircraft. The load model is verified by the variable pressure center loading condition, and the error caused by the pressure center change is clear. The model is optimized by adjusting the pressure center distribution of the modeling condition, and the accuracy and application scope of the load model are improved. The results show that the left wing and right wing of the longitudinal maneuvering flight load measured by the optimized load model are symmetrical, and the change of pressure center is within the verification range, so the measurement results are reasonable and reliable.

Abstract:As resource replenishment, airdrop is an important means. Compared with traditional land and water transportation, the airdrop can guarantee security efficiently for the material, and its capacity has an indispensable impact on the formulation of airdrop plans. At present, the capacity theory understanding of airdrop site scene is fairly superficial. Firstly, the operation characteristics of the airdrop site are studied and analyzed, and the definition of the capacity of the airdrop site is proposed in combination with the operation and structural characteristics of the airdrop site scene, Secondly, the possibility of aircraft chasing situation in the operation on airdrop ground is analyzed, and a capacity calculation model of the airdrop ground is established. Finally, taking a certain airdrop site as an example, the theoretical capacity of the airdrop site under this structure is calculated, and the reliability of the theoretical model is verified by Monte Carlo simulation. At the same time, the variation law of airdrop field capacity under condition of the probability change of aircraft at running speed and less than the minimum safety interval is explored. The research content of the paper can supply formulation of airdrop plan with scientific and reliable theory.

Abstract:Aluminum-silicon co-infiltration coating is usually sprayed on the surface of aeromotor blades which is at high temperature, so as to provide critical protection to prolong the service life of aeromotor blades. the thickness of the coating can be monitored through X-ray fluorescence non-destructive testing technology in a quick and expedient manner, but it is not accurate enough to detect the thickness of multielement double-layer aluminum-silicon co-permeated coating. Further the study of the relation between the thickness of multielement double-layer aluminum-silicon co-permeation coating and the X-ray fluorescence detection value. One method based on X-ray fluorescence nondestructive testing technology are proposed in this paper to reduce the error of X-ray fluorescence measurement. Moreover, the feasibility of the method will be examined by verifying the linear relation between the fluorescence pair value of Molybdenum and the thickness of the first layer. Lastly, based on the principle of X-ray fluorescence absorption and divergence, we will fit a computational model for the relation between the thickness of the intermediate layer and the fluorescence of molybdenum. We find the coating thickness of the experimental sample has a critical value, so that the relationship between thickness and fluorescence value is not monotonously increasing or decreasing. When the mass fraction of Molybdenum in the film is greater than that in the substrate, the fluorescence effect received by the instrument decreases first and then increases with the increase of the film thickness, with the critical value of about 14.3μm. The influence of interpenetration of elements on thickness measurement of Aluminum-silicon permeability coatings less than the critical value can not be ignored.

Abstract:To address the complex feature extraction problem in traditional rolling bearing fault diagnosis, an end-to-end rolling bearing intelligent fault diagnosis method based on a multi-attention mechanism is proposed by introducing channel attention and spatial attention mechanism using the feature that deep residual network can enhance the nonlinear characterization ability of the diagnosis model. Firstly, the vibration velocity and displacement signals are obtained by integrating the original vibration acceleration signal. Secondly, the three types of signals are combined into an image with feature enhancement and input to a deep residual network combined with a multi-attention mechanism for feature extraction. Finally, a multi-classification function is used to complete the rolling bearing fault classification. The validation was carried out on a local laboratory-bearing dataset, and the results showed that the diagnostic accuracy of the proposed method reached 97.50%. The feasibility and effectiveness of the end-to-end rolling bearing intelligent fault diagnosis method based on a multi-attention mechanism are verified, which can support the accurate fault diagnosis of rolling bearings.

Abstract:An evaluation level of the integrated combat scheme is determined on the basis of analyzing the relationship and capability influencing factors based on air and missile defense in integrated operation, constructing an index system for integrated operation capability, considering the uncertainty of experts' judgment information on evaluation objects comprehensively, applying intuitive fuzzy entropy to the index weights and expert weights, and utilizing triangular whitening function for evaluating grey clustering to indexes at different levels. The feasibility and scientism of the intuitive fuzzy entropy valuation method are verified by the example. The evaluation results can provide the auxiliary decision support for the integrated operation plan.

Abstract:Direct path interference, because of the small path loss, has a much greater interference intensity than the reflected echo of the target, which directly affects the detection performance of the passive bistatic radar target. Intelligent reflect surface (IRS) has the ability to control the propagation of electromagnetic waves. For this reason, an IRS is placed near the radar receiver. By optimizing the beamforming of the channel receiver and the reflected phase of the passive IRS, the direct path interference power entering the receiver can be constrained to a certain range, and the target echo power can be maximized, thus improving the target detection performance. The joint optimization problem is transformed into two homogeneous and non-homogeneous quadratically constrained quadratic programming problems by using alternating optimization method. The optimum solution for beamforming of receiver and reflective phase of IRS is obtained by using semi-definite relaxation method and Gaussian randomization method. Computer simulation analysis shows that the detection performance of passive bistatic radar under direct path interference can be effectively improved by using the IRS.

Abstract:The SAR-BagNet model is an interpretable deep learning model used for Synthetic Aperture Radar (SAR) image recognition. In order to maintain the interpretability of the SAR-BagNet model while also achieving high recognition accuracy, this paper uses the SAR-BagNet model as a foundation and incorporates spatial attention and coordinate attention mechanisms into the model framework. Experimental results on the MSTAR dataset demonstrate that the spatial attention and coordinate attention mechanisms enhance the SAR-BagNet model's ability to acquire global information. This enhancement effectively improves the model's recognition accuracy and decision rationality without compromising its interpretability.

Abstract:Aimed at the problems that as for missile-borne SAR at a large-angle dive movement stage, the echo signal has serious coupling and space-variance due to its complex flight characteristics to make the traditional focusing algorithm invalid, an improved frequency domain imaging algorithm of polar format algorithm (PFA) is proposed. Firstly, a range model of large-angle dive missile-borne SAR is established, and the range history is expanded in Taylor series. And then the high-precision two-dimensional spectrum of echo is derived by the series inversion method. Finally, the high-order cross-coupling terms are decomposed in the two-dimensional frequency domain. Moreover, a new two-dimensional interpolation mapping function is derived, greatly improving the performance of focused image. Compared with traditional PFA, the proposed algorithm is more suitable for large-angle dive missile-borne SAR, and is valid. 

Abstract:Aimed at the problems thatthe existing Uniform Linear Array (ULA) far-field narrowband non-coherent multi-target estimation algorithms is poor in adaptability to low Signal-to-Noise Ratio (SNR), small snapshots in adaptability, high in computational complexity, and existing Deep Learning (DL) approaches are difficult to effectively extract the complex-valued features of data,a Direction of Arrival (DOA) estimation method based on Deep Convolution Neural Network (DCNN) is proposed . This method is to transform the DOA estimation problem into an inverse mapping problem from the array output covariance matrix to the target DOA, and to utilize the Hermitian characteristic of the array output covariance matrix for extracting the real part, imaginary part, and phase characteristics of an upper triangular array,building input data of a network, and building a deep convolutional neural network with a three-dimensional convolution layer to extract data features, and the labels of the network correspond to the DOAs,realizing the DOA estimation of multiple sources. Theexperimental simulations show that the method can fully extract spatial features, improve DOA estimation accuracy and reduce the complexity of the algorithm.Under condition of low SNR and small snapshots, the estimation accuracy of the proposed method is significantly better than that of the MUSIC, the ESPRIT and the ML algorithms.

Abstract:Aimed at the problems that under some specified conditions, obtaining sufficient samples is so difficult that the accuracy of the BN parameters learned by the maximum likelihood estimation algorithm is often low, and multiple parent nodes collaboration to influence constraints is involved in some areas of practical application, a BN parameter learning method based on the modified multiplicative co-constraint under small data sets is proposed by drawing on the idea of PAVA order-preserving regression algorithm. First, the paper is to determine whether the parameters in the multi-parent part of the known sample data meet the needs of the multiplicative collaborative constraint. Secondly, both the left and right sides not to meet the needs of the multiplicative co-constraint are divided into wholes, and adjusted separately by using the PAVA algorithm. And then, for the adjusted whole, three correction methods with different weights are given to correct each parameter according to the amount of sample data corresponding to the combined state of different parent nodes, and gain mean final parameter learning result. Finally, the proposed method is validated by simulation using a classical grassland wetting network model. The experimental results show that the proposed method not only meets the needs of the multiplicative cooperation constraint under small data set conditions, but also the KL scatter is always lower than the other 2 methods in addition to that the running time is slightly higher than that of the other 2 methods by about 1×10-3s with minimal impact. Generally speaking, the proposed algorithm is superior to the other 2 methods in the comprehensive performance.

Abstract:At present, the countermeasure technology of malicious code is constantly changing, and new varieties of malicious code are emerging in endless streamto make the classification of malicious code face severe challenges. Aimed at the problemsthat features extracted are insufficient and low in accuracy by using current malicious code classification methods based on deep learning, a malicious code classification method (BiTCNSA) based on bi-directional temporal convolution network (BiTCN) and self attention mechanism is proposed. This method is combination of opcode features with image features to show different feature details, increasing feature diversity. The BiTCN is constructed to process the fused features, making full use of the pre and post dependencies of the features. The self attention mechanism is introduced todynamically adjust the data weight, further mining the correlation between the internal data of malicious code. The model is verified by using the Kaggle data set. The results show that the accuracy of this method can reach 99.75%, and the method is fast at convergence speed, lowin error, and better than the other models.

Abstract:Aimed at the problems that the current manned/unmanned cooperative combat mission is complex, the complexion of weapons and battlefield situation is changed increasingly, and the amount of information at the interactive interface increases rapidly, a model of manned/unmanned cooperative combat interactive mission based on PCTBTA is proposed. The model focuses on the influence of external environmental factors on the operator's human-computer interaction in manned/unmanned cooperative combat, and the default tasks of the operator under conditions of these influences. On the basis of comprehensively analyzing manned/unmanned cooperative combat mission process and its characteristics, constructing the conceptual model of interactive mission, and then determining five typical combat missions in manned/unmanned cooperative operation , and studying and analyzing the conceptual elements such as objectives, sequences, display variables, influencing factors, and operations of each task, an interactive mission model of manned and unmanned cooperative combat in a dynamic environment is established according to the conceptual model and conceptual element analysis of manned/unmanned cooperative combat. Compared with the CTT, GOMS and the other models, this model is capable of integrating external factors to change all the time in combat, and to further describe the interactive task process of manned/unmanned collaborative combat in a dynamic battlefield environment, providing a theoretical model for operators to efficiently complete interactive tasks in manned/unmanned collaborative combat, and laying a foundation for the optimization design of man-machine interface of various manned/unmanned collaborative combat information systems in the future.

Abstract:Now available research on the resilience in Command Information System is in ignorance of the impact of demand changes on its resilience process, and is unable to reflect the relationship between the resilience ability of the system after being attacked and task requirements. Three variables of normal performance, lowest performance and expected performance are introduced, and three resilience models of significantly increased demand, basically flat demand and significantly decreased demand are proposed. In view of the two stages of functional level decline and adaptive recovery in the process of resilience, a strategy of resistance enhancement and redundancy enhancement strategies is proposed, the measure effect function is defined, a resilience optimization model is established, and the approximate dynamic programming algorithm of the group knapsack problem is used to solve the problem. Taking a certain of Regional Joint Air Defense Command Information System as an example, a simulation experiment is conducted, and the impact of demand changes on the resilience optimization of the Command Information System is verified. The results show that in order to adapt to the operational requirements of different scenarios in complex and harsh battlefield environment, the Command Information System should select a reasonable resilience mode, and the comprehensive relevant factors, such as enemy's attack intensity, funding budget, coping strategies, and so on, should be taken into account.

Abstract:he messages sent by the data link often carry tactical information. In this paper, the specific content of the message being unknown, the feasibility of cognizing typical data link tactical behaviors is experimentally explored. Firstly, according to the characteristics of formatted messages, the structure of the V-series messages with a tactical message is simulated, and simultaneously, the modulated signal is of data link generated. And then the wavelet packet decomposition method is utilized for extracting the time-frequency features and constructing the time-frequency graph datasets. Deep learning is used to classify the signals containing different tactical messages, and the attention mechanism module CBAM applied to CNN is introduced to further improve the recognition accuracy.The experiments prove that even if the message structure is unknown, the corresponding tactical task can be recognized by identifying the physical layer signals containing different V-series tactical messages. Finally, based on the experimental results, a new data link tactical behavior reconnaissance process is proposed.