Abstract:At present,existing crack initiation methods and crack propagation methods which can be used to predict rubber fatigue life are not enough,a new method for rubber multi-axis fatigue life prediction based on incremental crack propagation and cracking energy density is proposed.First,a mathematical model of cracking energy density under the finite strain assumption is introduced.Second,assuming that the cracking energy density changes with crack propagation,a functional relationship between the energy release rate and the renewal cracking energy density is established.Finally,a multi-axis fatigue life predic-tion model of rubber structure based on incremental crack propagation is established.This model can sim-ultaneously achieve prediction of fatigue life and explicit expansion of cracks.The effectiveness of the pro-posed method is verified by the numerical examples.The results indicate that the fatigue life prediction model based on incremental crack propagation is good in prediction accuracy in consideration of high-strain conditions,and the prediction accuracy increases by about twice as compared with the classical crack ener-gy density life prediction model.

Abstract:In existing malware visualization classification models,there are inadequate accuracy and ro-bustness.For this reason,this paper proposes a malicious code visualization classification method CasKD-Net(Cascade DenseNet with KAN)based on an improved DenseNet.The CasKDNet is to realize the im-provements in accuracy and robustness by three key technologies.Firstly,a cascaded classifier structure is constructed to enhance the feature discrimination ability of texture similar families.Secondly,the KAN structure is used to replace the multi-layer perceptron in the DenseNet network,optimizing the non-linear expression ability of the feature extraction process and improving the overall accuracy of the model.Final-ly,the FFM image restoration algorithm is used to enhance the training set and improve the robustness of the model.It appears from the experimental results on the malicious code dataset Malimg that the CasKD-Net model achieves 99.69％of classification accuracy,and is superior to the existing research methods.Furthermore,in the context of white box attacks,the success rate of FGSM and I-FGSM algorithms at-tack against the CasKDNet only serves as 12.7％and 37.5％respectively,and the model is valid in pre-venting adversarial attacks.

Abstract:In order to make aero-engine exhaust system meet the needs of high radar stealth performance and thrust vectoring performance,taking the spherical convergent vector nozzle as the research object,the influence of the geometric parameters of the inner-flow shielding strut modification on the backward radar cross-section(RCS)of the nozzle is analyzed,and the iterative physical optics method is used to calculate the backward RCS of the spherical convergent vector nozzle under different strut inclination angles,bevel angles,and the number of struts.The calculation results show that within both the two detection planes,the inner-flow shielding struts can effectively reduce the backward RCS of the nozzle.With regard to the strut inclination angle,when the strut inclination is at an angle of 50° to the pitch detection plane,the RCS reduction is best in effect,i.e.reaching 71％.When the strut inclination is at an angle of 30° to the yaw detection plane,the best RCS reduction reaches 21％.With regard to the strut bevel angle,when the strut bevel is at an angle of 45° to the pitch detection plane,the RCS reduction is best in effect,i.e.reaching 64％.When the strut bevel is at an angle of 40° to the yaw detection plane,the best RCS reduction reaches 19％.When the number of struts is 16 in all,the best RCS reduction effect can be obtained within both the detection planes.

Abstract:Aimed at the problems that the transmission bandwidth in low earth orbit(LEO)satellite net-works is difficult to support the forwarding of large data volumes and make the links periodically discon-necting and reconnecting according to the needs,a satellite network multi-path routing optimization strate-gy is designed based on link state awareness.The dynamic topology of LEO satellites and the state of in-ter-satellite links are modeled,and an optimization of multi-path selection is constructed.In order to en-hance the efficiency of solving the optimization problem,a multi-path selection algorithm is devised based on state monitoring and path prediction.The algorithm is to utilize time slot partitioning of the topology for mitigating its dynamism and reducing complexity by screening inter-satellite links based on monitored states,and predicting link availability based on ephemeris data to avoid massive packet loss caused by link interruptions.On the computation phase,in comprehensive consideration of real-time link states such as transmission delay,transmission bandwidth,and transmission success rate,the optimal multiple paths are selected according to the load variations to increase the throughput of satellite networks and reduce trans-mission delay.The simulation results demonstrate that compared to the traditional routing algorithms such as contact graph routing(CGR),shortest path first(SPF),and equal-const multi-path(ECMP),when the load is 8 Gbps,the proposed scheme in the aspects of transmission delays,is 16.9％,11.4％,and 7.1％lower than the three algorithms respectively,and in terms of network throughput,is 34.4％,26.9％,and 15.6％higher than the three algorithms respectively.The transmission success rate is 15.3％,9.6％,and 5.6％higher than that of the three algorithms,respectively.

Abstract:The navigation parameters used in existing geomagnetic navigation research are most of original geomagnetic parameters-based.The selection of geomagnetic parameters may affect the efficiency of navi-gation,and there is man-man error in the process of selecting navigation parameters.For the above-men-tioned reasons,a search navigation method is proposed based on the geomagnetic feature extraction.The extracted geomagnetic features able to comprehensively describe the magnetic field information of the loca-tion are taken as new navigation parameters by the principal component analysis method.In combination with the existing evolutionary search strategies and gradient descent methods,the extracted geomagnetic features are fully utilized for taking as new navigation parameters to guide the carrier to continuously ap-proach the target and achieve navigation objectives.The experimental results show that the traditional search navigation method based on the original geomagnetic parameters has an iteration step of 647 for the objective function,while the search navigation method based on extracting principal component features has an iteration step of 564 for the objective function.However,the evolutionary gradient navigation method based on principal component features proposed in this paper has an iteration step of 238,signifi-cantly reducing the iteration step number,improving navigation efficiency and the consistency of sub ob-jective function convergence.The original magnetic field information is utilized further in autonomous re-mote navigation,and there are good prospects.

Abstract:With the increasing needs of data transmission rate in wireless communication,it has become a job with more challenge to study how to optimize signal detection from noise in multipath channels and to construct the best receiver in keeping with the certain criteria.From the point of view of optimal receiver design,the function of matching filter is derived and analyzed deeply,and a method of solving the whiten-ing filter by means of minimum phase channel is proposed.The effectiveness and robustness of the se-quence detection receiver scheme for eliminating inter-code crosstalk are verified,providing a performance limit reference for receiver design in multi-path channels.The simulation results show that:the perform-ance of sequence detection based on Viterbi algorithm is superior to that of MMSE equalizer,even more than 10 dB in the bad channel with spectrum zero.For the number of less than 5 paths,the sequence de-tection method based on Viterbi algorithm can not only be adoptable but its computational complexity can also be acceptable.

Abstract:Jamming resource optimization is an important link in current EW mission planning.Aimed at the problems that multi-objective optimization algorithm is easy to fall into local optimization and there is too much trouble in converges in three-objective optimization,a multi-aircraft jamming resource optimiza-tion method is proposed based on improved multi-objective moth-to-flame algorithm.Firstly,based on the multi-objective moth-to-flame algorithm,Tent chaotic map is utilized for initializing the population,in-creasing the diversity and uniformity of the solution and improving the search ability of the algorithm.And then,the induction of decision factor and Lévy flight is to make the algorithm accept not only the current solution with a certain probability,but also jump out of the current solution according to the disturbance and search again,enhancing the search ability of the algorithm.Finally,the widely distributed reference points are used to solve the convergence problem of the multi-objective moth-to-flame algorithm in the three-objective function.The simulation results show that this algorithm is better in convergence and pop-ulation diversity than the MOEA/D algorithm and the NSMFO algorithm,and the convergence result of this method is stable,achieving the purpose of assisting combat decision.

Abstract:A one-dimensional simulation model is established in aircraft evaporative cooling cycle system with the impact of refrigerant charge being exerted on the performance in evaporative cooling cycle system.The paper calculates and analyzes the effects of refrigerant charge on compressor speed,compressor dis-charge characteristics,compressor power consumption,COP(coefficient of performance),and sub-cool-ing degree at different temperatures within the flight envelope.The results show that whether to over-charge or undercharge exerts all an unfavorable influence on the performance of system.At different ambi-ent temperatures,there are differences in the optimal charge amounts corresponding to the minimum com-pressor power consumption and the maximum system COP in increasing with the rise at ambient tempera-ture.In addition,the sub-cooling degree is regarded as a key indicator for evaluating the optimal refriger-ant charge amount,whose stable section is only related to the charge amount within the range.The com-pressor power consumption and the system COP under different working conditions are good in perform-ance.The research results provide a basis for determining the optimal refrigerant charge flow in the air-craft evaporative refrigeration cycle system.

Abstract:Whether to synchronize high-precision for their clocks inter aircraft cooperation in aviation swarm is one of key technologies in imposing restriction on the effective cooperation among aircraft.And a high-precision clock self synchronization algorithm is proposed.Taking RTT synchronization as the princi-ple,a synchronization error solution equation containing relative radial motion velocity variables is derived by analyzing the impact of relative motion of aviation nodes on synchronization accuracy.On this basis,a pseudo code acquisition algorithm is proposed based on segmented correlation and frequency domain pro-cessing,and the algorithm has a comparative strong ability in anti-Doppler frequency shift.The simulation experiments show that this algorithm can effectively eliminate clock synchronization errors caused by rela-tive motion with the signal time of arrival(TOA)estimation and the Doppler frequency offset estimation being high in accuracy,and low in computational complexity.

Abstract:The distributed air defense and antimissile kill web are characterized by complex networks of heterogeneous nodes and multiple links,and are aimed at the problems that research on the functional het-erogeneity of nodes is insufficient,and assessment indexes are homogeneous in value evaluation of nodes,a hyper-network model is constructed for the reconnaissance,accusation and fire sub-network,and a method of node value evaluation is proposed based on the theory of hyper-network.The node topology,perform-ance index and connection relationship are taken into consideration by the method comprehensively from obtaining the overall effectiveness of the kill web through the improving weighted expert scoring and en-tropy-grey correlation-TOPSIS algorithm to measuring the node value importance by node deletion meth-od,achieving the quantitative analysis of the node value.The method is valid,providing a new idea for the value evaluation of the nodes in air defense and antimissile kill web.

Abstract:In order to solve the problems that dimensionality is high and feature extraction from flight tra-jectory data is inaccurate at the terminal area,this paper proposes a trajectory pattern recognition method based on LSTM-DAE spectral clustering.Firstly,the paper plans to achieve dimensionality reduction and to extract feature from the processed trajectory dataset by the LSTM-DAE network,and then proceed to even more accurately capture the nonlinear features of trajectory.Secondly,spectral clustering is employed by using the extracted trajectory features to complete pattern partitioning.Finally,an example analysis is conducted on the entry flight trajectory data at Tianjin Binhai Airport.The experiment shows that this method can accurately cluster high-dimensional flight trajectories after extraction,and can be divided into six categories of trajectory clusters,achieving still higher clustering quality.And the method can provide support for effectively identifying flight trajectory pattern features at the terminal area.

Abstract:In view of the problems that samples are limited in size and difficult in establishing a deep model for predicting the number of faults to evaluate the support performance of new mechanical equipment at the stage of testing and identification,a Score evaluation index is proposed by adopting the"transfer learn-ing"method.Large scale mature equipment data was used to assist in training the new equipment fault prediction model.Starting from the perspectives of samples,features,and models in transfer learning,with a focus on deep model-based transfer,this study conducts research on predicting the number of faults.The example shows that the precision of the fine-tuning based model deep transfer has increased by 46.55％and 164.87％respectively in the root mean square error and Score,while the standard deviation has decreased by 86.71％and 91.41％respectively.Far superior to the prediction methods based on sam-ple and feature in design applications of transfer learning and seven typical comparative models,the data-driven advantages of deep learning are fully utilized.With better performance in prediction accuracy,effec-tiveness,and stability,it is conducive to evaluating the support performance of new equipment and promo-ting the construction of equipment test and evaluation.

Abstract:Aimed at the problems that timing synchronization algorithm is sensitive to frequency offset caused by multipath fading and Doppler spread in aviation communication,and is relatively poor in per-formance under condition of low signal-to-noise ratio,a synchronization algorithm is proposed by using constant envelope zero autocorrelation(CAZAC).A leading sequence with conjugate symmetry features is designed by using CAZAC sequences,and based on the structural features of the leading sequence,the cross-correlation is utilized for designing a timing metric function for accumulating peaks.The correlation value at the correct timing position is used to estimate the decimal frequency offset,and lastly,a pair of i-dentical CAZAC sequences is used to complete the second decimal frequency offset estimation.The simula-tion results show that when the correct detection probability reaches 100％,the performance of the im-proved algorithm improves by 3 dB.

Abstract:Aimed at the problem that the Grey Wolf Optimization (GWO)algorithm is easily bogged down in local optimization in solving function optimization,this paper proposes a nonlinear adj ustment strategy by adopting sinusoid,logarithmic,tangential,cosine and quadratic curves.And at the same time a strate-gy on mutating position of the agents is presented,whose position is influenced by fitness value.The ex-perimental results for three standard test functions show that the proposed cosine and the quadratic curve strategies are superior to the classical linear strategy,and the others such as the sinusoid strategy,the log-arithmic strategy,and the tangential curve strategy are inferior to the linear strategy.