Abstract:A large number of studies shows that multi-agent reinforcement learning algorithm is difficult to converge when the number of agents increases.However,centralized training can overcome the challeng-ing non-stationary environment under the condition of predicting the action of the opponent,achieving still better learning results.In the light of predicting opponent's continuous decision-making,the time series opponent modeling method based on gated recurrent neural network is proposed.Based on the multi-agent reinforcement learning framework MADDPG,the opponent's state,environmental state,and our agent's action are utilized for constructing time series to serve as the input of gated recurrent units,aiming to cap-ture the time series information of opponent agents.A length-variable model inference method is presented to determine the optimal length of the input time series,avoiding the misuse of stale training samples in-curred by the adjustment of the opponent's strategies.The simulation results show that the proposed op-ponent model can effectively predict the opponent's future actions in combination with the length-variable model inference method.Compared with the SAM method,an improvement on the prediction accuracy is over 6％ by this method.

Abstract:In the light of the problems that there still remain numerous constraints and difficulties in com-prehensive optimization in wartime for forward assembly path planning of multiple flight formations,a collaborative planning method is proposed based on the improved fungal growth optimization(IFGO)algo-rithm.Firstly,a population optimization strategy is proposed,and the Adaptive Boundary Reflection Mechanism is introduced to accelerate convergence efficiency.Secondly,a new perturbation mechanism is constructed to further enhance global search capability.Finally,flight conflicts are solved by adopting a multi-hypha collaborative optimization strategy,making optimal solution be even more close in under the premise of no conflicts.This paper constructs a comprehensive cost function in consideration of threats,flight distance,and fuel consumption based on the function that is used to make comparative experiments between IFGO and other algorithms with its performance verified on CEC2022 test functions.The results show that making 10 times repeatedly experiments under simple and complex environments,its average cost function is 7.89％ and 9.62％ lower than the comparison algorithms respectively,and ranks first in rank mean among the CEC2022 standard test functions.

Abstract:In view of the problems that fight jets have to confront danger of complex battlefield environ-ments while carrying out strike missions in accordance with the needs of operational intention reached,an improved particle swarm optimization(PSO)algorithm is proposed.The algorithm,first,is to convert cartesian coordinates into spherical coordinates,and then,introduce adaptive weight factors to solve the problem that the PSO algorithm is too slow at converge speed and the outcome of calculation is optimum in local solution.It formulates various constraints along the flight path into different cost functions,with the PSO problem being to identify the path with the minimum cost value.The cost functions include the air-craft's track length,pitch angle variation,altitude difference variation,obstacle avoidance,and smooth-ness constraints.The results demonstrate that the PSO algorithm combined with spherical coordinates and adaptive weight factors is prior to the traditional PSO,the ant colony optimization(ACO),and the genetic algorithm(GA)in flight path planning within complex environments.

Abstract:In research and application of meter-wave radar technology,detecting low-altitude target is al-ways a highly challenging topic.Based on the specular reflection signal model of a monostatic meter-wave FDA-MIMO radar,and in view of the problem that accuracy in the joint range-angle estimation of ultra-low-altitude targets by the maximum likelihood(ML)and generalized multiple signal classification(GMU-SIC)algorithms is insufficient,a direct-wave matching multiple signal classification(DM-MUSIC)algo-rithm is proposed.This algorithm is to construct a direct-wave matching matrix to effectively eliminate the interference of reflected waves on the direct wave,and then restore the performance of the traditional MU-SIC algorithm.The results of simulation experiments show that under conditions of ultra-low-altitude,the DM-MUSIC algorithm has the best spatial spectrum resolution performance.Its joint range-angle estima-tion accuracy is significantly better than that of the ML and GMUSIC algorithms,and is also low in com-putational complexity.This enables the algorithm to process data more quickly and achieve more efficient detection and tracking of ultra-low-altitude targets.And this provides a feasible solution for the practical application of meter-wave radars in the field of low-altitude detection and has important value in both theo-retical research and practical applications.

Abstract:Three-dimensional wind field structures with single,asymmetric,and twin microbursts,and their impacts on flight safety are studied systematically.First,a computational fluid dynamics(CFD)model is constructed by using FLUENT to simulate the microburst-induced wind fields.And then,on the basis of the simulated data,airborne weather radar Doppler echoes are calculated to analyze the spatial dis-tribution and signal characteristics of each microburst type.Finally,an average F-factor is employed to quantitatively assess the threat level to aircraft.The results show that the single microburst produces a strong symmetric downdraft comparatively,the wind speed peaks of the asymmetric microburst and spa-tial shifts are closely related to the ambient wind direction,and the twin microburst generates intense up-drafts between converging downdrafts.Among the three,the twin microburst is the most risk to flight safety,whereas the asymmetric type is relatively the lowest.

Abstract:Traditional supervised classification method is usually used to an assumption that the probability distributions of training data and testing data are consistent.Such a method is significant limited,whereas it is caught in the domain discrepancy problems caused by heterogeneous polarimetric synthetic aperture radar(SAR)images.For the mentioned-above reasons,this paper innovatively proposes a polarimetric SAR image classification method in combination of Cloude decomposition with a deep domain-adversarial neural network(DANN).Firstly,the method is to utilize the features obtained from Cloude decomposi-tion for taking as an input of the DANN.Through the adversarial training of the domain classifier and the class classifier in the DANN,domain-invariant features are extracted,and then the classification of hetero-geneous polarimetric SAR images is achieved.The experimental results show that the method effectively overcomes the limitations of the traditional supervised method in domain adaptation.In two sets of trans-fer comparison experiments involving five actual polarimetric SAR images,the overall classification accu-racy is significantly enhanced compared to the direct transfer method and the DANN method utilized raw data as input.

Abstract:Aimed at the problems that high cognitive load and information oversaturation are present in the process of human-computer interaction for combat personnel,this paper explores a design method of hu-man-computer interface layout and display form based on intuitive interaction.Three kinds of intuitive in-terface layout and display forms are proposed for the main screen and secondary display interface of typical situational display interface,and a subjective-objective evaluation system is constructed through eye-track-ing data collection and subjective questionnaire survey.The experimental results show that the combina-tion of surrounded centralized layout with drop-down menu in the main screen interface,and the combina-tion of split-column 2-type layout with hidden menu in the sub-display interface are best in the optimal per-formance.Based on this method,the traditional interface is intuitively designed by adopting the CRITIC-TOPSIS method to evaluate the experimental data of interface perception(the closeness of the traditional interface serves as 0,and the intuitive interface serves as 1),realizing that the design of intuitive interfaces could effectively reduce the cognitive load of the battlefield personnel and enhance the efficiency of informa-tion processing.The research results can provide an innovative method and practice path for the field of di-rect-sense interface design,and its subjective and objective comprehensive evaluation system has a certain application value in the evaluation of interface design.

Abstract:In view of the problems that in detecting defects often show on surface of aircraft by using exist-ing visual-based method,there remains a low accuracy in detection,high parameters and expend time is more in computational consumption,and such a method is difficult to balance accuracy improvement with model lightweight,a new high-precision and lightweight aircraft surface defect detection method,called CSM-YOLO,is proposed.First,C2f module in the backbone network replaced by C2f-SCSA module is to dynamically enhance multi-scale features and improve the model's ability to capture,extract,and utilize key feature information,solving the feature loss caused by down-sampling.Secondly,the Slim-Neck fea-ture fusion network is improved with cross-layer connection and applied to the model neck,realizing the boosted computational efficiency and the detection accuracy and simultaneously cutting the information loss.Lastly,MPDIoU Loss is used to enhance bounding box regression accuracy and improve the detection precision of small target defects and cut false and missed detections.The experiments show that the CSM-YOLO enables to achieve high precision and lightweight.A maximum detection accuracy of 88.34%on surface defects can be reached,and there is a 2.92%improvement compared with the baseline YOLOv8n model,and the improvement in computation is obvious compared with the YOLOv3-tiny,the YOLOv5n,the YOLOv5s,the YOLOv7-tiny,the YOLOv9t and the YOLOv12n respectively.In aspect of model pa-rameters and computation,for the CSM-YOLO,there is a parameter of 2.67×106/s and a computational cost of 7.68×109/s,reducing the baseline YOLOv8n by 0.34×106/s and 0.41×109/s respectively.And the CSM-YOLO is balancing the accuracy improvement with the model lightweight.Moreover,the CSM-YOLO delivered significant performance gains on the aircraft surface defect detection dataset,offering an effective automated solution for surface defect detection.

Abstract:Aimed at the problems that the traditional rigid MIMO antennas are large in space occupation,complex in structure,and manufacturing process is tedious,two types of flexible MIMO antennas,a sym-metrically arranged two-element flexible MIMO antenna and an orthogonally arranged four-element flexi-ble MIMO antenna,are proposed in this paper.Prototypes of both antennas are successfully manufactured on a flexible PET substrate by adopting the low-cost,simple and easy-to-modify inkjet printing technolo-gy.The measured results show that the reflection coefficient of the two-element flexible MIMO antenna is a range from 2.74 to 12.47 GHz,while that of the four-element flexible MIMO antenna is a range from 2.7 to 13.4 GHz,both of which are in excellent agreement with the simulation results.For the two-ele-ment flexible MIMO antenna,the T-shaped branch structure is utilized for making its isolation meet the needs of 15 dB,while the quaternary flexible MIMO antenna achieves an isolation of greater than 20 dB through the orthogonal configuration of the ports.These antennas are characterized by a low profile,low cost,transparency,flexibility,and simple structure,along with good diversity performance,making them suitable for communication application in requiring multiple antennas.

Abstract:This paper establishes a real-time method of assessing their discharge state with discharge abili-ty of nickel-cadmium(Ni-Cd)batteries being at high-altitude and at low-temperature in operating condi-tions in affecting the reliability of aircraft in emergency power systems.First,an electrochemical P2D model is utilized for constructing a Ni-Cd battery model.This method enables core parameters to be sensi-tive to temperature,and recognizes to be capable of characterizing the discharge state,and the impact of temperature on the discharge state is translated into its effect on these core parameters.And then,a rela-tionship between these core parameters and the direct current internal resistance(DCIR)is established.On the basis of this relationship,a equivalent circuit model is developed,and the model parameter identifi-cation is performed by using the Hybrid Pulse Power Characterization(HPPC)method.The results show that by analyzing the correlation of model parameters(particularly DCIR)with temperature and DOD,a mapping relationship among temperature,internal resistance,and discharge state is successfully construc-ted.A method capable of effectively detecting the discharge state of Ni-Cd batteries at different tempera-tures is designed and implemented.This method can be simultaneously applied to real-time in-flight status monitoring and ground maintenance evaluations.And this method significantly improves the assessment accuracy of Ni-Cd battery in operational status,providing critical technical support for ensuring flight reli-ability.

Abstract:Different types of sensors differ from radiation mechanisms.The images among multi-modal re-mote sensors are characterized by significant source and geometric distortions.For the non-linear radiation problem in SAR images,though the phase consistency methods have strong robustness,this method is sensitive to scale changes in processing images.In order to achieve multi-scale alignment of visible light and SAR remote sensing images,a multi-scale phase consistency feature-based alignment method is pro-posed.First,a multi-scale spatial domain is constructed by using the Gaussian functions,and the detec-tion of phase consistency is performed to the images at different scale layers to extract feature points.A multi-scale fusion algorithm is used to fuse the phase consistency feature points with multi-scale features.And then,a description based on the maximum amplitude response and direction index of the Log-Gabor filter in polar coordinates is constructed for alignment.Finally,the correct matches are identified by using the nearest neighbor ratio and fast sample consistency.The experimental results show that the proposed algorithm is significant superior to the multi-scale alignment of visible light and the SAR images,impro-ving the alignment accuracy.

Abstract:This study addresses the significant impact of interrupted-sampling repeater jamming(ISRJ)on linear frequency modulated radar target detection performance by introducing a novel anti-jamming tech-nique grounded in wavelet packet decomposition.The proposed method innovatively constructs a jamming suppression framework based on invariant feature extraction,leveraging waveform prior information of ra-dar signals.This approach overcomes the limitations associated with traditional strategies centered around"jamming parameter estimation+jamming suppression",which are prone to loss of target information.Experimental results indicate that the method demonstrates superior capability in preserving target infor-mation under conditions of low jamming-to-signal ratio,high overlap of jamming and target signals,and complex jamming scenarios,outperforming conventional time-domain and time-frequency domain methods by achieving target information loss of less than 1.67%.The method significantly enhances target detec-tion accuracy,showing robust performance in complex jamming environments and effectively mitigating ISRJ-induced jamming,thus exhibiting substantial application potential.

Abstract:In view of the problem that it is very difficult for traditional methods to capture the intricate and nonlinear relationship between flight states and fuel consumption,this paper proposes a method of predic-ting the consumption of fuel for aircraft in dynamic time-based on the LSTM-KAN Network.First,eight key flight state parameters from QAR(quick access recorder)data in the terminal area-including altitude,true airspeed,and wind speed-the model employs KAN layers with B-spline basis functions in combination with a basic output structure are utilized for accurately capturing the nonlinear relationships between flight states and fuel consumption.And then,the KAN network logging on the final time step of the LSTM net-work is to achieve high-precision modeling of the dynamic time-varying patterns in the consumption of fuel for aircraft.The experimental results demonstrate that the model achieves a mean squared error(MSE)of 0.001,with 98.32%of the test set exhibiting a root mean square error(RMSE)below 0.09 1 kg/h.Ad-ditionally,the coefficient of determination(R2)reaches even more 0.989 7,significantly outperforming traditional models such as MLP(Multilayer Perceptron),standalone LSTM,and Transformer.The find-ings can be applied to optimization of airline fuel efficiency and airspace operations,thereby promoting greener practices in civil aviation.

Abstract:To address issues such as single-point failures and insufficient anti-interference capabilities in traditional centralized control architectures for distributed electric propulsion systems,a decentralized co-operative control method is proposed.First,a distributed architecture based on dynamic traction consen-sus is designed to achieve high-precision speed synchronization of multiple motors through information co-ordination,eliminating reliance on a central controller.Second,a sliding mode observer is developed to re-construct composite disturbances in real time,enhancing control performance under disturbances.Finally,simulations are conducted using a hardware-in-the-loop platform.The results show that compared to cen-tralized control,the proposed method improves synchronization accuracy by 5.3 times,reducing the maxi-mum error from 4.69%to 0.77%.Fault redundancy and disturbance compensation performance are sig-nificantly enhanced,eliminating yaw caused by force imbalances and enabling stable flight of DEP-UAVs.This research provides technical support for the design of highly reliable electric propulsion systems in un-manned aerial vehicles.

Abstract:In order to improve the durability of assembled grooved pavement,assembled airports should be investigated for wear resistance of macroscopic grooved structures of cement concrete grooved pavement necessarily to obtain the optimal characteristic parameters of macroscopic grooved structures.A tire-as-sembled grooved pavement interaction simulation model is used to analyze pavement wear under aircraft impact taxiing.Wear tests of pavement specimens are conducted by using a steel wheel wear testing ma-chine,and pavement wear resistance is analyzed by using point cloud scanning technology.Through nu-merical simulations of the macro structural parameters of different groove surface features,the pavement wear volume and tire stress are obtained.Optimized macro structural feature parameters are derived.Based on the experimental tests and point cloud scanning data analysis,the relationship between different macro structural feature parameters of the grooves and the wear rate is revealed,leading to the determina-tion of the optimal macro structural feature parameters.The results show that the wear rate of the pave-ment with rectangular grooves is higher than that of the pavement with trapezoidal grooves.When the groove space of groove is between 35～40 mm,the width of groove is between 15～25 mm,the depth of groove is between 2～4 mm,and the chamfer of groove is at an angle of between 25°～60°,the pavement wear rate and tire stress are comparatively low,helping extend the service life of both the pavement and the tire.

Abstract:Spectrum resources being short at L-band in digital aeronautical communication system(LDACS),this paper introduces a cognitive radio(CR)technology to improve communication efficiency by reasonably utilizing its idle spectrum on the premises of normal operation of distance measuring equip-ment(DME)being not affected.Firstly,the system framework of LDACS-CR is designed and DME spec-trum occupancy is modeled by utilizing a two-state Markov chain,and then cyclo-stationary feature detec-tion is utilized for perceiving DME signals in combination with estimating channel average periods to ana-lyze their spectral idle time by existing methods.Finally,the throughput of LDACS-CR system is analyzed under conditions of different scenarios.The research results indicate that by introducing cognitive radio technology,LDACS systems can achieve higher data transmission rates,providing new ideas for spectrum planning and deployment.

Abstract:Aimed at the problems that as for an effective assessment of drone swarm operations,index sys-tems are complex and incomplete,subjectivity is strong,and deviation is large in assessment methods,this paper proposes a top-down architecture design-combined weighting-cloud model evaluation method.Firstly,the swarm combat system is modeled on the general view,combat view,and the system view by the system structure design method,and then the equipment system is decomposed to construct the swarm combat effectiveness evaluation index system,and the combination assignment-cloud model evaluation method is applied to assign and evaluate the index system,generating the evaluation cloud model of indica-tors of all levels and the comprehensive evaluation cloud model in combination of the cloud features with the cloud evaluation cloud model to generate the evaluation cloud model of all levels.Finally,it appears from the experiments that the assessment method is valid,and is superior to the traditional methods.

Abstract:In view of the problems that intelligent detection technology in the process of strip surface defect detection is low in accuracy,missing in detection and false in detection,a GSS-YOLO object detection op-timization algorithm is proposed based on YOLOv8n.The model is to organically integrate the Neck net-work of the GOLD-YOLO algorithm with the Ultralytics to improve the detection accuracy of the model for defects of different sizes and shapes.In order to balance the gap in identifying the effect of different de-fect types and reducing the complexity of the network structure,two lightweight modules in the Slim-Neck structure are introduced,i.e.lightweight convolution VoVGSCSP and efficient channel attention mechanism SimAM,enable to improve the detection accuracy and generalization ability of the model,and to simultaneously limit the expansion of the computational and weight volume of the model.It appears from relying on the classical strip surface defect dataset NEU-DET,thaw experiment,lateral comparison experiment that the average accuracy of the model is 3.7%higher than that of the benchmark model,and the gap between the detection accuracy of various defects is reduced,thus the detection accuracy meeting not only the requirements,but also guaranteeing at running speed.In comparison with the current main-stream models,this model has a certain advantage in detection accuracy,and is in reference value to the application of defect detection in actual industrial production.

Abstract:In the context of multi-UAV combat scenarios,there is a method,i.e.air combatant corridor planning,in its way of enabling rapid traversal through controlled areas and reaching the combat zone,and this method is based on Delaunay mesh triangulation.First,a search map is constructed by discretizating the distribution structure of battlefield airspace and by using Delaunay mesh triangulation.Then,the qual-ity of the generated mesh is evaluated and detected by calculating the aspect ratio,and then the mesh is op-timized to improve its quality.Finally,a bidirectional search strategy is designed to enhance the traditional A* algorithm,facilitating cost-minimized path planning on the generated map and achieving corridor of the planned path.The experimental results show that the air combat corridor planning method can effectively avoid threats,is valid and superior.

Abstract:Aimed at the problems that data processing capability and decision-making speed are limited in kill net combat system in complex battlefields,as well as the problems that requirements of data are large,generalization ability is weak,and interpretability is low in the existing AI decision-making methods,this paper proposes a decision-making method based on knowledge-enhanced Large Language Models(LLMs).The method is to optimize the kill net decisions by directly leveraging LLM to understand the battlefield situation and semantics,through integrating three major modules,i.e.from the environmental state map-ping,the knowledge-enhanced LLM decision-making,and the decision text to the agent behavior conver-sion.The results show that this method can effectively understand battlefield situations and formulate decision schemes,significantly reducing the dependence on large amounts of data while ensuring good interpretability.