Considering the issues existed in the research results of current combat intention identification and predictionsuch as too few typestoo coarse granularityand lack of targets and objective of attack in the selection of combat intentionsas well as the demand for intention recognition in the future equipment development and mosaic war environmentthis paper analyzes the formation process of combat intention and the relationship between combat intention and the behaviorand introduces the explicit and implicit attributes of combat intention.Combined with mosaic operational requirements and the reality of multi-domain joint operationsthe combat intention is decomposed by typesprocessesstates and rolesand a three-level granularity model of combat intention is establishedwhich realizes the multi-levelmulti-attribute and multi-granularity detailed modeling of combat intentionand provides reference for the identification and prediction of combat intention.

In view of the shortcomings of traditional Ant Colony Optimization (ACO) algorithm in path planning of Unmanned Aerial Vehicle (UAV)such as slow convergence speed and easy to fall into local optimal solutionan Improved Adaptive Ant Colony Optimization (IAACO) algorithm is proposed.Firstlyin order to make ants move in the direction of the target point with greater probability and improve the search efficiency of the pathan angle guidance factor is introduced into the transfer probability of ACO.Thenheuristic information adaptive adjustment factor is introduced to balance the convergence and global search ability of the algorithm.Finallyby defining length index function and angle index functionthe objective function of route optimization is further established,and the global optimization of UAV route planning is realized.Experimental results show that the improved algorithm converges fasterand the generated track is smoother and shorter.

In order to accurately estimate the target number of Probability Hypothesis Density (PHD) filtersa cardinality compensation method based on Information-weighted Consensus Filter (ICF) and data clustering is proposedand ICF is used for information fusion.In a dense clutter environmentwhen the noise and clutter in the measurement are hightracking loss will occurand the performance of target number estimation will be degraded.For this reasona cardinality compensation process is added to the PHD filterbased on an information fusion stepthe estimated cardinality obtained from the PHD filter and the measured cardinality obtained through data clustering are used to obtain the final target number estimation.In order to verify the performance of the proposed methodthe simulation is carried out and it is demonstrated that the tracking performance of multiple targets is improved.

Aiming at the problems of the traditional fractal detection algorithm in Fractional Fourier Transform (FRFT) domain in sea cluttersuch as difficulty in determining the optimal order and unstable performancean improved fast detection algorithm of small targets on sea surface based on fractal transform domain is proposed.The new algorithm uses the fast estimation algorithm to quickly determine the optimal order of the target echo in FRFT transformand after FRFT transformthe energy is accumulated and the signal-to-clutter ratio of the echo is improved.On this basisthe Improved Multifractal Detrended Fluctuation Analysis (IMFDFA) algorithm is used to study the multifractal of sea clutter.The target detection is carried out on the measured data of IPIX radar.The experimental results show that the optimal order optimization algorithm proposed in the new algorithm has stable performance.Compared with the traditional fractal algorithmthe performance of the IMFDFA algorithm is obviously improved and more stablewhich is beneficial to the application of the algorithm in detecting small targets on sea surface to a certain extent.

Aiming at the problem of how the attacking missile strikes the ground target effectivelywhich defends the attacking missile by launching defensive missilea three-body game engagement kinematics model is establishedan objective function integrating miss distance and required acceleration is constructed and a differential game guidance law considering first-order kinetics autopilot is proposed.On the premise of three-point guidance and interception of defense missilethe guidance strategies of proportional guidance and differential game guidance of attack missile are simulated and verified respectively.Finallythe effectiveness of the proposed method is proved by analyzing acceleration constraints and terminal conditions.

UAV formation topology optimization is of great significance for improving the coordination of UAV cluster task execution and communication transmission efficiency.Firstlythe index of residual energy imbalance of UAV formation is proposed.Based on the length of formation communication chainthe influence factors of network delay and residual energy imbalance are brought into the generation system of UAV formation information interaction topologyand multi-objectives are considered comprehensively to optimize the UAV cluster information interaction topology.By constructing a satisfaction deviation membership functionthe goal programming model is established to realize the synthesis of multiple goals.At the same timethe multi-tree structure is used for classification in topology generationand the artificial bee colony algorithm is improved to solve the modelwhich can support the cooperation of large-scale UAVs.Finallythe formation composed of 16 UAVs is simulated and analyzedand the rationality of the model and the effectiveness of the algorithm are verified.

Compared with the conventional pulse Doppler radar signal processingSpace-Time Adaptive Processing (STAP) expands the signal processing dimensionso that the clutter and the target can be distinguished in the joint space-time domain.Based on sparse representation theory and the sparsity of clutter spectrumSparse Recovery (SR)-based STAP realizes clutter suppression under the condition of a small number of training range cells.Aiming at the performance degradation of SR-STAP methods with unknown yaw anglea nonlinear regression-based clutter reconstruction STAP method is proposed.Firstlybased on SR clutter spectrumoutlier degree is used as the convergence objective to iteratively eliminate the scatter points deviating from the ridgeand the coordinate weighted nonlinear regression is performed to realize accurate estimation of the parameters of the clutter ridge model.Thenbased on the results of the first screeningthe clutter spectrum is estimated accurately by the nonlinear regression method again.Finallythe clutter reconstruction and suppression are completed based on the above estimation results.The simulation verifies the effectiveness of the proposed methodand compared with the existing STAP methodsit achieves better space-time frequency response and SINR losseffectively improving the performance of clutter suppression and moving target detection.

According to the performance characteristics of reentry cross-medium unmanned aerial vehicle this paper puts forward its operational application scenario and application assumptionand models and analyzes its battlefield superiority and operational effectiveness from the following aspectssuch as the penetration efficiency and battlefield survivability of the reentry cross-medium unmanned aerial vehicleand its impact on detection windowinterception and trackingand combat resources.

Aiming at the space-time cooperative problem of UAV swarm attacking moving targets on the sea surfacea designated time cooperative control method is proposed.The relative motion between each UAV and its neighbors is transformed into an optimal tracker with a specified end time.Different quadratic cost functions are designed for the two control objectives with the specified time parametersnamelystrike configuration generation and strike configuration maintenance.By solving the control law by Runge Kutta methodthe UAV cluster can be driven to accurately generate the strike configuration at the specified timerealize simultaneous strikeand maintain the strike configuration in the subsequent specified time periodso as to realize continuous strike.The simulation results show that the UAV cluster can generate and keep the desired formation configuration at the specified momentand track the moving target.

Synthetic Aperture Radar (SAR) is a coherent imaging system.The images generated from SAR are often polluted by speckle noisewhich results in low accuracy in subsequent segmentation and recognition of SAR images.Aming at the problem of image pollutiona SAR image denoising network model Re-GAN is designedwhich combines Generative Adversarial Network (GAN) with the Residual Network (ResNet).The residual block of ResNet is added into the generator of GAN to enhance the ability of SAR image denoising.The combined loss function in the model can better preserve image details during noise reduction.In MATAR data setRe-GAN is compared with BM3D and wavelet denoising algorithm respectively.The experimental results show that Re-GAN has good performance in both visual effect and quantitative analysis.

In the complex electromagnetic environmentthe recognition rate of radar jamming signal is low under the condition of low Jamming Noise Ratio(JNR)and it is difficult to obtain a large number of jamming samples for training.To solve the problema radar jamming perception method based on Attention mechanismtransfer learning and residual neural network is proposed.The model takes the time-frequency domain data of interference and echo as inputautomatically extracts features and makes type judgment.The experimental results show that the model realizes the effective perceptual recognition of radar jamming under the condition of low JNR and small-sample training.The Attention mechanism and transfer learning can effectively improve the accuracy of perceptual recognition.The recognition performance is more robust and accurate than that of the traditional machine learning model and the unmodified neural network model.

In order to solve the problem of blind estimation of frequency hopping signal parameters with fixed frequency interferencea removal algorithm based on time-frequency energy value cancellation is proposedand the blind parameter estimation of frequency hopping signals is realized by least square method.Firstlythe K-means algorithm is used to select an adaptive threshold for the clustering of the energy cancellation valueand the noise and fixed-frequency interference are removed synchronously.Secondlythe maximum energy of the frequency axis at each moment of the reconstructed time-frequency matrix is extracted to obtain the envelope diagram of periodic oscillation.Finallythe least square method is used to linearly fit each extremum of the envelope diagram to realize the blind estimation of the frequency hopping signal parameters.Simulation results show that:Compared with similar algorithmsthis algorithm effectively removes fixed-frequency interferenceachieves high-precision estimation of frequency hopping parameters under the condition of lower SNRand maintains the relative accuracy of estimated parameters at a high level.

In order to achieve compund guidance functionssuch as autonomous guidancetarget positioning tracking and searchingetc.a design scheme of a small optical lens is proposedwhich can realize the integration of infrared and laser dual-spectrum and common aperture.The transmission optical system is adoptedand a wedge angle splitter is arranged in the optical pathwhich not only realizes dual-spectrum beam splittingbut also effectively corrects the system aberration caused by the 45° inclined flat plate.As a resultthe average Modulation Transfer Function (MTF) of the infrared channel is better than 0.45(30 lp/mm)and the diameter of diffusion spot in ±1.5° linear area of laser channel is 4.88~5.03 mmand the energy distribution is uniformwhich meets the requirements of the index.Finallythe high-precision optical centering adjustment and three-coordinate precise measurement are used to complete the assembly and adjustment of the lens.The average MTF of the full field of view of the infrared channel is better than 0.38the shape of the dispersion spot in the laser channel is symmetricalthe energy is evenly distributed in the linear regionand the lens effect is good.

Aiming at the flexibility requirement of resource organization in the design of distributed avionics system and improving the utilization rate of system resourcesa system management configuration method based on distributed avionics system is proposedwhich introduces functional componentsservice roles component instances and system management agentsand defines component-to-component messages and system configuration information tables.Through the system management strategythe restart of the component instancethe master-slave backup of the component instanceand the cold backup of the component instance are realized,thus increasing the task reliability of the system.The system reconfiguration of distributed avionics system is realized without changing the functions of application components and ICD interface.

A radar imaging system based on antenna scanning is proposed.This new system can image the target without radar or target motion.In the process of antenna beam angle scanningpulse compression coding modulation is applied to the radar transmitting signalwhich is synchronized with the beam angle scanningand then pulse compression is applied to the beam scanning angle dimension to improve the angle resolution of the radar.At the same timethe wideband signal is used to improve the range dimension resolutionand the two-dimensional(range dimension and angle dimension)high-resolution radar image of the target is obtained.Simulation is carried out for different signal-to-noise ratiosthe simulation results show that the new radar imaging system is effective.

Aiming at the shortcomings of ball and plate systemsuch as severe PID control oscillationtedious manual PID tuning and poor dynamic qualitya scheme combining Wavelet Neural Network (WNN) identification with WNN-PID self-tuning parameters is studied.Firstlyaccording to the strong coupling characteristics of ball and platea ball and plate system model consisting of two parts is established by Lagrange equation.SecondlyWNN-PID controller is constructed to overcome the problems of tedious manual tuning and poor stability of PID.Considering that gradient descent method and fixed learning rate are easy to fall into extreme valuethe momentum gradient and AdaDec algorithm are used to accelerate the training speed of the network.Thenthe convergence of the system is verified by Lyapunov stability theory.Finally the Matlab simulation shows that the stability and robustness of the proposed stragegy in ball and plate system are better than those of conventional PID and BP-PID strategies.

In view of the shortcomings of traditional matrix pencil algorithm in estimating the parameters of all-pole model under the condition of low signal-to-noise ratiothe calculated values are unstable and the accuracy of pole estimation is not high.FirstlyHankel matrix is introduced into the echo signal for signal decoherencethen the improved TLS-ESPRIT algorithm is used to estimate the poles of multi-band signalsand the improved TLS-ESPRIT based on Hankel matrix replaces the traditional matrix pencil to calculate the parameters of full-band all-pole modelso as to improve the accuracy of the fused signal.The final simulation and measured data show that the proposed method improves the fusion performance of multi-band signals at low SNR.

Aiming at the problem of longitudinal slippage disturbance in the trajectory tracking control of wheeled mobile robotsan iterative learning control method is studied.Firstlyin order to describe the longitudinal slippage disturbanceslippage parameters are introducedand the kinematic model of mobile robot with longitudinal slippage disturbance is established.Then an open-closed loopP-type iterative learning controller is designedseveral representative slippage disturbances are selected for analysisand the stability of the robot trajectory tracking under different longitudinal slippage disturbances is proved by theoretical derivation.Finallythe effectiveness of the designed controller is verified by a simulation example.

Aiming at the problem that the airborne control system is affected by both disturbance and noiseand the traditional disturbance observer is difficult to effectively suppress the noisea noise-reducing disturbance observer is introducedwhich can resist the influence of disturbance and noise at the same time.Firstlythe principle of the traditional disturbance observer is introducedand the reason why the traditional disturbance observer can not suppress noise is analyzed.Thenthe principle of the noise reduction disturbance observer is analyzedand it is found that the system can obtain better anti-disturbance and anti-noise capabilities by selecting appropriate filter.Finallythe controller and the disturbance observer are designedand the simulation experiments are carried out.The simulation results confirm the theoretical analysis that the system has both anti-disturbance and anti-noise capabilities.

Synthetic Aperture Radar (SAR) is a high-resolution space-borne (or airborne) radar systemand Automatic Target Recognition (ATR) in SAR image is one of the key technologies of intelligent image interpretation.Traditional SAR ship target detection algorithms are mostly limited by the scene and have poor generalization abilityso a detection model based on YOLOv3 network is designed.Multi-scale prior box is used to detect the targetand the optimal weight of the model is obtained through training to realize end-to-end target detection.The test results show thatcompared with Faster R-CNN algorithmYOLOv3 has better performance in accuracy and running speed.

A phase plate is designed to extend the depth of field of the initial optical structure.Several considerations and design requirements for the design of wavefront coding optical imaging system are proposed.According to these matters and the basic parameters selectedthe initial structure of an optical system is designed and optimized by Zemax.The initial structure is an imaging objective lens with an object distance of 4 mF number of 3 and effective focal length of 50 mm.The surface type of the selected cubic phase plate is extended polynomial.By setting a complex evaluation functiontaking the cubic coefficient of the phase plate as a variableand then optimizing the phase plate to obtain its parameters.Finallythrough the numerical comparison between the traditional optical system and the optical system with phase plateit is concluded that the depth of field of the design can be extended from 4 m to the range of 2.5 m to 8 m.