To solve the problem of the uncertainty of the searching width of the magnetic anomaly detector caused by the uncertain diving depth of the enemy submarine in practical combat, the concept of average searching width is put forward and the calculation model of the average searching width of the magnetic anomaly detector is built based on the performance characteristics of the magnetic anomaly detectorswhich can provide the basic parameters for combat plan and preparation. Two types, totally three kinds of searching methods are proposed for the anti-submarine patrol aircraft to carry out regional search with the magnetic anomaly detector. Two types, totally four kinds of action plans for use after the contact are given. The model is proposed for calculating the probability of successfully detecting the enemy submarines in specified areas by the anti-submarine patrol aircraft with the magnetic anomaly detector, and this model can be applied to inverse calculation.

The task assignment of the manned/unmanned aerial vehicle formation in ground Time Critical Target (TCT) attacking was studied.Based on the general framework of the manned/unmanned aerial vehicle formation system and the concept of time-sensitive characteristic function, the task assignment model of the manned/unmanned aerial vehicle formation for cooperative ground TCT attacking was built. By adopting the improved Genetic Algorithm (GA), the matrix coding was introduced to enhance the task assignment efficiency. The timing sequence problem of task assignment was solved by introducing the decoding method for the deadlock problem.A simulation of the proposed algorithm was carried outand the result proved that this algorithm has high task assignment efficiency.

The covariance matrix algorithms have strong robustness under the condition of large mismatch errors of MIMO radar.Howeverthose algorithms usually have high computational complexity.In view of the contradiction between high performance and low computational complexity, this paper proposes a method to modify the mismatch error model. The model is modified from two aspects. Firstly, based on bilateral constraining, the error set of the unilateral steering vector is set up to constraint the errors of the transmitter and the receiverand the mismatch error set of the joint steering vector was obtained. Secondly, based on the structural characteristics of the uncertainty set model, K-means clustering algorithm is used to obtain a small number of weighted characteristic sampling points to replace the previous large number of discrete sampling points, so as to reduce the number of sampling points needed for covariance matrix construction and thus reduce computational complexity. Simulation results show that the proposed algorithm has strong robustness under the condition of large mismatch errors and presents outstanding performance in improving the output SINR.

To solve such problems as target scale change, deformation and occlusion in multi-target tracking, this paper proposes a new algorithm with good robustness.This algorithm adopts the improved SVM classifier for online learning, and the tracking problem is considered to be a structural learning problem with the maximum interval and the updating mode is improved to find the optimal weight. At the same time, the mechanism for the online updating of the classifier is adjusted, and the accumulated error is reduced to a certain extent, in which way online learning gives play to its advantages more adequately. In addition, the algorithm adopts the correlation filter to adaptively adjust the size of the tracking box, and proposes the mechanisms of occlusion processing and data association, so that the numbers of the targets that reappear after the occlusion are not exchanged.The experimental results prove that the proposed algorithm improves the tracking accuracy and can fulfill multi-target tracking tasks in complex background environments.

To solve the problems of high autocorrelation sidelobe peakhigh cross-correlation peak, and low convergence rate of the algorithm in the orthogonal polyphase code design for Multiple Input Multiple Output (MIMO) radar, a method based on the Genetic Algorithm (GA) and the Harmony Search (HS) algorithm is proposed for the design of the orthogonal polyphase code of MIMO radar. Each time when the GA has been optimized to a certain extent after iterating several times, the HS algorithm is used to search possible solutions in its neighborhood. This hybrid algorithm possesses the global convergence ability of GA and the neighborhood search ability of the HS algorithm.By adding the HS algorithm, the convergence rates of the autocorrelation sidelobe peak and the cross-correlation peak are faster, and the mainlobe-to-sidelobe ratio is higher.Simulation results show that the proposed method is feasible and effective.

As a small surface intelligent task platformthe Unmanned Surface Vessels (USV) have such advantages as good mobilitystrong concealmentand a high degree of automationwhich have become a hot spot of research at home and abroad.The USV manipulating model is the basis for the realization of such control operations as autonomous navigation and intelligent obstacle avoidanceand its identifying accuracy directly affects the quality of the final control effect.In order to overcome the effect of “parameter cancellation” occurring in the identifying processand to improve its identifying accuracyan algorithm for model parameter identification based on the step-by-step processing of experimental data is presented.The working procedures of the algorithm are given in detailand the model parameters are obtained by a manipulating test operated on the lake.Finallyan experiment is implemented on the identified USV model and a real USV respectivelyand the correctness of the identified model is verified through a comparison between the experimental data and the simulation results.

As a small surface intelligent task platform, the Unmanned Surface Vessels (USV) have such advantages as good mobility, strong concealment, and a high degree of automation, which have become a hot spot of research at home and abroad. The USV manipulating model is the basis for the realization of such control operations as autonomous navigation and intelligent obstacle avoidance, and its identifying accuracy directly affects the quality of the final control effect.In order to overcome the effect of “parameter cancellation” occurring in the identifying process, and to improve its identifying accuracyan algorithm for model parameter identification based on the step-by-step processing of experimental data is presented.The working procedures of the algorithm are given in detail, and the model parameters are obtained by a manipulating test operated on the lake. Finally, an experiment is implemented on the identified USV model and a real USV respectively, and the correctness of the identified model is verified through a comparison between the experimental data and the simulation results.

In order to realize the attitude control of fixed-wing UAVs under the conditions of uncertainties and external disturbances, a control law combining the extended disturbance observer of the nonlinear sliding mode with the fast-response dynamic inversion was designed. On the basis of the design of the extended disturbance observer, an extended disturbance observer of the nonlinear sliding mode was designed by introducing the sliding mode theories, so as to make real-time estimation of the disturbances and their rate of change. The UAV attitude equation was divided into the slow loop of the posture angle and the fast loop of the attitude angular velocity. Then the dynamic inverse control law was designed respectively and the unknown disturbance was compensated based on the disturbance estimation.At the same time, the differential value of the command value estimated by TD (fast Tracking Differentiator) was added to the fast loop controller to improve its response speed. Finally, the stability of the composite controller was proved.The simulation results show that the designed composite controller can control the attitude movement of UAVs efficiently.

To solve the problem that the specular highlight covers the surface features during the imaging process and thus hinders many visual applications, this paper proposes a novel specular highlight suppression algorithm based on chromaticity analysis and L1-weighted norm regularization constraints.This method doesn't rely on such preprocessing procedures as clustering and segmentation. By using the distribution of image intensity, the Median Specular-Free（MSF）mechanism is put forward to effectively separate highlight components and suppress noise interference. Meanwhile, by combining the chromaticity analysis with local chromaticity correlation, the L1-weighted norm regularization is carried out. Then, the variables splitting strategy is introduced to realize fast optimization, and the inherent details of the targets were effectively restored, such as their texture, color and edges.Experimental results demonstrate that our method can realize accurate suppression and separation of highlight components, and preserve the original characteristics of the targets to a great extent.This algorithm is highly reliable with increased computational efficiency.

In order to solve the problem in carrier UAVs of the coupling of attitude and trajectory in the longitudinal channel at the end of carrier landing, a decoupling control strategy was proposed by using the elevator channel to control the attitude and using the engine channel to control the trajectory. The control method made up of dynamic inversion and adaptive control was applied to weaken the influence of unknown disturbances and guarantee the rapidity of the control system. In order to reduce the landing error caused by deck motion, the deck motion compensation network was designed.The simulation results show that the control scheme has a satisfactory control effect, and the controller has strong anti-interference ability, which can effectively reduce the influence of the ship stern flow.

To address the issue of the route planning of reconnaissance UAVs for multiple mission targets in the mission area, a route planning method is proposed based on improved Imperialism Competitive Algorithm (ICA) for multi-target reconnaissance.The problem of the reconnaissance order of multiple targets in the mission area is transformed into the Traveling Salesman Problem (TSP) model, and the optimal trajectory of each order is calculated quickly by using the improved Intelligent Single Particle Optimizer (ISPO) algorithm. The value related to the track length and the complexity of the track is the power of the state, and ICA is used to solve the problem.The simulation results show that the proposed method can work out the approximately optimal reconnaissance order and the trajectory satisfying the flight performance constraints.

There are problems in traditional compressive tracking algorithms of few extracted features, high sensitivity to illumination changes and inadaptability to the changes in target scale. To solve these problems, an adaptive-scale compressive tracking algorithm fusing HOG-like features is proposed.On the basis of Haar-like features, the new algorithm uses the method of fixed ratio to fuse HOG-like features, so as to reduce its sensitivity to illumination. The integral graph algorithm is used to accelerate the calculation of HOG-like features.In addition, a scale estimation method based on correlation filter is proposed, so as to find out the scale to which the scale filter has the biggest response and then take it as the new target scale.The size of the tracking window is timely adjusted and the feature extraction template is updated.Therefore, the problem of scale change is solved. The experimental results show that the improved algorithm can adapt to marked illumination changes, scale changes and in-plane rotation.Its tracking accuracy and robustness are improved greatly, which satisfies the requirements of real-time tracking.

It is widely agreed that a complete Synthetic Aperture Radar (SAR) target feature library should be built to improve the basic support capability of SAR image applications in China.At presentthe accuracy of the SAR image in the SAR target feature library built by the electromagnetic modeling simulation depends on the simulation parameters of the ground object.The simulation parameters can hardly be obtained by theory.To solve this problema method based on Convolutional Neural Network (CNN) is proposed to predict the best simulation parameters of SAR images.An 11-layer CNN regression system is builtwhose input is the SAR simulation image.Since the predicted simulation parameters are 4 dimensional a new loss function is proposed to improve the predicted accuracy of each dimension in the process of multidimensional regression.Through an analysis of the changes in the error amplitude of the parameters during the training of the neural networkit can be seen that the loss function can achieve the desired result in the prediction of all the 4 dimensions.A comparison between the real image and the simulated image shows a high similarity between themwhich validates the effectiveness of this method.

This paper addresses the issue of firing compatibility of the naval guns with the air-defense missiles in air defense and missile defense of warship formation. Taking the interception of the anti-ship missile by the naval gun during its full air route as the object of study and based on the quantitative description of airspace resource by using the airspace grid model，this paper adopts the ballistic differential equation to solve the fire control problem and builds a rapid calculation model of fire control units based on the airspace grid. Under the condition of the full-course interception of the anti-ship missile by the naval gun，the occupied airspace resource and time resource by naval gun projectile are calculated.The matching of the occupied spatial and temporal resources is carried out，and a new method based on airspace grid is given，which is helpful to solve the problem of firing compatibility between the naval guns at the launch end with the missiles.

To address the issue of autonomous time synchronization of Tactical Data Link (TDL) under the condition of limited resources, a time-difference data processing algorithm based on Kalman filter and an adaptive RTT message transmission strategy were designed, which effectively reduced the resource consumption rate of RTT messages.In addition, a corresponding solution was proposed for the problem of Kalman filter divergence caused by NTR role transfer, which ensured the stability of the time synchronization function.The experiment and simulation results have proved the effectiveness of the proposed method.

An adaptive neural network control method based on sliding mode control is proposed for the reaction-jet/aerodynamic compound control system of the missile. Firstly, to improve the stability of the missile, the aerodynamic part is designed by using sliding mode control, and the adaptive neural network control algorithm is used to eliminate the error in system modeling, so that the stability of missile overloading can be improved.Secondly, the variable structure method is adopted to design the reaction-jet part to improve the response speed of the missile.Meanwhilethe fuzzy control method is used to reduce the buffeting problem caused by the variable structure control, so that the stability of the missile can be further improved.The experimental results show that the proposed control method can effectively improve the response speed of the missile, and improve the stability and overload capacity of the missile. Therefore, the effectiveness of this method is verified.

To solve the problem that the traditional PRI sorting algorithm has difficulty in effectively identifying the key radar signals with complex PRI changes, a method of fast target recognition of key radar signals based on the characteristics of crystal oscillation is proposed. This method avoids the sorting procedures of the conventional PRI sorting algorithms, and uses the correlation between the PRIs to sort the pulse signal. This method can identify multiple targets simultaneously. According to the principle of PRI generation, whether the PRI change is complex or not has little effect on the recognition result.Through simulations and comparison, it is proved that the proposed algorithm can rapidly identify the pulse sequence of the key radar signal, and get a better recognition result of the radar signal with complex PRI changes.

In order to solve the problems of low locating precision and low recognition rate of the airport identification algorithm in sub-meter high-resolution remote sensing images, a new identification algorithm based on Deep Convolutional Neural Network (DCNN) is proposed.Firstly, the bi-cubic interpolation algorithm is used to down-sample the original single-phase remote sensing images and convert them into grayscale images, and the pre-processed images are obtained by fuzzy enhancement.Secondly, the edge information of the images is detected by using Canny edge detection operator, and the straight line segments are extracted by using probability Hough transform.The linear regions are preliminarily screened and merged by judging whether there are parallel lines. Then, DCNN is used for judging the merged regions to acquire the recognition probability of the corresponding regions. Finally, the airport area is obtained by analyzing the probability values of the candidate regions.Simulation experiments were made to the two kinds of remote sensing images with high resolution, the recognition rate was 100% and the mean locating accuracy was 87.53%, which proved the validity and versatility of the proposed algorithm.

The solid-state laser generated by using the LD direct end-surface pumping technology can be used in complex airborne environments, and it has high research value in engineering application. The static laser output with high light conversion efficiency can be obtained through the successful study of the design method of the high-efficiency LD direct end-surface pumping technology, which is of great significance to the subsequent realization of the high-energy LD end-surface pumping laser illuminant. This paper focuses on a kind of 20 Bar LD array with vertical and even distribution, and two kinds of optical coupling systems of quadrangular prism are simulated and designed by using ZEMAX software. The pump light of the end-surface pumping array is focused into the working substance by using the coupling system.The beam focusing positions of the two coupling systems are different.By changing the focusing position, the output status of laser energy in different states is studied. The experiment goes as follows:in view of this flat-flat cavity configuration and with the size of the working material (1at%) being 4 mm×4 mm×10 mm, the pump light is designed to evenly illuminate the front surface of the working material and the distance from the focus to the front surface of the working material is designed to be 5.5 mm.The result shows that at this timethe mode matching degree between the pump light and the oscillation light is the highest, in which state the maximum light conversion efficiency of the LD direct end-surface pumping solid-state laser can be as high as 40%.

The traditional Direction of Arrival (DOA) estimation algorithms require a large amount of sampling data, which causes high computational complexity. To address this problem, based on the compressive sensing theory, a beamspace based Regularized Multi-vector Focal Undetermined System Solver (RMFOCUSS) DOA estimation algorithm is proposed, which uses the spatial sparsity characteristic of targets of interest.The proposed algorithm maps the received compressed signals from the element-space to the beamspace, which overcomes the flaw that the sparse reconstruction algorithm cannot be used under the conditions of low SNR to some extent. Numerical simulations demonstrate that:1) Compared with the traditional CaponMUSIC and l1-SVD algorithms, the proposed algorithm can effectively carry out DOA estimation of the coherent signals with higher angle resolution and estimation accuracy;and 2) Compared with the element-space based RMFOCUSS and l1-SVD algorithms, the proposed method has a lower computational complexity.

To monitor the working conditions of UAV's actuator system in time, a novel fault-diagnosis approach was proposed based on Ensemble Empirical Mode Decomposition (EEMD) and the multi-class Support Vector Machine (SVM) with posterior probability. The actuator signals obtained under five typical working conditions were taken as the object of study, i.e., normal condition, loose condition, damaged condition, stuck condition and reverse condition.First, EEMD was made to the collected signals, which were decomposed into a series of Intrinsic Mode Functions (IMFs) with simple components. Then, the energy values of the components of each order were calculated, by which the signal feature vector was obtained. Finally, the multi-class SVM based on posterior probability was established according to the feature information, and thus the type of the aircraft actuator system fault was identified.Simulation results show that the proposed approach can be applied to the fault diagnosis of the actuator system.

The long-wave infrared detector is the key sensor of the airborne Infrared Search and Tracking (IRST) system, and its imaging performance directly determines the key parameters of IRST, including the target detection, and finally affects the infrared detecting distance of IRST. A long-wave standard imaging lens is designed to evaluate the imaging performance of the long-wave infrared detector with a focal length of 150 mm and a total length of 210 mm. The distortion rate of it is less than 2%, and the Modulation Transfer Function (MTF) is up to the diffraction limits. The imaging lens can accurately and quantitatively evaluate the imaging quality of the infrared detector during the screening and checking process.

The layout optimization of aircraft equipment cabins in limited space with various requirements of layout constraints is a complete NP problem to which is difficult to find an optimal solution. This problem is abstracted as a three-dimensional layout problem. This paper presents a new method for layout designing based on multi-constraint handling to solve the layout design problem.Firstlythe layout constraint priority is determined according to the layout space and the importance of constrained objects, and then the constrained objects are laid out with certain constraint vectors in order of priority. Secondly, unconstrained objects are grouped into categories and preprocessed according to the optimization goal and searching methods.The real layout is simulated to achieve the optimum in every step, so that the unconstrained object layout problem can be solved. Finally, the method of human-computer combination is used for tuning the layout, in order to achieve the orderliness and purposefulness of the layout design.