Aiming at air-ground collaboration of Unmanned Aerial Vehicle (UAV) with Unattended Ground Sensor (UGS) for target tracking, this paper proposes a searching and tracking method for moving targets based on local search trees in the traffic road network environment.In this method, the UAV is in close communication with an UGS to obtain the time information of the target passing through the node.Based on the information, the target motion speed is estimated and the subsequent target position is predicted.The path of UAV in target tracking is optimized through local recursive search.For the decision-making of sensor node accessing order with incomplete information in the tracking process, two kinds of node selection evaluation mechanisms are designed and analysis is made to their effects.Simulation results show that this method can capture targets with a higher probability when the targets moving path and speed change greatly.

Considering that the velocimetry accuracy will decline when the optical flow information is affected by foreground moving objects and noise in UAV visual navigation, we proposed an optical flow velocimetry optimization method based on the Mean-Shift algorithm.The method uses the pyramid Lucas-Kanade algorithm to calculate the sparse optical flow and the improved Mean-Shift algorithm for the clustering of the optical flow data, which eliminates the interference items in the data according to the consistency of speed.Experimental results show that the proposed method can reduce the effects of foreground moving objects and noise on optical flow velocimetry in UAV visual navigation, and improve the precision of UAV velocity measurement.

For the Scale-Invariant Feature Transform (SIFT) algorithm, the eigenvectors at the key points (the extreme points of a stable scale space) are computationally complex and usually have high dimensions, while the values of the polynomial deterministic matrix are limited.To solve the problems, an SIFT image registration algorithm based on the block polynomial deterministic matrix is proposed.The SVM-derived sparse representation method is used to reduce the high-dimensional descriptor vector extracted by SIFT to a low-dimensional sparse feature vector, which reduces the dimensions of key-point description vector.The Euclidean distance is used for similarity measurement of the key-point feature description vector.The contrastive analysis with the traditional algorithms shows that the improved algorithm can effectively improve the registration accuracy and enhance the real-time performance.

In order to overcome the shortcomings of the current methods in the nonlinear analog circuit fault diagnosis based on frequency response, and select a better test stimulus signal, we used Volterra series for modeling, and proposed a method for intelligent optimization of test stimulus by using the frequency basis search together with the genetic algorithm.Firstly, the comprehensive searching technique was adopted to search the best frequency point sets, which make the output frequency point sets of each kernel not coincident with each other.Then, the frequency basis was filtered by taking the sum of the Euclidean distances between the characteristic vectors of the fault status as the objective function.Thereby the optimal test excitation signal was obtained.A nonlinear circuit was used for verification.

Sea-land segmentation is of great significance for improving the accuracy of SAR target ship detection.Aiming at the fact that the traditional algorithm can not carry out high-quality sea-landsegmentation for the SAR image, this paper proposes a segmentation algorithm based on the improved SLIC superpixel partitioning and the Hierarchical Area Combination Criterion (HSWO).According to the statistical characteristics of SAR images, the SLIC and HSWO algorithm models are improved respectively at first.Then, the SLIC algorithm is used to segment the superpixels in the image, and the superpixel blocks are clustered according to the merging rules of the hierarchical regions.Finally, the segmentation between sea and land is implemented.Experiments show that the improved model proposed here has higher processing precision and efficiency.Compared with other algorithms, it is more suitable for the sea-land segmentation of SAR images and has certain engineering application value.

To improve the edge-preserving ability of image filtering, a guided image filtering algorithm based on robust adaptive weighting is proposed.Firstly, Gaussian filter is performed on the guided image and first-order differential method is used to judge the position of the edge, which can remove the noise as well as improve the robustness of the edge information extraction.Then, Otsu is used to segment the edge regions from the non-edge regions, the self-adaptability of threshold selection is improved.Finally, the improved piecewise function model is used to fit the ideal weight factor and control the degree of smoothness of different regions, which can realize the robust adaptive guided filtering and achieve the smoothness of the edge.The experiments of edge-preserving smoothing and matting are carried out.Compared with guided image filtering algorithm and the other two improved algorithms, the proposed algorithm has better robustness and edge-preserving smoothness.

Target recognition of conventional Synthetic Aperture Radar (SAR) usually adopts the batch learning method.However, in practical application, the training data of the system cant be obtained all at one time.When new training sample arrives, the whole system needs to be retrained when using the method of batch learning.In order to solve this problem, ROSELM, an incremental learning algorithm, is applied to SAR target recognition, and Particle Swarm Optimization(PSO) algorithm is used to optimize the initial weight of ROSELM to improve its stability and recognition rate.The experimental results show that:1) When new SAR target samples are obtained, the system updating can be implemented simply by updating the output weights without re-training;2) The algorithm is very fast and has a high recognition rate, which is a good choice for online updating of SAR target recognition system.

In SAR scene matching guidance system based on local invariant feature, the real-time performance of the matching of the local invariant feature points is of great engineering significance.A new scene matching algorithm based on two-tier matching strategy is proposed to improve the speed of scene matching algorithm.First, the first-level and the second-level feature points are extracted respectively from the real-time and reference images according to the threshold value.Then, make each of the second-level feature point subordinate to the closest first-level feature point according to their distance.Further, the first-level feature points are described by SIFT descriptor and using nearest neighbor method based on kd-tree to achieve the initial matching of the first-level feature points.RANSAC algorithm is utilized to get rid of the error matches.Finally, the second-level feature points which are subordinate to the first-level feature points after eliminating the false matching points are described by BRIEF descriptor and then are matched.Experimental results show that the matching speed of feature points is improved.

The current hyperspectral and multi-spectral image fusion algorithms have such defects as having large solution space, not considering the physical meaning of hyperspectral data, and being prone to local optimal solutions.To solve these problems, a hyperspectral and multi-spectral image fusion algorithm is proposed based on Minimum Volume Constraint and Coupled Non-negative Matrix Factorization (MVC-CNMF).In the process of separating the mixed pixels, the algorithm takes the physical meaning of the image into consideration and adds the minimum volume constraint of the endmember single body.Simulation results show that the proposed algorithm can effectively overcome the defects in the existing fusion algorithms, accurately match the endmember with the abundance of hyperspectral and multi-spectral images, and obtain high-spatial-resolution fused images.This algorithm is especially suitable for the hyperspectral images with a large number of endmembers.

In ship target detection of infrared images, the target is usually near the sea-sky line or coastline, thus the sea-sky line and coastline should be detected in advance to determine the potential area of the ship target, reduce the search range of target detection, simplify the data processing, and improve the detection speed.Aiming at the poor adaptability of the traditional sea-sky-line/coastline detection algorithm to different background images, we made an analysis to the features of sea-sky line and coastline, and proposed an algorithm for sea-sky-line/coastline detection by use of LSD line segment detection algorithm and clustering.Firstly, the local straight line contour of the image was obtained by the LSD line segment detection algorithm, and then the potential sea-sky-line/coastline area was obtained by K-means clustering.Finally, the real sea-sky-line/coastline position is determined by analyzing the texture features of the potential sea-sky line and coastline areas.The experimental results show that this method is highly adaptive to the detection of sea-sky line and coastline under various backgrounds and has high detection precision.

Fast steering mirror powered by piezoelectric ceramic has fine dynamic performance, and can meet the demand for high precision positioning in engineering, but its hysteresis effect restricts the further performance development. The mathematical model of hysteresis character based on PLAY operator has the advantages of being simple in structure, convenient for calculation, and high precision of model.However, the realization depends on parameter estimation precision, and it's difficult to obtain the physical quantity with a high precision. An algorithm is proposed to work out the parameters of PI inverse model by using geometric method. Experimental results show that the algorithm has good dynamic performance and high precision of modeling, and the algorithm based feed-forward control of PI inverse model can compensate well for the hysteresis effect of the fast steering mirror powered by piezoelectric ceramic.

In order to effectively improve the robustness against the direction error of arrival and the polarization parameter error, this paper addresses a spatial-polarization domain joint-robust adaptive beamforming algorithm.First of all, the Interference-Plus-Noise Covariance Matrix (IPNCM) is reconstructed on the direction-polarization angle interval of each interference signal.Then, the steering vector of the desired signal is estimated on its direction-polarization angle interval.The joint-robust beamforming weighting is designed based on the reconstructed IPNCM and the estimated steering vector for airspace-polarization domain.The simulation results show that the proposed algorithm is robust against the steering vector mismatch caused by the direction error of arrival and the polarization parameter error.

In order to obtain the law of the influence of motion factors of cone carrier on the Doppler frequency of radar fuze echo in its reentry flight, the motion model of the carrier is constructed accurately through the establishment of multi-coordinate system and by use of the conversion between the coordinate systems.Analysis is made to the influence of such motion factors as flight velocity, reentry dip angle, swing and cone on the Doppler frequency components of the radar fuze echoes in the time-frequency domain.In order to verify the correctness of the law of influence, the echo signal of the radar fuze is simulated by using Matlab software with the Short-Time Fourier Transform (STFT) as the central idea.The simulation results are consistent with the theoretical values.It shows that the motion model is reasonable and the law of influence is correct.

When hypersonic vehicle changes the flight state, the calculation process will be very complicated if the attitude control system is designed by using the traditional state feedback method, since the linearization calculation is repeated at different balance points to maintain the stability of the system.Besides, the actuator saturation should be taken into consideration due to the input constraint in high-speed flight.Aiming at the above problems, a two-step method was developed.Firstly, a model reference controller based on Linear Parameter Varying (LPV) system was designed.The system matrix and feedback gain of the controller were calculated by using singular value decomposition and linear matrix inequation, respectively.So the tracking response of the command signals could be realized after the flight state changes.Then, considering the case of actuator saturation, an anti-windup compensator was introduced, and LQR optimal control theory was used to calculate the compensator gain so as to limit the control input.Simulation result shows that:The designed controller has smaller overshoot and shorter adjustment time, and the calculation amount of the controller design is reduced effectively.

A new robust model-reference adaptive nonlinear inversion controller is presented for a statically unstable UAV to perform fast climbing and diving maneuver, which mainly deals with the aerodynamic and control couplings and uncertain disturbances during the maneuver, as well as the unstable pitching moment caused by static instability.Firstly, the nonlinear inversion control based on state-feedback is adopted to realize the decoupling between multiple control channels.Then, the robust adaptive controller is designed based on the decoupled closed-loop system to compensate for the nonlinear inversion error and the unstable pitching moment, and attenuate the uncertain disturbances, thus to guarantee the stability of the UAV in the process of fast maneuvering.The effectiveness and robustness of the robust model-reference adaptive nonlinear inversion controller are validated through nonlinear simulations of the fast climbing and diving maneuver.Compared with the typical robust servo LQR optimal controller, the designed controller has better decoupling capacity and uncertain disturbance attenuating performance.

A portable test system was designed in order to investigate the relationship between the Single-Event-Upset (SEU) affairs of SRAM-based FPGAs and the dose of high-energy particle in low-altitude environment.By using the test system, SEU tests of SRAM-based FPGAs were conducted at six different altitude sites in a certain area.The average elevation of the area is from 3000 m to 5000 m, which is appropriate for simulation of low-altitude flight environment.In the tests, the system gained lots of field data, which was then analyzed by using Matlab.According to the test result, the scientificity and practicability of this test system were verified and analyzed from such aspects as the storage structure of the SRAM-based FPGA, the SEU generating mechanism, and the working principle of the test system.The analysis result shows that the portable test system is effective, which can serve as a reference for the selection and application of SRAM-based FPGAs in the aerospace field.

The Head-up Vision System (HVS) integrates the Enhanced Vision System (EVS) with the Synthetic Vision System (SVS), and provides real-time location information for the pilot to improve flight safety during the aircraft's near-land operation. A new image fusion algorithm used in HVS is proposed in this paper. First, Laplacian pyramid transform is carried out for each source image.Then, the method of selecting the larger absolute value is used on the high-frequency part, and the method of local energy combined with local average gradient is used for the fusion of the low-frequency part. Finally, the fused image is obtained by inverse Laplacian pyramid transform.The experimental results show that this algorithm has good fusion effect in the fusion of EVS image with SVS image.

Determining the near-surface electromagnetic environment of HEMP is the basis for studying the HEMP effect of near-surface electronic devices.Since the ground has a reflection effect on the electromagnetic field, the electromagnetic environment near the ground is quite different from the electromagnetic pulse environment in free space.By analyzing the propagation characteristics of near-surface HEMP and the difference of soil electrical parameters, an inhomogeneous media influence model was established.The modified Fresnel reflection coefficient was used to calculate the model, the accuracy and simplicity of CST numerical simulation method were validated, and a simulation was made to the near-surface complex electromagnetic environment.The results show that:1) The waveform, ground height, upper layer thickness and incident angle have obvious influence on HEMP;and 2) Inhomogeneous medium makes the near-surface HEMP electromagnetic pulse environment different from that of the homogeneous medium.The conclusion supplies a basic basis for guiding the near-surface electromagnetic pulse protection.

By analyzing the typical operational processes of ship-borne anti-missile system in intercepting anti-ship missile and the damaging mechanism of the warhead of the ship-to-air missile,a quantitative analysis is made to the motion characteristics of the warhead preposed fragments,and the surface density model of the fragment is established. Pk/h function is used as the damage describing criterion to establish the damage probability model of the ship-to-air missile for one interception of anti-ship missile;and Markov chain is used for establishing the damage probability model of multiple interceptions.The feasibility and reliability of the model have been verified by the simulation analysis.Some suggestions are given for optimizing the intercepting strategy,which can provide a reference for the commanders to make operational plans.

Considering the external disturbances and parametric uncertainties existed in the attitude control of quadrotor Unmanned Aerial Vehicles(UAVs), we proposed an attitude control approach based on RISE.In the approach, a model-based feed-forward control term is used to achieve accurate nonlinear compensation, and a robust integral feedback term to compensate the unknown modeling uncertainty.The proposed controller can achieve stable and high-precision attitude control for quadrotors in presence of parametric uncertainties, external disturbances and noises.Lyapunov analysis reveals that the proposed method can guarantee the tracking errors to converge to the origin with asymptotic tracking performance.The effectiveness and robustness of the proposed method are validated through extensive simulations and comparisons.

The integrated INS/optical flow navigation method has found wide applications in many occasions.The accuracy of optical flow has direct influence on the performance of navigation parameters. When the optical flow has great error due to low illumination or other reasons, the accuracy of navigation parameters will decline greatly and thus cannot continue to navigate.To overcome this problem, a method for velocity forecasting based on Elman neural network is proposed.When the situation is suitable, the neural network model is trained online;and under special condition when the reliability of optical flow is very small, the neural network model that has already been trained is used for forecasting the velocity of the platform.In addition, the KF based on the INS dynamic error model can obtain the error vector by fusing the data of the INS with the optical flow, which is then used for correcting the navigation parameters. The trial based on the small quad-rotor air vehicle shows that, the forecasting of velocity from the neural network can closely approximate the true value in a short time, which demonstrates the correctness and effectiveness of the method. By comparing with the real value, the average error of attitude, velocity and position are 0.1%, 1%, and 2.4% respectively.

The opto-electronic load is an important reconnaissance equipment of the UAV system.To obtain a clear image, a new opto-electronic load control strategy is proposed for inhibiting the disturbance of visual axis.First, the structure and working principle of the two-axis and two-frame opto-electronic load are introduced.Then, analysis and classification are made to the disturbance factors affecting the stability of the visual axis.The dual rate-loop control strategy and an active disturbance rejection controller are applied on the opto-electronic load.The experimental results show that, the opto-electronic load using the proposed line-of-sight stabilization strategy can satisfy the technical requirements of disturbance rejection.

Aiming at the design features of the Line Replaceable Module (LRM) in the Integrated Modular Avionics, a method is proposed for the design of a Nand-Flash data storage module with high security, high reliability, and large capacity with ARINC653 compliant partitioning operating system.It has a three-layer structure including flash driver layer, flash management layer and file system layer.The method of emergency key data destruction is used to ensure the high security of the system, and the reliability of the system is guaranteed by adopting bad block management, wear leveling and trusted restoring file system.Read/write test, emergency key data destruction test and reliability test were made to the data storage module at different levels and with different file sizes, and the test results show that the module meets the design requirements and can be applied to the field of avionics requiring high reliability and high security.