In order to adapt to complex environments and changes of the fighters own stateand improve the performance of the target tracking algorithmthe original Tracking-Learning-Detection (TLD) algorithm is improved.Firstlythe improved Scale-Invariant Feature Transform (SIFT) feature matching algorithm is combined with the optical flow method in the original TLD algorithmand one of the two algorithms is selected adaptively according to different states of the fighter and the surrounding environmentso as to improve the robustness of the algorithm and its adaptability to complex environments.Secondlyparticle filter is introduced for optimization to predict the state of the fighter in real time and narrow the detection rangewhich can reduce the complexity of calculation and improve tracking speed and continuity.Finallyconfidence detection is set up to avoid misjudgment and improve tracking accuracy.The experimental results show that the improved algorithm has better tracking performance than the reference algorithm and can accomplish the task of long-time robust tracking when the fighter is in a complex scene or when its state is changing greatly.

To deal with the problem of excessive amount of computation in the collaborative positioning and navigation of UAV swarm in complex unknown environment this paper proposes two DRL methods for collaborative information screening of UAV swarm.The multi-UAV localization problem is mapped to a Partially-Observable Markov Decision Process (POMDP) and the requirements on positioning accuracy of all the UAVs are satisfied by use of the least collaborative information based on the accuracy threshold set by Cramer-Rao Lower Bound (CRLB).The simulation results show that compared with the non-screening situation the information screening strategy based on DRL can effectively enhance the real-time performance and operation efficiency of the collaborative positioning algorithm as well as reducing collaborative information redundancy while maintaining the positioning accuracy.

When GPS signals are lostthe accuracy of the integrated navigation algorithm will decrease.To solve the probleman efficient adaptive smooth switching fusion algorithm is proposed.When the GPS signal is under normal conditionsa high-order integrated navigation algorithm based on GPS measurement is used to estimate UAV attitude.When the GPS signal is lostthe high-order filtering algorithm is terminated and switched to a low-order attitude estimation algorithm based on quaternion.When the GPS signal is restoredthe estimated value of the low-order attitude estimation algorithm is taken as the initial value of the high-order filtering algorithmso as to realize smooth switching between the two attitude estimation algorithms.Simulation results and flight tests show that the switching strategy has good estimation accuracy and feasibilityattitude angle jumping is less than 0.5° during the switching process and the convergence time is less than 0.1 s which effectively improve the accuracy of UAV attitude estimation with or without GPS signals.

In order to fully utilize the operational potential of ground-attacking UAVsa UAV operational effectiveness evaluation model is proposed based on Firefly Optimization Algorithm (FOA) and Relevance Vector Machine (RVM).Firstaccording to the characteristics of UAVs and the requirements of the battlefieldan index system for operational effectiveness evaluation of ground-attacking UAVs is established from the perspective of mathematical statistics.Thenin order to overcome the shortcomings of a single kernel functiona polynomial kernel function and a Gaussian kernel function are used to construct a hybrid kernel function for RVM.Finallybased on the built evaluation index systemrelevant parameters of the multi-kernel RVM are optimized by using FOA algorithmand an operational effectiveness evaluation model of ground-attacking UAVs is established.Simulation analysis shows thatthe proposed model is superior to the RVM model with a single Gaussian kernel or a single polynomial kernel in various evaluation indices with high evaluation accuracythus the effectiveness and feasibility of it is proved.

Accurate azimuth measurement by laser detectors is a necessary condition for the seeker to guide towards the targetbut there is few theoretical research on the measurement accuracyso this paper studies the angular measurement accuracy of laser seeker.Firstlyan atmospheric transmission model of the guidance laser is establishedand it is verified that the light spot of the guidance signal received by laser detectors is uniform.Thenbased on the basic principles of azimuth detection by laser detectorsthe relationship between azimuth detection accuracy and three factorsi.e.the radius of the light spotthe position of the spots center and Signal-to-Noise Ratio (SNR)is deduced.Finallythe influence of different factors on azimuth detection accuracy is analyzed through simulation.The results show that angular measurement accuracy of the seeker can be improved by the following four methodsi.e.reducing the radius of the light spotshortening the distance between the centroid of the light spot and the center of the photosensitive surfaceincreasing the focal length of the detectors optical systemand increasing the SNR.

Interrupted-sampling repeater jamming is a typical radar active jamming mode.In order to counter the self-defense interrupted-sampling repeater jammingthe generation mechanism of false targets is analyzed.Based on thisand inspired by time domain correlation and frequency domain correlationan interrupted-sampling repeater jamming suppression algorithm based on fractional order correlation is designed and proposed.The algorithm takes the interfered echo in a coherent processing interval as the processing objectand changes the fast time position distribution of false targets by performing different fractional correlation processing on echoes with different repetition periodsand achieves the purpose of jamming suppression through coherent integration.Simulation results show that the proposed algorithm has certain suppression efficiency for interrupted-sampling repeater jamming.After suppressionthe critical Jamming-to-Signal Ratio (JSR) for effective detection of real targets is increased by about 25 dBand the suppression rate of false targets is more than 50%.

Aiming at the key technologies of visual collision avoidance in UAV obstacle avoidance system the YOLOv4-tiny target detection algorithm is improvedand a visual collision avoidance method based on target detection is proposed.With the separable convolution structure of MobileNetthe YOLOv4-tiny target detection network is optimized to improve the detection accuracy.The efficiency of the SURF matching algorithm is improved by removing the redundant information of the undetected target area.The target area is calculated in three dimensions by least squares methodand the calculated depth and pose information of UAV are treated as the basis of visual field scale of the collision avoidance determination areathe visual collision avoidance of UAV is realized through the integration of the pixel position of target detection.Finally the improved network model is analyzed by using standard data setthe mAP value is 69%and the UAV visual communication module is built for visual collision avoidance testwhich verifies the rationality of the method.

Military target recognition is a major research direction in the field of target detectionwhich is of great significance for early detection of enemy situation and accurate strike of the target.At presentthe mainstream military target recognition algorithms mainly use such unsupervised deep learning network models as YOLO and Fast RCNN.In these methodsthe manually-labeled datasets are inputted to the network and then processed to obtain the output.Howeverthe datasets have limited quantityunsatisfying accuracy and generalization ability.To solve these problemsthis paper proposes a military target recognition algorithm based on the unsupervised networkwhich effectively solves the problems of insufficient datasets and unsatisfying accuracy through Generative Adversarial Network (GAN).

Traditional safety surface generation algorithms have such problems as uncontrollable operation scalelong running timeand complex surfaceswhich can not meet the requirements of rapidity and reliability of airborne systems.In response to the above problemsthe safety surface generation algorithm is optimized by compressing the Digital Elevation Map(DEM)improving the terrain slope constraintand analyzing the longitudinal curvature constraint.Thentrajectory planning on the surface is conducted.Simulation results show that the improved safety surface generation algorithm reduces unnecessary calculations with efficiency increased by about 3~6 times in comparison with traditional algorithmsand the trajectory planned on the surface meets the requirements of helicopters in low-altitude penetration scenarios.

When the targets scale changes or when the target is occludedthe Kernel Correlation Filtering (KCF) tracking algorithm cannot ensure long-time tracking. To solve the problema scale-adaptive and anti-occlusion long-time tracking algorithm is proposed.FirstlyHOG features and color (CN) features are fused and a scale filter is added to estimate the scale of the target.Secondlythe Average Peak Correlation Energy (APCE) index is introduced to determine whether the target is occluded. When the target is occluded SVM classifier is used to recheck the position of the target after occlusion.Finallyaccording to APCE and maximum correlation response value of the position filterthe corresponding model updating strategy is selected.Two data setsOTB100 and UAV123are selected for experiment. The experimental results show that the improved algorithm can effectively deal with target scale change and target occlusionand realize long-time and stable tracking.

At presentthe testability models considering uncertain factors have the problems of high modeling complexitycomplicated structuredifficult hierarchical modelingweak learning abilityand poor expression of uncertain information.Oriented to electronic equipmenta hierarchical testability modeling method based on Bayesian network is proposed.The original Bayesian network model and the multi-signal flow graph model are taken as research objectswhose parametersstructural characteristicsadvantages and shortcomings are analyzed.According to their complementary relationship and testability modeling requirementsthe Bayesian network model is improved.The model replaces the fault and test relationship with the fault and signal relationshipand characterizes fault transmission by the connection between signal nodes.The models structure is determined by using the hierarchical structure of circuitsand the models parameters are established by using Bayesian networkwhose inference ability is used to calculate testability indexes.Amplification filter circuits are taken as an exampleand a model is built and used to calculate the testability indexesand evidence processing tests are conducted.Compared with the multi-signal modelsthe model can accurately calculate testability indexes with strong evidence processing abilitywhich verifies the effectiveness and practicality of the modeling method.

Global Positioning System (GPS) is widely used in navigationmeasurementemergency communication and other fields.In recent yearsthe scholars are making increasingly intensive researches on acquisition and tracking process of GPS baseband signals.Thereforedifferent acquisition and tracking algorithms are systematically summarized for comparison.According to the processing mode after the mixing of digital Intermediate Frequency (IF) signals with local signalsthe acquisition algorithms can be mainly classified as coherent acquisitionincoherent acquisition and differential coherent acquisitionand a comparison is conducted.According to the function of tracking loopsthe tracking algorithms can be classified as carrier tracking and code loop trackingand a comparison is conducted.The improvements on those acquisition and tracking algorithms in recent years are summarizedand the advantages and disadvantages of the improved algorithms are analyzed.Finallythe prospect of baseband signal acquisition and tracking algorithms is given.

The status and technological features of advanced airborne Electro-Optical Targeting Systems (EOTS) abroadsuch as Star SAFIRE 380X targeting systemEOTS targeting system and LITENING targeting systemare introduced.On which basisthe technologies related to advanced EOTS are analyzed and their development trend is given.

Firstlya pseudo-linear state-dependent model for the attitude and vertical channels of an unmanned helicopter is establishedwhich keeps the coupling state term.Thenthe optimal solution of the nominal model is found by online solving State-Dependent Riccati Equation (SDRE)and an integral-extended state-dependent controller is designed.As for the external disturbancesa finite-time disturbance observer is designed to online observe and then compensate for the disturbances in the coupling vertical and yaw channels.Simulations are implemented on the numerical platform of Yamaha-Rmax unmanned helicopterand a comparison with the classical robust H∞ controller is conducted.The simulation results show that the proposed method has the advantages of less dependence on parameters and faster responsewhich can rapidly stabilize inner-loop attitude and vertical velocity of the helicopter.

According to the antenna sharing requirements of mobile combat platformsa method for allocating the shared aperture of linear array oriented to radar-communication integration is proposed. Firstlya model of the shared aperture is establishedand a non-uniform linear shared array is designed according to the difference in operating frequencies between radar and communication.Secondlythe position allocation of array elements is transformed into a multi-objective optimization problemand the highest sidelobe level of radar array pattern and the channel capacity of communication array element are selected as optimization objectives.Finallythe improved simulated annealing algorithm is applied to optimization calculationand relevant simulation experiments are completed. The simulation results show that the highest peak sidelobe level of radar and the capacity of communication channel can both meet the working requirementsand antenna aperture sharing is realized.

A new laser ranging technology based on gas absorption spectroscopy is proposed.The frequencies of reference optical signals and measurement optical signals carrying gas absorption spectroscopy information are mixedand Fourier transform is used to extract optical path differenceso as to obtain the distance information.According to the deduced ranging principlesa ranging system based on laser methane remote sensor is designedand its ranging process is simulated and analyzed by Matlab software.The simulation results have verified the feasibility of the gas laser ranging technologywhich supplies a promising direction for developing gas laser radar.

MAV/UAV cooperative operations are becoming a research hot spot in various countries.In response to current demand for effectiveness evaluation of cooperative operationsthis paper discusses a method for evaluating the effectiveness of MAV/UAV cooperative operations based on Hopfield network.By analyzing the typical models of MAV/UAV cooperative operationsand integrating the content of “Observation-Orientation-Decision-Action (OODA)” mission loopthe index system for effectiveness evaluation of MAV/UAV cooperative operations is established.The evaluation data obtained by Extensional Analytical Hierarchy Process (EAHP) and the fuzzy evaluation method is taken as the reference standard of Discrete Hopfield Neural Network (DHNN).Based on the associative memory function of DHNNan effectiveness evaluation model for MAV/UAV cooperative operations is designed.Example analysis and simulation verification are conducted.The simulation results show that this method can realize an objective and effective evaluation of the effectiveness of MAV/UAV cooperative operations.

Time Triggered (TT) network can ensure the reliability and real-time performance of data transmission in safety-critical systems.The basic principles of monitoring and fault-tolerant mechanisms of TT network are analyzed.The monitoring mechanisms based on time windowsemantics and data flow are designed respectivelyand the synchronous fault-tolerant mechanism based on redundant resources and the high-integrity fault-tolerant mechanism based on COM/MON architecture are designed.A verification environment for these monitoring and fault-tolerant mechanisms is builtand relevant verification tests are conducted.The results show that the analysis and design of monitoring and fault-tolerant mechanisms can enhance the safety level of TT network.

Infrared and visible image fusion can effectively compensate for the weakness of a single sensorand produce a fused image with better visual effect and higher definition.When designing the rules of fusing detail layers and base layers of two imagesthe fusion methods based on multi-scale decomposition often only consider single-scale information in detail layers and base layerswhich may lead to a lack of effective features in the fused image.To solve this problemthis paper proposes a novel fusion rule based on the random walk algorithm to fuse base layers and detail layers.This fusion rule estimates two saliency maps and obtains one of them with multi-scale information for the fusion of base layers and detail layerswhich can effectively fuse the multi-scale information in each layer into the output image.Thereforethe fused image is more beneficial for human-eye perception.

In order to improve the initiation control accuracy of guided missiles and obtain a more accurate initiation delay timean integrated information fusion method based on measurement data of infrared seeker and laser rangefinder is proposed.In multi-mode information processingdue to different starting-up time and sampling frequencies of the two sensorsand the time difference between observation datathe measurement data of the two sensors is not on the same time base.Thereforea time alignment method based on interpolation and extrapolation is presented in a typical projectile/target encountering environmentand the measurement data of the sensors is applied to the calculation of a delay time model.On the basis of the modela centralized data filtering algorithm based on the improved particle filter is proposed.A comparison with the traditional extended Kalman filter algorithm is conducted.The simulation results show that the accuracy of the detection angle is improved by 80.45% and the measurement accuracy of the optimal firing angle is improved by 78.61%thus the initiation control accuracy of the integrated fuze is improved.

The technology of RT cluster management is describeda function of Bus Controller (BC) oriented to the addressing of 128 terminals is realizedand the innovation of the technology is introduced.Based on the RT cluster management technologya time-triggered dynamic network scheduling mechanism is proposed which ensures the certainty of communication delay and time offsetand realizes bus data transmission with high reliability and certainty.By setting up the system test environmenttwo key technologiesRT cluster management and dynamic network scheduling mechanismare tested and verified.The system operates stablyand its function and performance are consistent with theoretical analysis.

Based on the concept of electromagnetic maneuver warfareand from the aspect of the maneuver confront between our airborne active phased array radar and the opposite sides electronic reconnaissance systemthis paper analyzes the meanings of radio frequency stealth and electromagnetic maneuver of airborne radarand discusses the agility and entropy of electromagnetic maneuver used in radio frequency stealth of airborne radar.Based on thisin the electromagnetic environment of air combatthree kinds of electromagnetic maneuver strategies in radio frequency stealth of airborne radarnamelyminimum timeminimum power and maximum signal uncertaintyare summarizedand the corresponding operation modes of electromagnetic maneuver are presented.The range of study on electromagnetic maneuver warfare of airborne radar is expanded by applying it to the study of radio frequency stealth of airborne radarand the proposed strategies and operation modes have certain application value.

In the field of intelligent transportationthe real-time three-dimensional(3D) object detection in road scenes is of great significance to ensure the safety of vehicle driving.The use of radar point cloud and image data fusion is able to achieve the effect of complementary advantages.Howeverin order to achieve higher detection accuracythe 3D object detection algorithms using radar and image data fusion usually adopt the two-stage networkbut the computing speed is slower than that of the one-stage networkand the speed of detection system is very important in practical application.Aiming at the above problemsa real-time detection method of 3D object is designed based on the improvement of one-stage network RetinaNet.The 3D anchors are mapped to the feature maps of the point cloud and image respectivelyand ROI pooling is used to convert the regions cropped by the anchor on the feature map into the same size and fuse them.Finallythe regression parameters and categories of the target boundary box are outputand the target prediction boundary box is obtained by adjusting the anchor box.Experiments on KITTI dataset show that the proposed network is superior to the comparison algorithms in the accuracy and time consumption of pedestrian and vehicle detection.