Research achievements of our project team on precision docking control for probe-and-drogue autonomous aerial refueling of UAV are reviewed. First, the bow wave model of the receiver and the control-oriented drogue motion prediction model are introduced, which are established by deep learning method combined with CFD and high-order complex model simulation data. Then, through in-depth analysis of the successful experience of manned aerial refueling, the backstepping high-order sliding mode anti-disturbance docking controller, the translation and rotation combined docking controller and the direct/preview composite guidance docking controller are creatively put forward. Simulation results show that the proposed method greatly improves the docking accuracy of the autonomous aerial refueling.

In order to improve the accuracy of tracking target trajectory of anti-ship missile with snaking maneuverability, a target tracking method combining CT turning maneuverability model with nonlinear filtering is proposed.Firstly, CT model is used to successfully simulate the trajectory of S maneuver, and then the tracking effects of EKF and UKF on S maneuver target is compared and analyzed.Although UKF is superior to EKF in precision, its operation time is significantly increased compared with EKF.In order to improve the real-time performance of UKF algorithm for target tracking, the mini-skewness sampling is adopted to reduce the number of sampling points, the distribution of sampling points is optimized by means of proportion correction and the linear operation is carried out directly by combining the actual state equation with the measurement equation.Simulation experiments show that the tracking effects of the improved UKF on the S maneuver target is similar to that of UKF, significantly better than that of EKF, and the computing time is reduced by 50% compared with that of UKF.

Aiming at the classification and recognition of radar jamming signals in complex electromagnetic environment, the time-domain and frequency-domain characteristics of noise jamming, dense false target jamming and combined jamming signals are studied. Comparison and analysis are made to different characteristic parameters, and the parameters with large difference are used to establish the identification feature vectors. Thus a Support Vector Machine (SVM) recognition structure is constructed for classification and recognition in simulation. The results show that, the feature parameters of the recognition structure are easily extracted, and the operation speed is fast, which has a high recognition probability to noise jamming, dense false target jamming and combined jamming, and has a certain application prospect.

Aiming at the scale changes, illumination changes, motion blur and background clutter in the target tracking process, a target tracking algorithm combining multi-feature fusion with scale adaption and based on Background-Aware Correlation Filters(BACF)is proposed. Firstly, various feature information of the target and the background are extracted, and weighting fusion is made to the feature distribution weights according to the response graph, to complete the target position prediction. Then, a single HOG feature in the discriminant scale space tracking algorithm is replaced by two complementary features of the Histogram of Oriented Gradient (HOG) feature and the Color Names (CN) to estimate the target scale. Finally, the adaptively updated threshold is selected to update filter model. The simulation results show that the proposed method has higher success rate and accuracy.

In order to solve the problem of parameter tuning in the design of sliding mode variable structure control law, a parameter optimization method based on reinforcement learning is proposed. Firstly, according to the system, the corresponding sliding mode control law is designed, and the range of parameter selection is given. Based on the Actor-Critic structure, the parameter online adjuster is designed. TD-error method is used to solve the problem and gradient descent method is used to calculate the updated value of neural network weight. Finally, the longitudinal channel system of fixed wing vehicle is taken as an example to simulate and Verify through experiments, which shows that the proposed control method reduces the blindness of control parameter tuning and improves the dynamic performance of the system significantly.

The traditional techniques relying on artificial experience for radar emitter feature extraction are cumbersome and have insufficient degree of distinction. To solve the problems, an improved Dual Path Network (DPN) is proposed to automatically extract features and make identification. Firstly the one-dimensional time-domain signal is transformed into two-dimensional time-frequency domain, and then directly input into the DPN for identification. Thus the identification problem of radar emitter is transformed into an image recognition problem. At the same time, considering the problem of feature loss due to too many network layers in DPN, a Triple Path Attention Module (TPAM) is proposed for re-sampling the radar emitter feature map. Then the TPAM is embedded into DPN to form a TPAM-DPN for identifying the radar emitter. Experiments on six common radar signals show that the features extracted by this method are more conducive to improving the radar emitter identification accuracy and are more time-efficient.

Aiming at the problem of long-range air defense early warning of large surface warship formation, the deployment of air defense sentry ship is studied. Based on the in-depth analysis of the advantage of air defense sentry ships in long-range early warning, the deployment models of single and two air defense sentry ships are respectively established. And the detection capability, fire strike capability, the speed of shipborne air defense missile and the penetration target are taken into account. The superiority and feasibility of the model are demonstrated through simulation calculation, and the conclusion of optimal deployment of air defense sentry ship of large surface warship formation is obtained, which can provide references for deployment of air defense sentry ship formation.

Silicon micro-gyroscope has been widely used in medium/low precision inertial stability platform and navigation system. By assembling gyroscope array, the output precision can be improved effectively. The output compensation method of gyroscope array is studied, and a Kalman filter structure based on gyroscope array is built up. A Moving Horizon Estimation (MHE) algorithm combined with U-D square root filter measuring refresh method is designed to improve the output precision of MHE. The experiment shows that the output error can be reduced by data fusion of gyroscope array. In static-state experiment, the output precision of MHE under horizon length N=1 is similar to that of U-D filter, while both of them are superior to traditional Kalman filter. Under low-dynamic test condition, benefitted by fusing the measuring output of preceding sampling points, the output precision can be further improved by MHE with horizon length N>1.

The Unmanned Aerial Vehicle (UAV) formation team collaboratively performing reconnaissance and tracking tasks is a hot research topic in recent years. As for the track planning problem of UAV formation in tracking target process, the optimal observation configuration of UAV formation collaboration in target tracking is derived, and the observation conditions are determined. Secondly, as for the target tracking problem of UAV, the construction steps of the navigation vector field are presented, and the control conditions of UAV formation for cooperative target tracking are described. On the basis of Lyapunovs navigation vector field, the collision avoidance function is established, and the collision avoidance potential field is constructed, which are superimposed with the navigation vector field to a form fusion vector field. Through experimental simulation, the fusion vector field constructed in this paper is proved to be effective in solving the obstacle avoidance problem of UAV formation in the process of collaborative target tracking.

Aiming at the spatial-temporal search problem of how to find the missing target quicklya recursive Bayesian method based search decision-making framework and an adaptive search strategy are proposed.In the decision-making frameworkthe upper limit of the searchers speedthe possibilities of false alarm detection and missed detection of the sensor are taken into considerationand the expression of the confidence function evolution under the decision framework is theoretically derivedbased on which an effective search theory is proposed.Since the impact of different strategies on decision-making can be studied and compared under this decision-making frameworkit provides an effective platform for comparison and analysis of different strategies.Finallythe effectiveness and practicability of the decision framework and proposed strategy are verified by a large number of numerical simulations.

Aiming at the operational characteristics of ground synthetic units and the shortcomings of traditional firepower distribution model, a firepower distribution model with minimum firepower waste is established. The model aims to roughly meet the damage requirements without excessively pursuing the absolute damage probability in order to avoid excessive concentration and waste of firepower. A model solving method based on improved genetic algorithm is proposed to enhance the ability of the algorithm to optimize the model. The simulation results and comparative analysis show that the model is reasonable and feasible, which can provide a better solution for the firepower optimization allocation of the ground synthesis unit.

In signal detection of the complex environment, the clutter environment where a target is located can not be pre-determined, at this time the use of targeted single CFAR detection strategy tend to have varying degrees of loss, and flexibility is poor. Therefore, a multi-strategy CFAR detection algorithm based on adaptive threshold selection is used, combining the advantages of CA, GO and ACCA detection algorithms, and the corresponding strategies are selected reasonably by judging the current clutter background. In addition, the distributed CFAR system using the composite algorithm as a local detector is studied and the fusion criterion adopts “and” and “or” rules. Finally, the performance of the algorithm is verified by simulation.

Aiming at the problem of attitude calculation drift caused by the error of the MEMS based attitude measurement system, an attitude algorithm based on variable-step-size momentum gradient descent is proposed. The algorithm improves the attitude calculation accuracy from two aspects of sensor data denoising and attitude calculating. In sensor data denoising, an improved recursive amplitude-limited mean filtering algorithm is proposed to reduce the computational complexity of microprocessors. In attitude calculating, an adaptive step size is designed based on gradient descent method, and momentum gradient is used to optimize the direction of each iteration so as to minimize the error after each iteration, and dynamic limiting filter is used to suppress angle oscillation. Results of the static and dynamic experiments show the effectiveness and superiority of the proposed algorithm.

The ship targets in the forward-looking infrared image under sea-sky scene may face problems such as low signal-to-noise ratio and large size difference. In order to better highlight the targets and suppress the background, a forward-looking infrared ship target enhancement method based on multi-scale local contrast is proposed. Firstly, the scene content in the forward-looking infrared images under sea-sky scene is analyzed. Then a new local contrast measure, called Relative Regional Contrast Measure(RRCM) is defined. By changing the size of the grid window, a Multi-scale RRCM(MRRCM) is constructed. Finally, the scale-invariant local contrast image is obtained. The experimental results show that, the proposed method can well adapt to the multi-type and multi-size ship targets in the forward-looking infrared image, which improves the signal-to-noise ratio and signal-to-background ratio of the target signal significantly. Compared with other methods, the ship target enhancement effects of the proposed method is better, especially for small and dim ship target regions. In addition, the proposed method has high operation speed, and can be further accelerated by parallel processing and integral image.

Atmospheric visibility has an important impact on aviation, navigation, land transportation, military activities and so on.The image method based visibility measurement is a new method, which has a certain degree of error in the measurement of visibility because of the errors of the imaging brightness and the dynamic range.The relative errors of visibility measurement caused by target brightness gain error, target brightness zero error, sky background brightness gain error and sky background brightness zero error are analyzed in detail.The total relative errors of visibility caused by the above four errors are obtained.The feasibility and validity of the image method based visibility measurement are verified.

In the past, the equipment simulation training system mainly focused on single-machine training, or generated and transmitted signals by physical objects, which made it difficult to implement in the practical application. In view of the above shortcomings, we put forward a design scheme of simulating training system based on ZeroMQ message oriented middleware, designed a practical and simple message passing standard, and developed an equipment simulation training system based on the communication model. The whole computer simulation system of the equipment simulation system is implemented and the interconnection between different operators is realized, which provides training conditions for improving the actual combat capability of the troops, and also provides a reference model for later simulation training system design of other kinds of equipment.

Aiming at the problem of how to deal with the evaluation information scientifically in the evaluation of military aircraft overhaul efficiency, a method of military aircraft overhaul efficiency evaluation based on intuitionistic fuzzy method is proposed, and the corresponding index system is established. Based on the similarity between intuitionistic fuzzy numbers, the judgment matrix similarity is introduced to determine the expert weight;the intuitionistic fuzzy set weight aggregation operator is used to gather the expert weighting information to get the index weight;and the linear weighting method and intuitionistic fuzzy method are used to gather the overhaul efficiency evaluation information to get the evaluation values. The overhaul efficiency of a certain aircraft is evaluated, and the rationality and effectiveness of the method are verified, which provides a method support for further improving the efficiency of aircraft overhaul.

In order to meet the anti-disturbance performance of the Self-Propelled Anti-Aircraft Gun (SPAAG) turret pitch servo system, the servo controller of the system was redesigned. The composition and working principle of the servo system were studied, and the mathematical model of each module of the system was established. The model compensation Active Disturbance Rejection Control (ADRC) for pitch servo system was proposed. By using the Recursive Least Squares (RLS) method of variable forgetting factor, and by inputting the cross-axis current and the actual rotational speed, the resistance torque and impact load of the servo system were obtained, which were used for compensation of the ADRC. The system was simulated by the S-Function module in the Matlab/Simulink simulation environment. The simulation results show that this algorithm can generally increase the ability of extended state observer for estimating disturbance, and can enhance the robustness and control precision of systems.

Aiming at the problems of large amount of sensor information, strong correlation and serious redundancy in the process of fault diagnosis, a grey Kohonen network fault diagnosis method is proposed. Firstly, the grey correlation analysis method is used to analyze the relevant parameters of the index parameters in each state of the equipment, and the grey correlation analysis matrix under each state is established. Secondly, through the introduction of grey value, the strong and weak relationship of the correlation is quantified to achieve the purpose of eliminating the similar information. Finally, fault diagnosis is carried out through the Kohonen network. The example verification results show that this method can maintain a high fault diagnosis rate while eliminating the similar information, and at the same time achieve rapid diagnosis.

In view of the fact that the armoured vehicle carrying obstacle-breaking rocket weapon may vibrate on a rough road, while a large vibration will also be generated when launching obstacle-breaking shells, which may easily cause the deviation of the direction angle of the next obstacle-breaking and affect the obstacle-breaking accuracy. With consideration of other disturbances and uncertainties, the obstacle-breaking weapon system can be taken as a nonlinear time-varying system. Based on the robustness and adaptability of fuzzy control and the adaptive and self-learning ability of neural network, a PID control algorithm based on fuzzy RBF neural network is proposed. At the same time, the structure parameter values of fuzzy neural network are initialized by K-means hierarchical clustering, and the fuzzy neural network is trained by using the LM algorithm. The simulation results show that the method can effectively improve the anti-interference ability, the accuracy of the obstacle-breaking and the speed of the gun adjustment.

In the existing methods, only the maintenance decision of a single device is taken into consideration, the maintenance strategy of the system and the joint inventory optimization of spare parts are neglected. The paper proposes a system's maintenance and inventory joint decision model based on remaining useful life prediction under imperfect maintenance. Firstly, based on the Wiener process, the degradation of the equipment affected by imperfect maintenance is modeled, and then the remaining useful life probability distribution of the equipment is derived. Secondly, the detection interval, the preventive maintenance threshold, the maintenance times threshold and the inventory are used as decision variables to establish a cost-time ratio joint decision model, which realizes the optimization of system's maintenance decision cost. Finally, numerical simulation experiments show that the maintenance cost of the proposed method is reduced by 1. 5%, compared with that of the method based on a single decision variable.

Aiming at the parameter perturbation and rudder surface faults in supermaneuver flight of thrust vector aircraft, a fault-tolerant control method based on terminal sliding-mode is proposed. Firstly, the rudder surface fault model of thrust vector aircraft is given. Then, the terminal sliding-mode control laws are designed for the attitude angle loop and angle velocity loop respectively, and the projection adaptive law is introduced to estimate and compensate for the parameter perturbation and the rudder surface faults. Finally, the rudder deflection command signal is obtained, and the Lyapunov stability analysis of the system is carried out. The Matlab/Simulink simulation results verify that the proposed fault-tolerant control law can realize the tolerance for parameter perturbation and rudder surface faults of supermaneuver flight, and the adaptive law can accurately estimate the rudder faults.