To overcome the shortcomings of the traditional method of on-call search for the unknown-heading target by using single-layer bounding array, an active search method using multi-layer bounding array is put forward, which is weakly dependent on the judgment of the enemy submarine's possible sailing speed. The parameter determination model is also constructed. The proposed method can shorten the process of discovering targets, complete search tasks in a limited time, and improve the search efficiency.

In the background of the joint operations of ground synthetic units, a threat assessment indicator system based on cluster targets was established. Aiming at the problem of the imbalance of state in the traditional threat assessment method caused by invariant weight, the interval number theory was introduced to determine the index interval invariant weight, and the variable weight theory was used in combination with the combat situation to determine the index interval variable weight, so that the weight can be dynamically adjusted with the change of the battlefield. By combining interval variable weight with grey correlation method, the least-squares grey-correlation threat assessment model for interval number was constructed. Finally, the effectiveness of the model and the method was verified by the example simulation.

Because the Active Disturbance Rejection Control (ADRC) technology adopts too many controller parameters in the attitude control of quadrotor aircraft, it is difficult to obtain a set of optimal solutions. To solve the problem, a PSO algorithm with variable-weight hybridization is proposed. The algorithm is mainly composed of two parts. First, according to the distance between the particles in the particle group and the global optimal particle in the iterative process, the inertia weight is dynamically adjusted, and the coefficient is set to control the degree of influence on the inertia weight. Second, the hybridization is introduced. Within a specified number of iterations, if the global optimal value of the particle group does not change continuously, the specified number of particles are hybridized to increase the particle diversity and avoid falling into local optimum. The quadrotor model is built and simulated by Matlab/Simulink. The results show that the optimization algorithm is effective for ADRC controller parameter setting, which can ensure the control quality of the quadrotor and improve the design efficiency.

To solve the problem of individual identification of distance deception jamming emitters, a new method for individual identification of jamming sources based on envelope front difference characteristics of the individual emitters is proposed. Firstly, the signal is denoised by using a combination of wavelet denoising and sliding window processing. Compared with the denoising processing only using one method, it can obtain better denoising effects. Then, the idea of cross-correlation algorithm is used to realize the position alignment between the received signal and the template signal. Finally, the mean of the difference amplitude of the envelope rising edge is introduced as the characteristic factor, and the classification of individual emitters is realized by using K-means clustering algorithm. The simulation result shows that the proposed method performs better on identification than the included angle cosine algorithm proposed in reference [10].

Unmanned helicopter is undertaking more and more important combat missions, and deck motion is an important factor affecting the safety of its autonomous landing, thus the accurate estimation and compensation of deck motion becomes an urgent problem. To solve the problem, a method of deck motion estimation and compensation based on adaptive AR model and optimal preview control is proposed. When designing the deck motion estimator based on AR model, the time-variant factor is introduced to design the updating law of the adaptive model parameters, so as to optimize model performance. Simulation results show that the proposed method improves the accuracy of deck motion estimation to a certain extent, and its computation process is relatively simple. Then, the prediction signals generated by the deck motion estimator are introduced into the preview controller as the foreseeable future information, and the deck motion of the landing point is compensated based on the optimal control theory. This method solves the problem of phase delay of deck motion compensation system effectively and improves the deck motion tracking accuracy to a certain extent, so that it can increase the success rate of unmanned helicopter autonomous landing.

The salient texture structure of actual images can provide priori information for Block Compressed Sensing (BCS) algorithm and optimize the algorithm. Based on this, a new Adaptive Block Compressed Sensing (ABCS) algorithm based on saliency is proposed. The saliency in the proposed algorithm is built on the theory of gray-level spatial-dependence matrix and Weber's theorem. The deterministic Orthogonal Symmetric Toeplitz Matrix (OSTM) is adopted to measure the target image. The adaptive block strategy minimizing the average entropy, the block-vector generation method maximizing the angle second-order moment and the adaptive sampling rate setting under the synthetic feature are proposed. The analysis and verification are carried out by using different basic reconstruction algorithms. Experiment results show that, compared with the traditional algorithm, the proposed algorithm performs better on different indexes, is easy to implement by hardware and has universality and stability for different reconstruction algorithms and test images.

In order to realize fast and stable control of the linear variable parameter system, a robust model predictive control algorithm based on polyhedral invariant sets with maximum control laws is proposed. When offline, by setting a group of progressive values approaching the stable point, according to the principle that the distance from the state variable to the stable point is roughly the same, a sequence of state variable sets is constructed, and the larger control law in each set is obtained through optimization. The laws are then optimized in a backstepping way to obtain a series of maximum control laws,and are combined with the input and output constraints of the system to obtain the offline polyhedral invariant sets. When online, the actual control law of the system is obtained by linear interpolation optimization according to the position of state variables of each sampling period in the polyhedron invariant sets. The detailed steps of the offline and online algorithms and the closed-loop stability proof of the system are given. The simulation results prove the effectiveness of the proposed algorithm, and indicate that the proposed algorithm makes the closed-loop response of the system more rapid and stable.

In order to solve the problem that the accuracy of image recognition is not high in the training process of convolutional neural network by using traditional quadratic cost functiona convolutional neural network algorithm based on cross-entropy cost function is proposed.By mathematical derivationit is proved that the cross-entropy cost function is more accurate than the quadratic cost function in image recognition.Based on MNIST dataset and CIFAR-10 datasetand using AlexNet convolutional neural networkthe quadratic cost function and the cross-entropy cost function are adopted to train the image recognition model respectively.When the recognition accuracy and loss value of the digital image are stablethe cost function is tested several times by using the test dataand comparison is made to the recognition accuracy of the two functions.The simulation results show that the proposed method can not only improve the accuracy of digital image recognitionbut also has a faster model training speed than the traditional cost function.The process of training deep neural network model is obviously shortened.

In order to solve the problem that the accuracy of image recognition is not high in the training process of convolutional neural network by using traditional quadratic cost function, a convolutional neural network algorithm based on cross-entropy cost function is proposed. By mathematical derivation, it is proved that the cross-entropy cost function is more accurate than the quadratic cost function in image recognition. Based on MNIST dataset and CIFAR-10 dataset, and using AlexNet convolutional neural network, the quadratic cost function and the cross-entropy cost function are adopted to train the image recognition model respectively. When the recognition accuracy and loss value of the digital image are stable, the cost function is tested several times by using the test data, and comparison is made to the recognition accuracy of the two functions. The simulation results show that the proposed method can not only improve the accuracy of digital image recognition, but also has a faster model training speed than the traditional cost function. The process of training deep neural network model is obviously shortened.

Aiming at the control speed and accuracy of the AC servo system of the ship-borne rocket launcher, a composite control strategy combining RBF neural network based PID control with fuzzy control is proposed. The composite control strategy adjusts the weights of the two control methods on the controlled object according to the error information, so as to switch the two control methods without disturbance, and realize the continuous and stable output of the final control quantity. The simulation results show that the control strategy combines the advantage of RBF neural network based PID control with that of fuzzy control, and has high control precision, fast system response speed and small oscillation, which can better meet the performance requirements of the AC servo system of the ship-borne rocket launcher.

Multi-polarized antenna systems are capable of reducing device size while maintaining low inter-antenna correlation, and thus have great potential for wireless communication systems of small UAVs. A polarization Multi-Input Multi-Output (MIMO) channel model is proposed based on the geometric theory of double cylinders. The geometric method is used to simulate channel depolarization, where the MIMO signal strength of the scatterer is determined by a simplified ray tracing method. The modeling method provides a mechanism for analyzing and calculating the MIMO polarized Channel Impulse Response (CIR), which can further deduce the time-frequency characteristics of cross-polarization discrimination (XPD) and MIMO channels. In addition, numerical simulation analysis of appropriate modeling parameters is performed.

The principle of the phenomenon that the infrared imaging system may defocus with the change of temperature is analyzed, and a linear infrared athermalization detection system is proposed. The characteristics of linear infrared imaging are analyzed. To solve the problem that the traditional methods of SMD, TenenGrad, Laplacian are prone to be disturbed by the stripe noise of the linear infrared system and considering the shape of the target plate, a method of definition evaluation based on the improved Sobel operator in seven directions is proposed. The experimental results show that this method can effectively overcome the interference of the stripe random noise of the linear infrared imaging system on the definition evaluation factor. Comparing with that of the traditional algorithms, the response curve of the proposed algorithm has obvious variation tendency and small fluctuation. It can better meet the monotonic and unbiased requirements of autofocusing.

Aiming at the problem that the tracking accuracy decreases due to the change of target scale in the process of target visual tracking, a kernel correlation filtering tracking method based on GM (1, 1) scale estimation is proposed. The GM (1, 1) grey prediction model is introduced to predict the target scale based on the kernel correlation filtering tracking algorithm, and the scale estimation process is designed to estimate the current target scale, which effectively solves the problem of tracking performance declining caused by obvious scale changes. Finally, the OTB-100 video sequence dataset is used to test the performance of the algorithm. The test results show that the algorithm still has good tracking performance when the target scale changes significantly. It can realize stable and accurate target tracking, and effectively improve the tracking performance of the kernel correlation filter tracking algorithm.

Radar wave measurement is the only way to obtain high-accuracy wave information that can be carried out on ship, reflecting the real-time and actual wave conditions of the ship's current navigation area. The nonlinear evolution of waves can be well described by the model based on the mNLS equation. The radar wave measurement system is used together with mNLS equation. The radar wave measurement spectrum is selected with the application conditions of ε≤0.3 and Δf/f0≤0.2 to construct the initial conditions for simulating the nonlinear evolution of waves.

The principle of an electric servo system with double pendulum nozzle of a certain type of a missile is analyzed, and the nonlinear characteristics and cross-coupling interference of two vertically-mounted electromechanical actuators are studied through the motion analysis and modeling of the system. The decoupling operation of the expansion, contraction and displacement of the actuator under random swing angle instructions of the controller is given. The nonlinear relationship between the swing arm and the nozzle swing angle is studied. Based on this, a mathematical model of a single servo mechanism with nonlinear links is established, and the controller parameters are adjusted. The time-domain based step response method and the frequency-domain based sinusoidal correlation analysis method are used, and a simulation is carried out on the Simulink platform. The result is compared with the real experimental data, which verifies the accuracy of the model and lays a solid foundation for further research on the performance test of this type of servo mechanism.

IMA1G has been widely applied in military and civil avionic systems, and the era of IMA2G will come soon. IMA2G has overwhelming superiority over IMA1G in such aspects as function, power consumption, volume, weight and cost. The academia and industry at home and abroad attach great importance to IMA2G. Considerable progress has been made in relative research areas, but there are still many challenging issues. This paper gives a detailed illustration of the progress of IMA2G in key technology fields in foreign countries, and then introduces our research scheme, achievement, test and validation on the IMA2G. Finally, further research directions are given.

The calculating method of the operational effectiveness of laser angle deception jamming is obtained by analyzing the whole jamming process of the laser angle deception jamming system. Then, by analyzing the factors affecting the success rate of laser angle deception jamming, the corresponding probability is obtained and the operational effectiveness of laser angle deception jamming is finally calculated. The analysis and calculation of operational effectiveness can provide a decision guidance for laser angle deception jamming commanders to better implement laser angle deception jamming, so as to enhance the operational effectiveness of the laser angle deception jamming system.

The Micro Electronic Mechanical System (MEMS) inertial components have some shortcomings such as low accuracy and big noise.In order to improve the fusion accuracy of MEMS gyro array, a gyro array fusion algorithm is put forward based on the conventional algorithm of support degree fusion, which removes abnormal support points and uses limited memory information of time to carry out weighting for the fused data. The algorithm uses the weighted idea of spatio-temporal fusion of gyro array to optimize the traditional weight assignment of the support degree. Thus the stability and reliability of gyroscope array output are guaranteed, and the fusion result is closer to the true value of the output. The experimental result shows that the proposed fusion algorithm is superior to support degree fusion algorithm and the average-value estimation algorithm.

High-resolution is the trend of infrared imaging technology. In order to implement high-resolution infrared imaging in real-time, a technical scheme based on SOPC technology is proposed. Hardware and software co-design architecture is constructed with the core of CPU Micro Blaze to complete such image processing procedures as non-uniformity correction, image denoising, and histogram equalization enhancement, so as to solve the problem of high-speed processing of large-data images and meet the real-time requirement.Compared with traditional hardware architecture of FPGA+DSP, this technical scheme has lower power consumption and is miniaturized with high reliability and high generality. The designed high-resolution infrared image processing system has the advantages of high definition, high contrast and clear image details.

Airborne electronic hardware is assuming more and more system functions. Considering the cost, using COTS IP becomes a trend, which also brings challenges to airworthiness certification. The usage position and risk of COTS IP in airborne electronic hardware design are analyzed, and the current airworthiness certification requirements are studied. A suggested evaluation process is proposed based on the authority documents and certification experience, and the airworthiness certification elements are summarized. The research results are helpful to solve the airworthiness certification issues of COTS IP, and provide a reference for the existing airborne electronic hardware design assurance methods.

The track interruption problem that appears in the space-target tracking process will seriously disturb the track fusion and increase the loads of the sensors.Therefore, the study on the spatial target intermittent track correlation algorithm is of great significance.In the traditional intermittent track correlation algorithm, the track obtained by linear extrapolation prediction has big error, which may lead to inaccurate association. Therefore, a spatial target intermittent track correlation algorithm based on trajectory prediction is proposed in the paper. The algorithm tracks the space target based on the dynamic equation, and makes curve-fitting to the filter update value before the interruption to obtain the predicted initial value point. The trajectory prediction is then combined with the target dynamic model. The predicted data is correlated and matched with the state update values of the first few moments after the interruption, to achieve the correlation fusion of the track before and after the interruption. Finally, the simulation results verify the effectiveness of the proposed algorithm.

In order to overcome the difficulty of controller design and parameter tuning for complex high-order controlled object, a design method for fractional-order robust controller based on model order reduction is proposed. Firstly, the complex high-order model is approximated to a reduced-order fractional-order model with time-delay links. According to the Nyquist curve characteristics of the original model and by using the sequential quadratic programming method, the parameters of the reduced-order approximated model are obtained. Based on this, the structure of a fractional-order controller is designed, and the new numerical calculation methods of the robust index of maximum sensitivity and the controller parameters to be tuned are deduced. Finally, the controller parameters are tuned based on the performance index in composite time domain. Simulation results show that the proposed parameter solving method for model order reduction does not need global optimization, has a fast convergence speed, and the reduced-order model approximates the original system very well. The designed robust controller enables the original system to have a good control quality.