To solve the problem that energy integration cannot be realized effectively due to the range migration and Doppler migration of the echo of high-speed maneuvering targets, this paper proposes an algorithm of coherent integration via the Second-order Keystone-Biorthogonal Fourier Transform (SKT-BFT) based on velocity compensation.Firstly, a phase compensation function of velocity is constructed to correct the linear range migration caused by velocity by searching the velocity.Secondly, the second-order Keystone transform is used to correct the range bending caused by acceleration, which can also be called second-order range migration.Finally, the slow temporal-dimension signal is regarded as a Linear Frequency Modulation (LFM) signal, and the energy integration is realized by using the biorthogonal Fourier transform to estimate the chirp slope.It can be proved by simulation that this method can effectively realize the integration under the condition of low Signal-to-Noise Ratio (SNR) and the detection of a single target and multiple targets, which has better performance than other algorithms.

The time-varying formation tracking problem for second-order multi-agent systems under external disturbances is studied.Firstly, an extended state observer is designed to estimate the external disturbances of the system, and a time-varying formation tracking control protocol with a disturbance compensation term is proposed.Then, a formation vector generation model is designed by constructing a tracking differentiator, which improves the response process of the system to formation instruction control input.Finally, the formation-tracking feasibility condition is given and the stability of the system under the control protocol is proved.A comparative numerical simulation is carried out and the theoretical results are verified.

In order to improve the tracking accuracy of the photoelectric tracking system, a method for designing the coaxial tracking system of the coarse tracking link is proposed.Based on this, the control method of the fine tracking system is studied, and an Adaptive Backstepping Sliding Mode (ABSM) controller is designed.The controller is applied to the fine tracking system to form a dual composite shaft system based on coaxial tracking.The route data of three types of anti-ship missiles is generated by computer simulation, and the tracking effects of the composite axis system are simulated.The simulation results show that the system can accurately and stably track the target and meet the accuracy requirements of the tracking system of shipborne laser weapons.

In order to improve the real-time performance of intelligent surveillance cameras in single target tracking while keeping high tracking accuracy, a tracking method based on the triplets siamese neural network is proposed.Firstly, the backbone of the siamese network is reconstructed based on DenseNet to make full use of feature maps of different levels with less parameters and computation amount.Secondly, the mask branch is added to the siamese network.Then, the similarity scores between the features of target template image and those of the search image are reordered, and the mask is directly generated and refined according to the candidate response window corresponding to the maximum score.Finally, the loss function of the algorithm is defined.The proposed algorithm is evaluated on OTB50/100 and VOT2018 benchmark data set.Experimental results show that, compared with the original SiamMask algorithm, the proposed method is more lightweight while its accuracy is improved, the average frame speed is increased by two times, and the real-time performance is better.

The current blind beamforming algorithms have poor adaptability.When used in multi-objective analysis, the cascading mode has a complex structure, the parallel mode has poor robustness, and the time-frequency analysis has difficulty in estimating the steering vector.In this paper, a multi-objective blind beamforming algorithm is proposed based on the Hough-Iterative Fourier Transform（HIFT）.The algorithm firstly solves the estimation problem of the steering vector by the short-time Fourier transform and the Hough transform, then implements the waveform control by controlling the side lobe level and the transitional zone through the design of a beam pattern projection operator. Finally, the parallel output of the blind beam of multiple targets is obtained.The experimental analysis results show that:1) This algorithm is different from most of the existed beamforming algorithms, and has no need for inverse and small calculation cost; and 2) The side lobe and other parameters can be designed flexibly according to the actual situation, and the beam output performance is stable.

In the field of detection and resolution of UAV conflict, when a UAV flies towards the target point at the speed re-selected by using the reciprocal velocity obstacle method, the problem of a too large turning angle will occur.To solve this problem, an improved method is put forward.Firstly, based on the analysis of the principle of Velocity Obstacle (VO) method, the conflict resolution problem is described in detail, and a mathematical model is established by using the geometric relationship.The relative velocity between the UAV and the intruder and the angle of relative position between the UAV and the intruder can be used to judge whether there is a conflict.Then, the Reciprocal Velocity Obstacle （RVO） method is used to re-select the UAV speed and predict the UAV position of the next moment, and the B-spline curve is used to smooth the conflict resolution path.In order to shorten the path, the strategy of not restoring the original track is adopted.Instead, the UAV maintains the current speed to fly towards the target point until it reaches the end point.The simulation results show that the algorithm is feasible and effective, and the problem of detection and resolution of UAV conflict is solved.

In current electronic war, Towed Radar Active Decoy(TRAD) can capture the tracking resolution cell in the half-power beam of the radar, and the tracking gate is transfered from the target to the decoy. Under this circumstance, a new technology by which the target will be recognized correctly under the interference of decoy is in urgent need. The first step of realizing that is to separate the target from TRAD. Considering that the logistic regression model is a classification algorithm based on linear regression, so by using the logistic regression probability model, a new multi-Bayesian model based on MCMC is formed to separate the target from TRAD.

To solve the trajectory tracking control problem of the six degree-of-freedom parallel platform with uncertain modelling errors and external disturbances, a fuzzy non-singular fast terminal sliding mode control method is proposed.A non-singular fast terminal sliding surface is adopted and the universal approximation property of the fuzzy system is used to establish a fuzzy system on each degree of freedom.The adaptive law of the output of the fuzzy system is designed by Lyapunov function, and the adaptive adjustment of the switching term in the non-singular fast terminal sliding mode control is realized.The simulation results show that this method can effectively track the trajectory of the parallel platform with modelling errors, and has certain advantages in restraining output chattering, improving convergence speed and reducing tracking errors.

To overcome the shortcomings of traditional nonlinear filtering algorithms, i.e., the EKFs tracking accuracy decreases rapidly at high maneuvering and low data rate, and UKF has small linearization error but poor real-time performance at low data rate, an interactive multi-model filtering algorithm based on data rate control is proposed.This algorithm adaptively allocates the weights of EKF and UKF according to the target detection rate and the system non-linearity in different combat modes, which effectively overcomes the shortcomings of only using one filtering model in previous algorithms.The simulation results show that the proposed algorithm solves the problem that the tracking accuracy and stability of the system decrease too fast when the data rate of the EKF algorithm decreases while the target is maneuvering.Compared with that of the UKF algorithm, the running time of the proposed algorithm is greatly shortened and the radar resource consumption is reduced when the tracking accuracy of the system decreases little in the maneuvering phase.

This paper studies the problem of the stealthy engagement maneuvering strategy in continuous state space based on deep reinforcement learning.A stealthy engagement maneuvering strategy based on Double Deep Q Network (DDQN) by using the Markov decision process is established.The method of generating a target value function by DDQN solves the over-fitting problem of the traditional DQN.The training sample is obtained by the method of random sampling according to the priority, which accelerates the training of the neural network.The greedy coefficient is set according to the method of exponential decline, which solves the “exploration and utilization dilemma”of traditional reinforcement learning.An angle factor is introduced in the design of the reward function to make it more consistent with the actual combat situation.The simulation results show that DDQN has good convergence and can effectively generate the stealthy engagement maneuvering strategy.

This paper proposes a hybrid control algorithm combining a fast continuous non-singular terminal sliding mode control strategy with linear active disturbance rejection control strategy for trajectory tracking control of a quadrotor aircraft with modelling uncertainties and external disturbances.The dynamical model of the quadrotor aircraft is obtained through Newton-Euler equation, and the model is divided into two control loops of an inner loop and an outer loop.The inner loop adopts the fast continuous non-singular terminal sliding mode control to ensure the fast convergence of the attitude.The outer loop uses the linear active disturbance rejection control to reject the external disturbances and realize high-performance trajectory tracking.Meanwhile, the stability of the proposed controller is analyzed.Lastly, the effectiveness of the proposed control strategy is verified by simulation cases and flight tests.The results show that the proposed controller has fast response speed, good robustness, high control accuracy and strong disturbance-rejection ability.It can meet the requirements of quadrotor aircraft trajectory tracking control.

To overcome the difficulty in the design of such top indexes as Reliability, Maintainability, Supportability and Testability (RMST) in the development process of complex modern equipment, an integrated RMST design method based on efficiency is proposed.Firstly, the meaning of RMST index and the definition of equipment efficiency are analyzed, the main factors affecting equipment efficiency are selected, and the object of trade-off optimization in the integrated RMST design is clarified.Secondly, according to the relationship between the selected RMST index and equipment efficiency, the comprehensive expression of equipment efficiency model and the cost model of equipment development are established respectively.Finally, taking the cost of equipment development as the constraint and the maximization of equipment efficiency as the decision object, a trade-off optimization model of the integrated RMST design is established.The example analysis shows that the trade-off optimization model is effective and can provide a reference for determining RMST index in the development process of complex modern equipment.

The development of new operation targets and operation environment poses a new challenge to air-to-air missiles.The role of the reaction jet control technology on air-to-air missiles is briefly stated in consideration of the characteristics of future air-to-air missiles.In view of the overall design of air-to-air missiles, several requirements and technical difficulties of the reaction jet control technology are put forward.The study can serve as a reference for the development of future air-to-air missiles.

The gestures of the ground commander are an important part of the air command movement.In view of the current low timeliness of gesture segmentation and inaccurate short sequence segmentation, this paper proposes a real-time dynamic gesture segmentation method.Firstly, OpenPose is used to extract the key position information of the hand, calculate the angular feature and length feature sequence of the hand vertex and base point, extract the subsequence through the sliding window, and compare the mean of the subsequence with the mean of the previous subsequence.The threshold method is used to find a point that satisfies the threshold condition from the subsequence, i.e.the sequence division point.This paper constructs a database of gesture instructions.Compared with that of manual annotation, the segmentation of the proposed method is more accurate, and the accuracy is above 90% with high timeliness.

Rainy weather will greatly worsen the image quality and hinder the subsequent processing of the image.In order to realize raindrop removal in the image, a single-image raindrop removal method based on Conditional Generative Adversarial Networks (CGAN) is proposed.This method adopts the basic framework of CGAN, uses the raindrop image as additional condition information and adds Lipschitz constraints.The network model is trained by combining condition adversarial loss, content loss with perception loss to repair the raindrop area and reconstruct the image.The experimental results show that the proposed method has better raindrop removal effects than the existing algorithms, and can avoid image blurring on the basis of ensuring the raindrop removal effect.

To solve the problems in missile data acquisition and recording, such as difficulty in distinguishing the signature, unclear data source, the allograph and wrong signature, a digital verification scheme of missile data based on the sequential multiple digital signature technology is designed.Firstly, a sequential multiple digital signature method based on Schnorr digital signature algorithm is proposed, and its correctness and security are analyzed.Based on this, a digital verification scheme of missile data is designed according to the actual working conditions of the army.Finally, a digital verification system of missile data based on multiple digital signatures is designed, and a security analysis is carried out.The results show that this scheme can realize the digital verification of missile data and has high security and reliability.

Under the weather conditions such as fog, the original scene acquired by the optical camera will have serious image degradation, for example the contrast and visibility of the image are reduced and the image is blurred.At present, in the field of image dehazing, the dark-channel priori algorithm has a good effect in solving the above problems.But in view of the computational complexity and real-time performance of the algorithm, it cannot be directly transplanted into the FPGA.Therefore, based on the dark-channel priori algorithm, this paper first uses a fast estimation method for atmospheric transmittance to reduce the calculation amount and at the same time weakens the Halo phenomenon of the traditional algorithm.Then, the auto level method is used to perform contrast stretching on the dehazed image for improving the dehazing effects.Experimental results show that the proposed algorithm not only meets the requirements of real-time dehazing, but also effectively improves the ability of image dehazing.

By using the calculation method of CEP, the deviation of the landing point of the laser-guided bomb is obtained.Based on the method of sequential truncation tests, the truncated number is obtained through Monte-Carlo simulation.By analyzing different CEP indexes, the delivery precision of laser-guided bombs is evaluated.It is verified that increasing the number of delivered bombs within the set range of truncation number can improve the delivery precision of laser-guided bombs.

Fault tolerance is a basic requirement of the multi-sensor integrated navigation system.In order to improve the ability of the integrated navigation system to detect abrupt and gradual faults, a fault-tolerant algorithm combining principal component analysis with integrated navigation is designed.The theory of multi-sensor integrated navigation system is introduced in detail, and the fault detection and identification based on principal component analysis are introduced in detail.The fault tolerance performance of the integrated navigation system is analyzed.The experimental results show that the fault-tolerant algorithm which combines principal component analysis with integrated navigation has good detection ability for both the abrupt and gradual faults of the navigation sensor, and the fault-tolerant performance of the navigation system is obviously improved.

The influence of jamming on the performance of TACAN system is studied.Firstly, based on the models of effective navigation area under the blanketing jamming of radio-frequency noise and saturated asking jamming, the influence on the radius of effective navigation area of the beacon station under the action of single jammer and multiple jammers is analyzed.Secondly, the error of angle measurement under the envelope jamming of the same frequency and of different frequencies is deduced.Finally, the analysis results are verified by simulation.The simulation results show that:1) Multiple small-power jammers are more effective in affecting the navigation area of the beacon station than the single large-power jammer;2) The saturated asking jamming is better than blanketing jamming;and 3) The envelope jamming of the same frequency and of different frequencies will cause the error in accurate measurement interval, and about 40° angle measurement error will occur.

The problem of fault detection for the aeroengine's nonlinear distributed control system is studied.First, an nonlinear model in the form of polynomial combination is established for the aeroengine by using the identification method.Then, in consideration of the network factors, a model with random time delay is established for the aeroengines nonlinear distributed control system.After that, the nonlinear distributed control system is regarded as a random system, for which a nonlinear fault observer is designed.The fault observer error systems mean-square asymptotically-stable working condition in the form of Linear Matrix Inequality (LMI) is derived by applying the Lyapunov stable theory and the mean-square asymptotically-stable theory of the random system, and a feasible LMI solution is also obtained.Finally, the specific steps of fault detection are given, and the validity of the proposed fault detection algorithm is verified through simulation.