Aiming at the problem that the grey wolf algorithm is easy to fall into the local optimum,which leads to the low precision of path optimization,an improved grey wolf algorithm is proposed and applied to the static 2D UAV path planning research.Firstly,a nonlinear adjustment strategy of control parameters is introduced to balance the global search and local search of the algorithm.At the same time,the grey wolf algorithm is combined with Powell algorithm to improve the optimization effect and path planning accuracy.Finally,the test and simulation show the optimization accuracy and path planning effect of the improved grey wolf algorithm.

Aiming at the problem that the error modeling method of inertial devices affects the error compensation accuracy which affects the navigation system accuracy,an error modeling method of inertial devices based on statistical characteristics restoration of device noise is proposed.Firstly,the high-frequency noise error is restored by power spectral density modeling,and then the low-frequency noise error is restored by the performance calculation method corresponding to its statistical characteristics.It not only avoids the problem that the power spectral density modeling method is weak in characterizing the low-frequency noise of devices,but also avoids the large amount of preliminary workload and calculation in the machine learning method.Compared with the existing error modeling methods,the experiment show that the error between the device mathematical model output and the real device output is reduced by an order of magnitude,and the device error compensation accuracy is higher in pedestrian navigation application.The navigation position error is reduced by 35.261% and the heading angle error is reduced by 31.198% after compensation,which effectively improves the accuracy of pedestrian navigation system.

In order to solve the problems that the simplified autopilot is often used to replace the inner loop in the formation keeping control of quadrotor formation which can not fit the real model,and the decline of control accuracy due to disturbance and difficult controller parameter ajustment,a sliding mode formation controller based on RBF and BP neural network is proposed.Firstly,the Leader-Follower concept is used to model the formation problem of quadrotor,and the inner-loop trajectory tracking controller is designed based on RBF neural network approaching uncertainties and interference terms.On this basis,the formation keeping controller is designed according to the communication relationship between UAVs,and the parameters of the controller are adjusted by BP neural network.Finally,the stability is proved by Lyapunov method.The simulation results show that the proposed controller has good anti-interference ability,and can effectively improve the accuracy of formation keeping and keep the a stable and expected formation.

Carrier aircraft landing is a system engineering with strong nonlinearity,strong coupling,and complex environmental interference.The automatic landing task under the fault condition of the carrier aircraft is extremely risky and challenging.A sliding mode fault-tolerant control method based on Extended State Observer (ESO) compensation is proposed for the deterioration of control system performance caused by random actuator drift failures and partial damage failures during landing.The method adopts ESO to accurately estimate the unmodeled dynamics of the system,the aft wake disturbance and the fault state in real time,and then a sliding mode control law is designed to ensure error convergence,and the differential signals of all orders required by the control law are obtained by using a high-order Linear Tracking Differentiator (LTD).The simulation results show that the method designed in this paper has good tracking performance,anti-interference ability and fault tolerance.

Aiming at the problems of strong randomness of sampling point expansion and unsmooth path in the path planning process of RRT (Rapid-exploration Random Tree) algorithm,a constrained random sampling based RRT algorithm-CRS-RRT (Constrained Random Sampling-based RRT) is proposed.In this method,the potential energy function of gravitational field in artificial potential field method is introduced to constrain random sampling points to sample near the target point,so as to guide the random tree to grow towards the target point and improve the planning speed of the algorithm.Combined with the strategy of removing redundant nodes and Minimum Snap track smoothing method,a safe and smooth track that meets the dynamic constraints of UAV can be quickly generated in a complex three-dimensional environment. The simulation results show that the algorithm can effectively improve the speed of trajectory planning and shorten the trajectory length.

Aiming at the cooperative take-off and landing scenarios of carrier-based UAV,the mission profile of carrier-based UAV formation during take-off and return is given.The flight route design requirements of each flight stage are put forward,and the flight routes of carrier-based UAVs in take-off waiting stage,approach waiting stage,landing stage and wave-off waiting stage are designed respectively,and the whole flight route of autonomous cooperative take-off and landing of carrier-based UAV formation and its corresponding flight key points are obtained.Finally,the designed flight route is simulated and verified by Matlab,and the simulation results verify that the designed flight route meets the corresponding design requirements.

Aiming at the problems that traditional methods are inconsistent in simulation time and easy to fall into local optimization in solving UAV path planning problems in complex environment,an improved cellular ant colony algorithm is proposed on the basis of grid map.Firstly,in order to unify the simulation time step,the hexagonal grid map is used to model the flight space;Then,an improved cellular ant colony algorithm is proposed for path planning.The algorithm introduces the concept of potential field to modify the heuristic function,adopts the differential search strategy to guide the ant colony to search for the target quickly,and designs an adaptive pheromone update method to select the optimal route.The experimental results show that the model and algorithm proposed in this paper solve the problem of non-uniform simulation time in rectangular grid map,effectively improve the speed of path optimization and global search ability,and avoid the algorithm falling into local optimization.

Aiming at the problem that feature matching algorithms,such as moving grid statistics (GMS),have concentrated mismatching,and the matching fails due to too few matching points when the viewpoint changes greatly,a feature matching algorithm using improved GMS algorithm combined with RANSAC to calculate homography matrix is proposed.Firstly,the original GMS algorithm grid is assigned weights,and the variable threshold is improved to obtain enough matching points.The matching point set obtained by the improved GMS algorithm uses RANSAC to fit the homography matrix,and the Brute Force Matching(BFM) set is finely selected to generate the final matching point set.Experiments on public feature matching datasets show that the matching precision and recall are improved by 22.69% and 18.58% respectively,and the algorithm is more suitable for scenes with large viewpoint changing.

The existing rain removal methods cannot completely remove rain streak,and the image details are lost after rain removal.Therefore,an image rain removal method based on multi-resolution fusion dense network is proposed.The main body of the network is composed of multiple multi-resolution parallel fusion modules,which always keeps the spatial accurate high-resolution representation and receives a lot of context information from the low resolution.A multi-scale feature fusion module based on the selective convolution kernel mechanism SKNet is used to effectively aggregate features from streams with different resolutions by nonlinear method.An improved residual module is used in different resolution streams,and the convolution of multiple scales of adjacent layers is used to obtain rich rain ripple information.Dense connections are used between modules to enhance feature propagation between different modules.Experiments show that the evaluation index of the proposed method on synthetic and real rain image datasets is improved compared with other rain removal methods,and more detailed information can be retained while removing rain patterns.

Aiming at the problems of small object scale and low resolution in remote sensing image object detection,a remote sensing image detection method based on residual adversarial object detection algorithm is proposed in this paper.The feature information of the image is reconstructed through residual adversarialism,so as to realize the improvement of image resolution.Based on image feature information which is extracted by Backbone network,the features are fused by Neck structure.Finally,the CIoU_Loss function is designed to increase the regression accuracy,and improve model performance.Experimental results show that,compared with other algorithms,the mean precision,mean recall,mean F1-score and mean mAP value of this algorithm are improved by 8.15%,6.9%,7.15% and 6.75% respectively.The algorithm has high accuracy in object detection in low-resolution remote sensing images,and has good effect on small object detection in remote sensing images.

Aiming at the problem of low detection accuracy due to unclear targets of ships at sea,this paper proposes a fusion method of ship infrared and infrared polarization images based on a deep learning framework to enhance the weak targets of ships at sea and improve the detection accuracy.Firstly,the source image is divided into ship contour part and feature part,and the contour part is fused by weighted average strategy,and the ship polarization image is denoised by using non-local mean.The feature part is extracted using VGG network,and then the fused image is reconstructed.Compared with the traditional image fusion methods,the proposed method can retain more infrared and polarization features of the ship,so that the fused image information is enhanced and has better contrast and signal-to-noise ratio,which provides a new method for ship target detection in complex sea surface background.

Aiming at the command and control decision-making problem in complex battle scenes,a command and control decision-making method for multi-task process is proposed based on the decision tree theory.Firstly,the judgment process of OODA loop theory is analyzed in detail,and the judgment model is generated by constructing decision tree.The decision tree invokes the multi-task process service in the decision process and generates judgment results.The commander makes command decisions according to the judgment results,so as to realize the intelligent decision-making of the machine under the control of human.Through the verification and analysis of this method,it shows that this method has the characteristics of high-level interpretability,accurately expressing the commander′s operational intention and flexibly constructing operational strategies.

The influence of temperature change on the zero offset of optical fiber gyro is one of the key factors that restricting its performance.The BP neural network can improve the accuracy of temperature compensation to a certain extent,but the BP neural network has local minimum problem.In this paper,Ant Colony Optimization (ACO)BP neural network algorithm is used to compensate the drift of fiber optic gyro,and the initial parameters of BP neural network are optimized.The experimental results show that using ACO-BP neural network to compensate can improve the zero offset stability of fiber optic gyro by about 80% in the temperature range of -40 ℃~60 ℃,and the compensation effect is better than that of previous BP neural network.

This paper proposes a Segment Random and Angle Random (SRAR) algorithm for UAV′s three-dimensional route planning under complex space and multiple constraints.SRAR algorithm uses random segments and random angles to determine the way points.Meanwhile,the angle of directional line segment is specified,and the circular obstacle avoidance method,route smoothing method and multi-route planning method are designed to improve the quality of route planning,and meet the diversity requirements of route planning.This paper also proposes an unassociated iterative updating method to solve the variable-length solution,which can obtain the variable-length solution under multiple constraints.By verifying the route quality in complex space,and comparing with ant colony algorithm,genetic algorithm and particle swarm optimization algorithm,the results can effectively prove that the SRAR algorithm is robust,efficient and universal for UAV′s route planning.

Accurate path following is necessary for UAV mission planning.With respect to the problems of AOGL path following algorithm,which relies on prior knowledge or trial selection,lacks theoretical guidance and may not be optimal,a path following algorithm based on PSO-AOGL is proposed.By multiplying the empirical values of each element in the weight matrix with the coefficient to be optimized,the modified guidance law is derived by substituting the Riccati equation,and the objective function is constructed to find the optimal solution.Under different disturbance wind speed conditions,numerical simulations are carried out for linear and circular expected tracking paths respectively.The results show that PSO-AOGL has better tracking effect and can effectively reduce the tracking position error.

Aiming at the high precision tracking problem of photoelectric tracking turntable,a backstepping sliding mode control system and algorithm are designed.Sliding mode control is introduced in the algorithm based on backstepping control,which improves the robustness of the controller.Through the simulation analysis,it can be seen that the performance of the algorithm is greatly improved compared with that of the backstepping control algorithm and the PID control algorithm.The step tracking response speed is increased by more than 21%,and the steady-state accuracy of the sinusoidal tracking and the steady-state accuracy under random disturbance are improved by two orders of magnitude,the pulse interference stabilization time isreduced by more than 0.7 s.The tracking accuracy of the photoelectric tracking turntable are well guaranteed.

Taking the shielding enclosure of airborne electronic equipment with apertures and slots in the center as the research object,a BLT (Baum-Liu-Tesche) equation which can calculate the shielding efficiency of the enclosure under plane wave irradiation is established.Aiming at the problem of large error in Robinson aperture array impedance calculation,the eccentricity coefficient Cmk is introduced to modify the aperture array impedance,and the modified BLT equation can accurately calculate the shielding efficiency of the enclosure with aperture array at any position.The calculation of shielding effectiveness of shielding enclosure at any polarization angle is realized by vector decomposition of electromagnetic wave electric field.The results show that the shielding effectiveness of the shielding enclosure increases with the increase of the number of apertures and slots.The increase of polarization angle enhance the electromagnetic shielding ability of shielding enclosure.By comparing the calculation results of BLT equation with the results of Robinson method and CST simulation,it is proved that BLT equation is accurate,which provides an effective method for the shielding effectiveness of shielding enclosure of computer-borne electronic equipment.

According to the characteristics of close distance and large yaw angle of aerial submarine search buoy,a route planning algorithm combining Dubins path and double ant colony algorithm is proposed.Compared with the traditional route planning that the linear route is firstly determined and then smoothed,the algorithm converts the straight path into the Dubins path in the iterative pathfinding stage,which reduces the risk of not reaching the target under the constraint of the turning radius,and the Dubins distance is used as the benchmark for judging,which is closer to the global optimal than the straight-line distance.Moreever,the parameters of yaw distance disturbance are added to guide ants to choose targets with smaller straight-line distance and yaw angle,and use the negative feedback of pheromones in different ant colonies to promote the search for new paths and improve the wayfinding ability.The simulation results proves that the algorithm accelerates the convergence speed of planning,effectively shortens the total distance of the route,and the average shortening range is more than 14.6%.

Large-scale complex systems such as industrial equipment and weapons usually go through storage,standby,testing,operation,maintenance and other states in the whole life cycle,and the internal wear rate of equipment is different in different states,which brings new challenges to the problem of remaining useful life prediction of equipment.In view of this,the degradation model of multi-state switching equipment is established based on Wiener process,and the remaining useful life of multi-state switching equipment is derived by combining semi-Markov model and discrete Markov chain model,and a Monte Carlo simulation method is proposed to solve it.In addition,expectation maximization algorithm and maximum likelihood algorithm are used to identify the parameters of the proposed model.Finally,the effectiveness of the proposed method is proved by simulation experiments and practical cases.