Aiming at the problem of keeping and changing the cooperative formation of manned/unmanned aerial vehicle (MAV/UAV) system in three-dimensional space, a method of MAV/UAV formation control with hierarchical structure is proposed.According to the MAV/UAV command and control mode and operation process, the control structure of MAV/UAV formation system is designed, in which the tracking controller of manned aircraft trajectory is designed by using dynamic surface control method, which solves the problem of repeated derivation of virtual control signal in backstepping control.Based on the limited centralized distributed formation control strategy, the UAV formation controller is designed, and the cooperative formation flight of multiple UAVs and a manned aircraft is realized.The stability of the controller is proved based on Lyapunov theory.Finally, the simulation of MAV/UAV formation in three-dimensional space is carried out, and the simulation results verify the effectiveness of the designed MAV/UAV formation control strategy.

Aiming at the problem of UAVs autonomous decision-making in close air combat, an autonomous guidance method for UAV based on Proximal Policy Optimization (PPO) algorithm is proposed.The rewards are reshaped, such as distance, angle, speed and mission constraint, a three-degree-of-freedom model of UAV is established, and the state and action of reinforcement learning are constructed on the velocity coordinate system.The simulation training is carried out on the model of PPO algorithm combined with the fully connected neural network(standard PPO algorithm) and the PPO algorithm combined with the long short-term memory network (improved PPO algorithm) respecitively.According to the training results, it can be proved that, compared with the standard PPO algorithm, the improved PPO algorithm proposed in this paper can handle the UAV autonomous guidance tasks that are highly correlated with time series more effectively.

Aiming at the problems of slow real-time tracking and low precision of unmanned Autonomous Underwater Vehicle (AUV) in complex hydrological environment, an adaptive predictive tracking control method based on Laguerre function is designed.Firstly, the trajectory tracking problem is transformed into quadratic programming design based on predictive control.Secondly, in order to solve the control variable mutation caused by AUV trajectory change and external disturbance, an adaptive predictive controller is designed by combining recursive least square algorithm.Then, the controller is reconstructed by Laguerre function to solve the slow response speed of the system caused by the large amount of computation.Finally, the simulation proves that the method can improve the response speed, anti-interference capability and robustness of the system.

In order to solve the problem of missed detection and false detection caused by occlusion in the process of target detection, a military target detection algorithm with image restoration module is proposed, which realizes the automatic recognition of military targets through deep learning, and then combines image restoration to achieve image enhancement of occluded targets.For the target detection module, a convolutional attention mechanism is added on the basis of YOLOv4 to enhance the sensitivity to target recognition and the feature extraction ability of the network.And the cross-iterative batch standard layer is adopted to improve the training efficiency of the model.The image restoration module is a double generator model designed based on the generation confrontation network.Considering that the integrity of the target image outline has a certain impact on image restoration and target detection, an edge generation network is added.The image restoration module aims to restore the part of the target that is occluded.The experimental results show that the target detection accuracy of the target detection algorithm with image restoration module reaches 79.63%, which effectively solves the problem of missed detection and false detection in the case of occlusion.

Aiming at the problems of motor failure and composite disturbance in coaxial multi-rotor UAV, a fault-tolerant control algorithm is designed by using adaptive neural network observer.Firstly, the motion model of the coaxial eight-rotor UAV including motor fault and composite disturbance is established, then the composite disturbance is approached by neural network and the fault factor is estimated by adaptive law, and the adaptive neural network observer is designed to estimate the system state.Finally, the backstepping fault-tolerant control law is designed for the attitude angle loop and angular velocity loop,and a filter is adopted to filter the virtual command signal, which suppresses the differential explosion and realizes the gradual stability of the UAV.The experimental results show that the proposed method has better stability and accuracy compared with the adaptive fault-tolerant control method, and the maximum tracking error is only 0.1°, which effectively compensates for the influence of composite interference and rotor motor fault,and improves the flight stability and fault-tolerant performance of the UAV.

Aiming at the problem that the non-linear system with non-uniform sampling is difficult to control due to external disturbance, a model-free adaptive quasi-sliding mode control algorithm is proposed by using data-driven strategy.Firstly, the parameter estimation error of pseudo-Jacobian matrix in the process of system dynamic linearization is regarded as an external disturbance.Secondly, the controller of the dynamic linearization model is designed by using sliding mode control.The algorithm is characterized by utlizing the inherent characteristics of sliding mode control to reduce the influence of external disturbance on the control function of the system, which effectively improves the control effect.Analysis is made on the stability of the control algorithm, and the effectiveness of it is verified by simulation of a discrete-time nonlinear system with non-uniform sampling.

This paper aims at the beam scheduling management problem of phased array radar in the process of multi-target tracking, and in the traditional beam scheduling, only the target with the smallest error among all targets is optimized, which leads to the problem that some targets with larger errors will not converge or even be lost when facing high maneuvering targets.In view of this, a radar beam scheduling algorithm based on maximum average covariance is proposed, which combines the advantages of the Interactive Multi-Model (IMM) filtering algorithm with that of phased array radar without inertial sampling,to realize beam scheduling for stable tracking of multiple targets,and study the effect of expected covariance and process error covariance (noise error) on the tracking process.The simulation experiments show that the algorithm can effectively keep the covariance of multiple high maneuvering targets within the expected range and achieve dynamic convergence, and can affect the confidence and convergence speed of the tracking process by adjusting the expected covariance and process error covariance, so as to achieve stable tracking of multiple targets.

The research on metamaterials has developed greatly since it was put forward more than twenty years ago, and the absorbing property of metamaterials is one of the research hotspots.Aiming at this, the proposal of metamaterials absorber is briefly introduced, the development of metamaterials absorber is summarized, and several basic types of metamaterials absorber are emphatically introduced, such as the lumped elements loaded metamaterial absorber, multi-unit or multi-layer metamaterial absorber, multi-type materials composite metamaterial absorber.Finally, the application prospect of metamaterial in stealth technologies is discussed, and some challenges and suggestions for the practical application of metamaterial in stealth technologies are also given.

Aiming at the formation control problem of distributed swarm systems, a distributed formation control method under the condition of directed topology is proposed.Firstly,by introducing auxiliary functions into the control protocol and using variable substitution, the formation control problem is simplified to the stability problem of the autonomous system, which simplifies the complex control problem.Secondly, the control gain matrix is designed by solving linear matrix inequality, which is simple in form and low in complexity.Then the stability of the system is analyzed by Lyapunov method.The simulation results show that the multi-UAV system can effectively realize formation tracking control under the distributed control protocol.

Anomaly target detection is a research hotspot in hyperspectral imagery processing.Aiming at the problems of current anomaly target detection algorithms, a new anomaly target detection algorithm is proposed by combining low-rank tensor decomposition with sparse representation for hyperspectral imagery.The algorithm utilizes the spatial spectrum and spectral characteristics by solving the problems of background, anomaly target and noise in hyperspectral imagery.Firstly, the algorithm uses the low-rank tensor decomposition to restore the original hyperspectral imagery, so that the image quality is improved, and the anomaly target becomes prominent and easy to be detected.Then, the sparse difference index is used for anomaly target detection to obtain the required anomaly detection results.Finally, simulation experiments are carried out by using real hyperspectral images.The results show that the new anomaly target detection algorithm has the characteristics of high detection accuracy, low false alarm rate and good robustness.

Aiming at the attitude tracking control problem of quad-rotor UAV under input constraints, a parametric control method is proposed.Firstly, a second-order, full-actuated model of quad-rotor attitude error is established, and a parametric controller is adopted to make the attitude error system equivalent to a first-order steady closed-loop system, which make the attitude errors converge to the tracking signal.Considering the input constraints and for better tracking performance of the quad-rotor, an intelligent optimization method is adopted and the particle swarm optimization algorithm is used to optimize the parameter matrix in the control law.The simulation experiment shows that compared with the backstepping method, the proposed parametric controller can track faster the attitude commands with input constraints, which has better transient performance, thus can realize high-precision attitude tracking of the quad-rotor UAVs.

In recent years, with the maturity of artificial intelligence, information network and other technologies, avionics system has developed rapidly, and discussions about how to apply avionics system to combat in the future have emerged in an endless stream.By integrating various innovative activities of avionics system with three-level analysis method, a reasonable and sustainable methodology for the future is constructed to study the field of avionics system.Firstly, the background and present situation of avionics systems is systematically reserarched, and then the three-level analysis method model is deeply analyzed, the application of avionics system is analyzed in detail from three aspects, the reduction of the efficiency of manned avionics system and its reasons, the continuous improvement of the current avionics system and the induction of two types of innovations, and a highly autonomous unmanned avionics-led vision in the future and its three driving factors.

Aiming at the problems such as the difficulty in locating accidental faults and the inadequacy of system diagnosis capability in fault diagnosis and equipment maintenance of airborne weapon control system, an integrated diagnosis platform of weapon control system built on the basis of open system architecture and commercial off-the-shelf component is proposed.The platform, oriented to muptiple types of aircraft, meets the II-level repair system and is able to provide the comprehensive support in the full life cycle of equipment from design, operation to maintenance, which lays a solid foundation for solving the contradiction between equipment complexity and diagnosis/maintenance, maximizing the diagnosis efficiency, reducing the maintenance cost and ensuring the readiness of weapon control system.

In the process of shooting range test, the autonomous and stable target tracking of photoelectric theodolite plays an important role.The current tracking mode using single information source is not suitable for different task types and environmental characteristics of shooting range.Based on the photoelectric theodolite tracking system in the shooting range as the research platform, combined with the local or external information data measured in real time during the tracking process, a multi-source optical measurement information fusion tracking method and the corresponding data processing algorithm is proposed.The maximum estimation error of tracking is 2.49′ and the estimated random error is 8″, which meet the tracking accuracy requirement of photoelectric theodolite test task.It is proved that this method can guarantee the stability and reliability of tracking process effectively.

Facing the complex three-dimensional environment, the computational complexity of the traditional path planning algorithm is extremely increased, and the original effect is lost.Reinforcement learning can be independent of the accurate environment model, and its overall efficiency is much higher than that of the traditional algorithms.Aiming at the path planning problem of AUV in 3D environment, based on the establishment of collision avoidance detection model and gym simulation environment, the optimized PPO algorithm design and model training of improved network structure are carried out.Through simulation experiments, the accuracy and effectiveness of the algorithm are verified.

Many applications rely on the accuracy of data understanding, and semantic image segmentation effectively solves this problem, which provides the necessary context information for scene understanding based on pixel level.A U-Net network structure, ECAU-Net, is proposed based on ResNeXt50.Compared with the general convolution operation, ResNeXt50 has a stronger feature extraction ability.In fusion process, Efficient Channel Attention module(ECA-Net) is introduced to further enhance the ability of feature representation to discriminate scene segmentation.In addition, the introduction of dilated convolution in the overall network structure expands the receptive field of the image without changing the size of convolution kernel, thereby maximizing the performance of the network.The experimental results show that on the CamVid dataset, compared with U-Net, ECAU-Net has improved the Acc, Acc class, MIoU and FWIoU respectively by 2.1%, 8.6%, 8.2% and 3.2%.

Aiming at the low recognition accuracy of radar signals under the low Signal-to-Noise Ratio (SNR), we propose a method based on multi-time-frequency images fusion and dilated residual network.Firstly, radar signals are transformed into different time-frequency images by various time-frequency analysis methods, and these time-frequency maps are fused and processed.Then, a new network model is constructed, which combines the dialated residual network with Distinguishing Feature Fusion Extraction (DFFE) module to identify 10 types of radar signals.Simulation results show that when SNR is -6 dB, the overall recognition accuracy of the proposed method for 10 types of signals reaches 98.7%.

In order to improve the regularity, safety and stability of local dynamic obstacle avoidance for Unmanned Surface Vehicle(USV), a local dynamic obstacle avoidance method based on drift angle estimation is introduced.Firstly, the velocity space of Dynamic Window Approach(DWA) is selected according to the International Regulations for Preventing Collisions at Sea(COLREGS).Secondly, the collision-free predicted trajectory is selected based on the predicted movement state of the obstacles.Finally, considering the ocean disturbance, the lateral position error is used to estimate and compensate for the drift angle.The simulation experiment results verify the effectiveness of the method.

Avionics Full Duplex Switched Ethernet (AFDX) introduces virtual link to realize logical separation of physical bandwidth resources.The asynchronous arrival and multiplexing output of data frames may cause the delay jitter of virtual link, which eventually leads to the uncertainty of end-to-end delay analysis of traffic.Aiming at the deterministic problem of AFDX data frame transmission jitter and delay fluctuation, this paper designs an AFDX end system scheduling strategy based on time window, which can effectively reduce the transmission jitter of data frame and provide certainty.The analysis and simulation of the scheduling strategy show that the scheduling strategy can effectively restrict the worst-case jitter of the data transmitter, and improve the certainty and reliability of the whole AFDX network.

Due to the high complementarity of visual and inertial sensors, fusion of visual and inertial data is a common method for attitude measurement.With the development of airborne and vehicle-mounted equipment, the traditional attitude measurement on a stationary base cannot meet the needs of some practical application scenarios.Because of the interference of the movement information of the base, it is difficult to measure the attitude of the target on the moving base accurately and quickly.In order to make use of the advantages of visual/inertial fusion, and realize the relative attitude measurement of a target on moving base, this paper proposes an attitude measurement system based on monocular vision and dual gyroscopes.The establishment of coordinate system in measurement system and the relative relationship between coordinate systems are analyzed, and the scheme of coordinate system normalization is put forward.The cubature Kalman filter is used to fuse the visual with inertial data to achieve high-precision, wide-range, fast and stable attitude measurement.The experimental verification is carried out on the platform of vision and inertia fusion attitude measurement system, and the results prove the effectiveness of the system coordinate system normalization scheme and fusion algorithm.The experimental results show that the built system has high accuracy, and the RMSE of the fused pitch angle and azimuth angle is less than 0.1°.