To achieve efficient and accurate multi-target search of a UAV swarm,and with consideration of the wide detection ranges of high-altitude UAVs and the strong maneuverability of low-altitude UAVs in the swarm,a novel cooperative target search method is proposed for heterogeneous UAV swarms,which can realize rapid regional search and accurate multi-target search according to the characteristics of different UAVs.First,the high-altitude UAVs use a region search algorithm based on digital pheromones to determine the targets presence area,which realizes area overlapping in short time.The low-altitude UAVs then proceeds to the target area guided by the high-altitude UAVs and use a precise search algorithm based on an Improved Wolf Pack Algorithm (IWPA) to accurately locate the target.To address the limitations of traditional wolf pack algorithm,such as insufficient information sharing,fixed step size,and susceptibility to local optima,improvements have been made by using particle swarm optimization,adaptive parameter adjustment,and differential evolution methods.These enhancements improve the algorithms global optimization capability and convergence speed.Simulation experiments demonstrate the effectiveness and superiority of the proposed model and method for target search tasks in complex environment.

The formula of target tracking Bayesian Cramér Rao Lower Bound (BCRLB) in the range direction is derived,and the BCRLB in the range direction is used as a quantitative indicator of target tracking performance.For dual aircraft target tracking scenarios,BDWF,a BCRLB algorithm based on range direction Weighted Fusion is proposed using interactive multi-model Kalman Filter (KF).The simulation experiment shows that:1) The target tracking trajectory of the proposed algorithm fits well with the real trajectory,and the Root Mean Square Error (RMSE) during the tracking process can quickly converge to a stable value;and 2) Compared with the target tracking algorithm based on signal-to-noise ratio weighted fusion,the target tracking accuracy of the proposed algorithm is higher.

In complex electromagnetic environments,the image transmission signal of non-cooperative UAV is prone to be covered by interference,and traditional detection methods cannot effectively identify the signal,so a UAV image transmission signal recognition method based on the improved AlexNet model is proposed.According to the time-frequency spectrum characteristics of the image transmission signal in interference coverage scenarios,this method improves and optimizes the AlexNet model and deepens the network structure without increasing computational complexity by splitting the convolution kernel,reducing the number of nodes in the fully-connected layer,and adding a global average pooling layer,which effectively improve the image transmission signal recognition capability.In indoor anechoic chamber and real-world field environments,the time-frequency image datasets under different interference intensities are prepared to train the model.The results show that,when the Signal to Interference plus Noise Ratio (SINR) is -15 dB, the improved AlexNet model can still maintain a verification accuracy above 90%,and the unit training time can be shortened by more than one second compared with other CNN models.

The allocation of electronic interference resources of ship formations plays a very important role in modern naval warfare.In order to solve the problem of allocation of interference resources for self-defence anti-missile operations in naval formations,the data fusion and threat ranking of targets are completed by the joint probabilistic correlation algorithm and the technique for order preference by similarity to an ideal solution.The interference resource allocation index is proposed,and the interference resource allocation model is constructed by combining fuzzy theory.Secondly,to address the problems of slow solution speed and difficulty in convergence of the large-scale target interference resource allocation,the clustering idea is introduced to improve problem solving speed and stability.By improving the K-means clustering method,a principle of interference resource allocation is to ensure that the interference resources of a ship cannot be exhausted.Finally,the established interference resource allocation model is solved based on genetic algorithm,and the significance of introducing clustering to the large-scale optimal allocation problem is demonstrated by comparing the situations with and without introducing the clustering method.The simulation verifies the real-time performance and stability of the solution of the optimal allocation method after the introducing the clustering.

To solve the problem of blurring and degradation of images captured in foggy conditions and to overcome the limitations of traditional dehazing algorithms based on prior of dark channels,a dehazing algorithm using wavelet transform and using prior theories of dark and bright channels is proposed.By using the prior theories of dark and bright channels,the corresponding atmospheric light value and transmittance are obtained respectively,and the transmittance of dark channels obtained from multi-scale filter windows is fused by using the wavelet transform.The final atmospheric light value and transmittance are obtained by using linear fitting function and guided filter function,and the clear haze-free image is restored by using the atmospheric scattering model.Finally,the subjective and objective analysis methods are used to simulate and compare the proposed algorithm with the previous representative dehazing algorithms.The experimental results show that the proposed algorithm can effectively remove the haze in the original image on the premise of ensuring the real-time performance,and it avoids the color distortion caused by excessive enhancement in the traditional dehazing algorithms.It performs well in terms of subjective visual effects,and is also excellent in terms of objective evaluation indexes,which verify the feasibility,effectiveness and superiority of the proposed algorithm.

To address the problems as degradation of algorithm performance because the target is blocked by interference and confusion of target and interference in infrared target recognition,a new airborne infrared target recognition algorithm is proposed based on residual dense connection attention.Firstly,to fuse shallow and deep features across layers,obtain fusion depth features,enhance feature reuse performance and strengthen semantic information,an improved residual dense block is proposed.Secondly,to enhance the adaptive expression ability of the fused depth features,a parallel mixed attention block is designed.Finally,the test on a large quantity of infrared datasets shows that the algorithms average recognition accuracy is increased by 1.9 percentage points compared with that of the GoogLeNet algorithm,which proves the validity of the algorithm.

With the advantages of high transmission rate and great security,the Aeronautical Mobile Airport Communications System (AeroMACS) has become an important part of the airport and air-ground communication network.Aiming at the problems of outdated Channel State Information (CSI) caused by fast time-varying channel during the high-speed movement phase of aircraft take-off and landing,and worsening of communication quality caused by large Doppler frequency shifts,a channel prediction method based on transformer neural-network is proposed by using a multi-head self-attention mechanism.Modulation Coding Scheme (MCS) for WiMAX and 5G dual-mode of AeroMACS is adjusted according to the Signal-to-Noise Ratio (SNR) predicted in real time.Simulation results show that,compared with the other three artificial intelligence methods,the proposed transformer network-based channel prediction method achieves higher accuracy and enhances the total throughput of the system,which can effectively cope with the problem of outdated CSI and improve the system communication performance.

Aiming at the problem that a single traditional discriminator Generative Adversarial Network (GAN) tends to ignore the brightness information and edge information of infrared light in the outdoor scene at night,a fusion algorithm of infrared and visible images based on attention mechanism and dual discriminators is proposed.Firstly,in order to pertinently obtain target information of infrared images and background texture information of visible images,a channel attention mechanism is introduced into the generator network.Secondly, the GAN with two discriminators is used,and a new discriminator input is designed to improve the training stability while better preserving the source image information.Finally,the loss functions are set as adversarial loss,structural similarity loss and gradient loss to constrain the discriminator for generating fusion images with rich details.The experimental results on the TNO dataset show that the fusion image obtained by this algorithm has more significant gradient changes and clearer edges,which is more in line with human visual effects.

In modern warfare,a large amount of command and control and intelligence information is transmitted through satellites.The anti-jamming ability of satellite communication earth stations is the basis for ensuring the smooth flow of information.As for the case that multiple small earth stations cannot communicate due to jamming,the Minimum Variance Distortionless Response (MVDR) algorithm based on the reflector antenna array is proposed.Firstly,the elements of antennas at satellite communication earth stations are analyzed,the gain expression of the reflector antenna is given,and a mathematical model for the reception of signals by the reflector antenna array is established.Then,the mechanism of the MVDR improvement algorithm for beamforming is studied,the calculation formulas for the output Signal to Interference plus Noise Ratio (SINR) and optimal weight vector of the reflector antenna array are derived,the reasons for the poor anti-jamming effect of uniform linear arrays at certain angles are analyzed,and a linear array whose deployment approximates the minimum redundant array with mutually prime element positions is proposed to address this issue.Finally,the effectiveness of the proposed algorithm in improving the anti-jamming performance of the array is verified through multiple simulation experiments.The simulation results show that the proposed algorithm can improve the satellite direction gain of the reflector antenna array,increase the output SINR of the array,and effectively suppress the downlink jamming of satellite communication.

The nycterohemeral star survey in short-wave infrared band is expected to realize the all-time automatic navigation in near-earth space,in which the star map recognition algorithm is one of the key technologies for realizing the all-time navigation.The traditional triangle algorithm is prone to redundant matching and miss-matching in recognition due to its low dimension of matching features when the number of navigation stars increases.To address this problem,this paper proposes a dual-feature based shortwave infrared star map recognition algorithm,which selects dual high-dimensional features of triangles area and tangent circle radius as matching features,and reduces the computational complexity of matching recognition by constructing a K-vector index of area features and applying K-vector lookup method.In addition, an optimized selection strategy of observation triangles is proposed to reduce the computation cost in the matching process and improve the recognition speed of the algorithm.The test shows that:1) The recognition rate is higher than 95% when the noise of star point position is less than 2 pixels;and 2) The recognition rate of the algorithm can reach 87.6% when the number of pseudo-stars does not exceed 50%.The feasibility of the proposed algorithm is verified by actual star observation test.Compared with the improved triangle algorithm,the algorithm has obvious advantages in the recognition speed,recognition rate and noise resistance capability.

Waveguide display technology is a new type of display technology and is important in AR field.At first,the principle,classification and characteristics of waveguide display technology are expounded,the technical characteristics are presented,and the development trend of technology is sorted out.Then,the application status of the technology in the field of airborne Head-Mounted Display (HMD) is summarized.An analysis is made on the advantages and disadvantages of the airborne waveguide HMDs and the challenges of their application.Based on this,the future development direction of promoting waveguide display technology to the airborne field is given and its development prospects are predicted.

MAGIC CARPET is a very successful carrier landing control technology for carrier-based aircraft in the United States in recent years,and will be the main landing mode in the future.The stability of MAGIC CARPET control is studied.Firstly,the control structures of MAGIC CARPET landing and conventional landing are analyzed.Then,a real-time human-in-the-loop simulation experiment is conducted,which consists of two groups:MAGIC CARPET control and conventional control,with each group divided into two situations of immobile throttle and auto-throttle-on.The experiment shows that for the stability performance,including flight path stability and controllability,speed stability,and angle of attack stability,the order from high to low is:MAGIC CARPET control with auto-throttle-on,MAGIC CARPET control with immobile throttle,conventional control with auto-throttle-on,and conventional control with immobile throttle,which reveals that MAGIC CARPET has excellent stability performance.Finally,the stability mechanism of MAGIC CARPET control and conventional control is comprehensively analyzed from the perspectives of D-V curve,signal-flow graph,complex domain,etc.The research results provide reference for aircraft landing and direct force control.

Aiming at the problem that it is difficult for the existing remote sensing image denoising algorithms to fuse shallow image features into deep image information,a remote sensing image denoising network that fuses multi-scale features is proposed,which consists of an asymmetric convolution block,a dilated attention block,a residual projection block and a residual fusion block.First,asymmetric convolution is used to preliminarily extract features and reduce a large amount of information redundancy in the network.Then,the multi-scale features are extracted through the dilated attention block to learn rich context information, and the fusion of the extracted multi-scale features is more conducive to denoising and retaining more edge texture details of the image.The residual projection block collects a large amount of contextual and spatial information from the multi-scale features,and finally the residual fusion block generates a residual image to remove noise.Experimental results show that the proposed network outperforms several advanced image denoising algorithms in both quantitative and qualitative evaluations on the NWPU-RESISC45 and UCMerced~~LandUse remote sensing image datasets.

The color fusion image quality based on scene understanding measures the complexity of observers accurately understanding scene content through fused images,which is usually obtained through subjective assessment.In order to improve the stability of the subjective assessment,the abstract concept of color fusion image quality based on scene understanding was concretized into two performance indicators:segmentation of regions and discrimination of objects.Taking village scenes as example,15 people were organized to subjectively assess the quality of 42 color fusion images of the village scenes without thermal targets.The subjective scores were analyzed by using structural equation model.The prediction model of color fusion image quality based on scene understanding was established and the correctness of the model was verified.The results show that the color fusion image quality based on scene understanding in village scenes can be predicted by two factors of color harmony and color naturalness,and the image quality is more manifested in the performance of discrimination of objects.

o solve the problem that it is difficult to establish the mathematical model of the system controlled by adaptive discrete-variable structure,an improved model-free adaptive control method based on compact format discrete-time nonlinear system,ICF-MFAC,is proposed.Firstly,the error feedback output control item is added to the CF-MFAC method with only one time-varying integration term,so that the controller can set internal parameters online and thus improve the dynamic performance of the system.Then,the convergence of the ICF-MFAC method is proved by rigorous mathematical derivation.The theoretical analysis indicates that the ICF-MFAC method improves the utilization rate of pseudo-partial derivative estimates and accelerates the response speed of the system.Finally,simulation result of the test on quad-rotor aircraft shows that:compared with the CF-MFAC method,the control error of the improved control method is reduced by 6.12% with no overshoot and small tracking error,while the control effect and robustness are always maintained.

Aiming at the problems of detail loss,color distortion,and incomplete dehazing of image dehazing algorithms,this paper proposes a dehazing algorithm with channel estimation of the upper and lower boundaries of the light veil and adaptive fusion of atmospheric light.Firstly,the upper boundary of the atmospheric light veil is estimated according to the minimum channel of the foggy image,and the lower boundary of the atmospheric light veil is obtained based on B channel of the foggy image,RG channel difference compensation,and linear transformation.The upper and lower boundaries are optimized for approximation,and the two boundaries are fused by the wavelet transform. Secondly,the average value of the three channels of RGB and the middle channel of brightness are used to obtain different rough atmospheric light,and the average brightness value is used as the regulator for adaptive fusion.Finally,the haze-free image is recovered by using the haze removal model.The experimental results show that the images recovered by the proposed algorithm are thoroughly defogged,clear and natural,and can retain more image details with good effectiveness and robustness.

The problem of across range unit walking can deteriorate the performance of Orthogonal Frequency Division Multiplexing (OFDM) modulated radar in high-speed weak target detection.Therefore,a high-speed weak target detection algorithm for OFDM radar is proposed based on Prewitt operator and Radon transform.Firstly,the problems of across range unit walking caused by multi-pulse accumulation and of the noise interference on Radon transform are analyzed.Then,the Prewitt operator is used for convolution operation to suppress noise and retain echo energy,and Radon transform is used to accumulate energy along the range unit walking diagonal.Theoretical analysis and simulation results show that:Compared with Hough transform and Moving Target Detection (MTD),the proposed algorithm has better high-speed weak target detection performance at low Signal-to-Noise Ratio (SNR),and it can effectively achieve multi-target detection.

In order to achieve unmanned-mode intelligent collection of tank target information on the frontline battlefield in information-based warfare,an improved algorithm based on YOLOv4 is proposed for ground tank target detection.Based on the original YOLOv4 object recognition algorithm,the multi-layer feature stitching module is employed to enhance the transmission and flow of feature information.The global information acquisition module is used to better capture global feature information.The multi-scale information fusion module is utilized to expand the scale of feature fusion.The decoupling detection head module is added to decouple the tasks of target classification and position regression,enabling more thorough network learning.The experimental results show that:1) Compared with the YOLOv4 algorithm,the improved YOLOv4~~Modify algorithm achieves higher recognition accuracy,with a 10.2 percentage points increase in Recall and a 4.3 percentage points increase in mAP;and 2) The improved YOLOv4~~Modify algorithm can accurately identify tank targets of different scales in complex environments,addressing the original algorithms drawback of missed detection for small tank targets,and providing visual technological support for information-based warfare.