With the continuous development of Multifunctional Phased Array Radar (MPAR),its characteristics of multifunctional states and strong self-adaptive capability make the traditional behavior identification method unable to adapt to the needs of modern warfare.In this regard,a radar behavioral state identification method based on the modified D-S evidence theory is proposed.The method takes the spatial motion state and antenna scanning mode as auxiliary features and fuses the double-chain HMM estimation results through the modified D-S evidence theory to obtain the final identification results.The simulation experiment results show that the proposed method can effectively solve the evidence conflict problem which cannot be handled by the traditional D-S evidence theory,and it shows stronger robustness in the case that the accuracy of radar word extraction decreases.

The consistency of uncertain multi-agent systems based on dynamic event triggering mechanism under network attacks is studied.Traditional event triggering mechanisms cannot effectively save network resources by changing the triggering rules.Dynamic event-triggering mechanism is introduced.On the basis of the dynamic variables of the mechanism,the event triggering conditions are dynamically adjusted according to the running state of the system,so as to reduce the frequency of control updates.The mathematical model of the closed-loop control system is established under the delay control input,network attacks and the trigger of dynamic events.The consistency of multi-agent systems under dynamic event triggering control is proved by Lyapunov stability theory and Linear Matrix Inequality (LMI) theory.Finally,the effectiveness of the proposed method is verified by simulation experiments.

OFDM signal is one of the commonly used transmitting waveforms in integrated radar communication system.This paper analyzes and compares the characteristics of two kinds of OFDM signals of different sequences(ZC sequence and Oppermann sequence),and analyzes and compares the performance of the two waveforms from the perspectives of autocorrelation,Peak-to-Average Ratio (PAR),ambiguity function and Doppler tolerance.Theoretical analysis and simulation experiments show that the signals of ZC sequence and Oppermann sequence both show concentrated main lobe in autocorrelation function,but the side lobe of ZC sequence is higher than that of Oppermann sequence in autocorrelation function.The PAR of ZC sequence is better than that of Oppermann sequence.ZC-OFDM signal and Oppermann-OFDM signal have the same “oblique blade” ambiguity function.The Doppler tolerance of ZC-OFDM signal is better than that of Oppermann-OFDM signal.In conclusion,ZC-OFDM signal has better radar detection performance than Oppermann-OFDM signal.

The traditional Feature-Based (FB) signal modulation recognition algorithms suffer from the problems of low recognition accuracy,difficult feature extraction and poor generalization performance.To solve the problems,a signal Automatic Modulation Recognition (AMR) method based on MSPP-CNN is proposed,which combines Convolutional Neural Network (CNN) with Multi-Scale Pyramid Pooling (MSPP).In the proposed method,multi-scale pyramid pooling is used to improve the nonlinear feature extraction ability of the model for different modulation signals,so that the model has better feature representation and generalization performance.In the construction process of the CNN model,different convolution,pooling and activation methods are used to optimize and verify the model,so as to ensure the rationality of model structure and parameters.The experimental results show that the recognition accuracy of the proposed method under signal-to-noise ratio of -18 dB,0 dB and 18 dB reaches 56%,62.98% and 92.04% respectively.In addition,a comparison with other traditional feature extraction algorithms and CNN algorithm is conducted,which verifies the effectiveness and high recognition accuracy of MSPP-CNN.

Aiming at the resource allocation problem in two-layer scheduling of the IMA architecture,a two-layer resource allocation method based on chaotic Logistic mapping and adaptive NSGA-Ⅱ algorithm is proposed.Firstly,according to the task scheduling and partition scheduling requirements of the IMA architecture,and in consideration of resource allocation balance,a partition resource allocation model and a node resource allocation model are constructed respectively.Secondly,the improved NSGA-Ⅱ algorithm is used to optimize the model,chaotic Logistic mapping is introduced to generate the initial population to improve the ergodicity of the initial population,and the adaptive crossover and mutation operators are used to enhance the search performance and convergence rate of the algorithm.Finally,numerical examples of different scales are selected for experiment,and the experimental results show that the proposed method can effectively solve the resource allocation problem of the IMA architecture,improve the allocation efficiency and optimize the allocation result.

Aiming at a class of nonlinear systems with unknown time-delay and dead-zone inputs,an adaptive asymptotic tracking control algorithm based on novel dynamic surface is proposed.Firstly,a nonlinear filter with a positive time-varying integral function is constructed,which is different from the traditional low-order linear filter.It can not only avoid the problem of “differential explosion” and reduce the computational burden,but also effectively compensate for the boundary layer error.Secondly,the radial basis neural network and a new inequality are utilized to deal with the unknown nonlinear time-delay terms.Finally, the neural network error and the unknown parameters of the dead zone are estimated online.Theoretical analysis shows that the semi-global uniform boundedness of all signals in the closed-loop system is achieved and that the tracking error converges to zero.The simulation results have verified the effectiveness of the proposed control algorithm.

Phase coding is a typical pulse modulation form of pulse compression radar,and the jamming of phase coded pulse compression radar is a research hotspot in the field of electronic warfare.In view of the shortcomings of traditional noise jamming against phase coded pulse compression radar,a multi-phase refactoring jamming modulation technology is proposed in confrontation of phase coded pulse compression radar.The radar signal samples obtained from reconnaissance are divided into several segments,which are reconstructed to generate coherent jamming signals.The jamming signals can obtain coherent processing gain of radar signals and the jamming power utilization rate is high.The simulation results show that this technology can modulate the jamming signals according to radar working modes,and generate the expected jamming signal waveform with good jamming effects.

In order to solve the problems of color distortion,low contrast and fuzzy details of underwater images caused by medium scattering and absorption,an improved algorithm of dark channel prior for underwater image enhancement is proposed.White balance processing is adopted to correct the blue (green) color deviation of the underwater image,and homomorphic filtering is conducted on the L component of the image in LAB space,so as to obtain a dark detail highlighted image.In RGB space,CLAHE processing is conducted to enhance the image contrast and solve the image atomization problem,and MSRCR processing is conducted to improve image color saturation and equalize image brightness.The fusion weight coefficient is calculated based on the dark channel prior image,and weighted fusion of the three images and detail enhancement are conducted to obtain the final enhanced image.The experimental results show that the proposed algorithm can effectively remove color distortion,and the enhanced image presents high contrast and clearer details.

The continuous research on panoramic vision in the field of vision shows that the accuracy of moving object detection based on image sequence obtained by fisheye lens is low,and the robustness is not high when disturbed by noise.Aiming at the above problems,an improved method of moving object detection based on panoramic vision is proposed.Firstly,the algorithm uses the five-frame difference method to process the image,and uses the difference between adjacent five frames to separate the foreground from the background,thus effectively reducing the target-hole problem.Then,the ghost phenomenon in detection is effectively overcome by improving the adaptive learning rate and updating the number of Gaussian distribution in the Gaussian mixture model.Finally,the target result is obtained by morphological processing.Experiments show that the improved algorithm provides more reliable detection results than the traditional panoramic image moving target detection does,and the target detection rate is higher.

In modern military operations,the research and development of manned/unmanned aerial vehicle cooperative operation have vital impact on the control of battlefield situation and the improvement of combat capability.Aiming at the problem of manned/unmanned aerial vehicle cooperative operation,the high-level abstraction modeling method of metamodel is used to analyze the architecture and mechanism of manned/unmanned aerial vehicle cooperative operation,and three aspects including operation concept,process control and operation mission activities are discussed.The metamodels of operation concept,process control and operation process are established to analyze the conceived design of a specific mission,and operational architecture design is improved profoundly from two aspects of operational execution modes and combat interaction optimization.In-depth research and analysis on the operational modes of manned/unmanned aerial vehicle cooperative air combat are conducted,which provides a reference for designing and building a manned/unmanned aerial vehicle cooperative engagement system.

With the rapid development and application of UAVs,the popularity of UAVs has also caused certain security risks to public security,military security and personal privacy.UAVs has the characteristics of high flying speed and small volume,so how to accurately and quickly find and locate the position of UAVs is a challenge.A real-time detection algorithm for YOLOv3 UAVs based on SandGlass Bottleneck Block is proposed.Firstly,the original three feature scales are extended to five feature scales to make full use of multi-scale information to help improve the detection accuracy of small targets.Then,the SandGlass Bottleneck Block is stacked as the backbone network part of the method,and the SandGlass Bottleneck Block is taken as a lightweight network to accelerate the model,which uses the channel attention mechanism to focus on more important channels in the splicing part after upsampling unformation and suppresses unfavorable information.In order to verify the proposed algorithm effectiveness,a UAVs data set is generated based on complex urban background.Experimental results show that the proposed algorithm can achieve 98.92% accuracy and a recall rate of 98.76%,achieves a real-time detection speed of 62.37 FPS on 1080Ti graphics card,the model weight is only 5.38 MiB,which further provides the possibility for real-time target detection on embedded platforms and mobile devices.

In order to meet the requirements of the next generation 6G network for large capacity,high speed and low delay on optical fiber transmission system,a second-order multi-pumping Raman fiber amplifier is designed by using quartz fiber as transmission medium and artificial bee colony algorithm.The Runge Kutta method and the shooting method are used to solve the power coupled wave equation of the second-order multi-pumping Raman fiber amplifier,and then the artificial bee colony optimization algorithm is used to optimize the different arrangement structures of the four pumping beams.Through Matlab simulation,the best bi-directional pumping structure FFBF is obtained from 14 types of bi-directional pumping structures.The average gain of the structure is as high as 24.8 dB in the bandwidth of 100 nm,and the gain flatness is only 0.78 dB,which provides a reference for the design and optimization of Raman fiber amplifier in 6G network.

UAV object detection can be used in anti-UAV scenarios.To facilitate algorithm deployment on embedded devices,a lightweight model is often required.YOLOv4-tiny,the lightweight version of the YOLOv4 object detection algorithm,has a fast detection speed with relatively simple network structure and low detection accuracy.In order to further improve the detection accuracy,the model of YOLO-L2 is proposed.The backbone network of YOLOv4-tiny is selected for feature extraction,and a path aggregation network based on coordinate attention is used for feature fusion.In the process of fusion,a set of learnable coefficients are used for weighting.A cascade residual module named ResBlock-L2 is embedded in the deepest feature layer to enlarge receptive fields and fuse features with different receptive fields.The bounding box loss function MEIoU is proposed to replace CIoU.Compared with YOLOv4-tiny,the improved algorithm improves the mAP by 3.19% and 3.95% respectively in VOC dataset and self-made UAV-L dataset, and meets the real-time requirements.

Aiming at low efficiency of the existing noise image classification,a noise image classification algorithm based on the improved Darknet is proposed.The 1×1 convolution layer in the output part of Darknet network is removed,the number of convolution kernels in Layer 19 is changed to four,and Softmax layer is added to the end of the network,so that the classification function of the network is realized.Dropout is introduced after the passthrough layer and after Layer 6,Layer 7 and Layer 8 respectively,and L2 regularization is introduced in the convolution layer,so as to avoid network over-fitting.Layer 10 and 11,Layer 12 and 13,Layer 15 and 16,Layer 17 and 18 of the network are changed into four residual blocks to avoid gradient disappearance when updating the weights in back propagation.20000 images are taken from CIFAR-10 data set,and four kinds of noise,that is,Gaussian,salt,speckle and Poisson,are added respectively after 128×128 size transformation.One-hot coding is carried out for each image according to its category.Finally, the images and labels are made into a training set,a verification set and a test set.The experimental results of the four algorithms show that the accuracy of the improved Darknet network for color noise image classification can reach 0.904,which is much higher than that of the other three algorithms.

To deal with the difficulty of remote sensing image target detection in foggy weather,an improved method based on Mask R-CNN is proposed.The defogging algorithm is added on the basis of Mask R-CNN, which improves the mean average precision by 18.71% in foggy weather and effectively improves the target detection effect in fog.In order to further improve the mean average precision of multi-scale target detection in remote sensing images,a recurrent neural network based on optimal feature combination is used to replace the structure of feature pyramid,which reduces the loss of feature information in the transmission process.The Region Proposal Network (RPN) is redesigned to generate the sizes of the candidate box,and Soft-NMS is used to screen the candidate box to reduce the regression error of the candidate box.Experimental analysis shows that the mean average precision and recall of the improved algorithm are improved by 5.37% and 6.37% respectively.

Aiming at the jumping error of phase unwrapping in gray code phase-shifting fringe projection 3D measurement,an adaptive compensation method for continuous phase is proposed.Based on the binocular structured light 3D measurement system,the continuous phase map of the object to be measured is obtained through the phase-shifting method in combination with the positive and negative gray code method.Then,the wrong pixels are searched for on the continuous phase map.Finally,the improved adaptive median filter method is used to compensate for the jumping error pixels in phase unwrapping and the invalid pixels generated by noise and shadow.The experimental results show that in comparison with the traditional gray-code phase-shifting method and the method of applying fixed-value median filter to continuous phase map,the proposed method can compensate for most of wrong pixels on the continuous phase map while retaining the details of the object to be measured as much as possible,which improves the stability and accuracy of 3D topography measurement.

Aiming at strict-feedback nonlinear systems subject to external mismatched disturbances,the fixed-time control problem is studied.Based on backstepping control strategy and Lyapunov stability theory,the design steps of a novel disturbance observer and controller are given,which can realize global fixed-time stability.The traditional fixed-time backstepping control algorithms often suffer from singular problems caused by repeated differential operation of the virtual control laws.To address this problem,a smooth backstepping controller is designed without using the fractional power items.The disturbance estimation is compensated into the nominal controller in the design,which realizes high-precision,fixed-time stability of control object under external disturbance,and guarantees that the signals in the closed-loop system can converge to stability within fixed time.Finally,the effectiveness of the proposed control scheme is verified through comparative simulations.

As for aircraft remote sensing images with diversified scales and dense targets,the detection accuracy is relatively low,and the model is complex,which is not easy to deploy.To solve the problems,a remote sensing aircraft detection model based on the improved YOLOv4 is proposed.The K-means++ algorithm is adopted to optimize the anchor frame of target samples to improve size matching between the prior frame and the target,which reduces the missed detection rate.Focal loss is introduced in the loss function to reduce the weight of simple negative samples in the training process.Convolution kernel pruning and inter-layer pruning are fused for sparse training of the convolution kernel and batch normalized BN layer,which simplifies network structure and reduces the amount of parameters.Through experiments,the improved YOLOv4 foreign object detection algorithm has an mAP of 92.23% on the public UCAS-AOD and RSOD remote sensing data sets,and the detection speed is increased to 130.24 frame/s,which is conducive to rapid detection of aircraft targets in remote sensing images in actual industrial scenarios.

In order to improve the coverage efficiency and adaptability of multi-robot collaboration search in complex environments,a multi-robot path planning strategy for collaborative full-coverage search is proposed.Firstly,the Cooperative Coevolving Particle Swarm Optimization (CCPSO2)is used for sensor deployment in the target area.Secondly,the improved K-means method is used to cluster sensor deployment points,so as to achieve effective task area division.Finally,the deployed sensor location is taken as the waypoint to solve the Traveling Salesman Problem (TSP),and the enclosed path of each robot is obtained,so as to realize collaborative full-coverage search.The experimental results show that the proposed method can obtain a more evenly-distributed coverage path for each robot while ensuring good obstacle avoidance and minimizing the coverage period,which realizes effective collaborative full-coverage search of multiple robots,and can effectively adapt to the external complex environments with good robustness.

Aiming at the problems of small size,large number of targets and complex background in UAV aerial images,a UAV aerial target detection algorithm based on improved YOLOv4-tiny is proposed.On the basis of the original network,the algorithm expands the scope of detection scale,improves the matching degree for targets of different sizes,and fuses deep semantic information with shallow semantic information from bottom to top to enrich the feature information of small targets.Meanwhile,the attention mechanism module is introduced to conduct secondary screening of the feature information of the region of interest at each scale behind the backbone network.So as to filter the redundant feature information,and retain the key feature information.Compared with that of the original network,the average accuracy of the proposed algorithm is improved by 5.09% on the basis of real-time performance,and the experimental results show that the proposed algorithm has good comprehensive performance.

Velocity-loop control is widely used in the electro-optical tracking system for motor driving.Good velocity-loop response is helpful to improve the tracking accuracy of the electro-optical system.The motor of the system is driven by the Pulse Width Modulation (PWM) voltage outputted by the inverter.As the demand for accuracy and electro-magnetic compatibility of the electro-optical tracking system is increasing, inverter output filter must be used to suppress the electro-magnetic noises caused by PWM voltage.Traditionally,the design of the filter enhances electro-magnetic noise suppression based on the improvement of insertion loss,but the analysis of the impact of the filter on motor control is inadequate.To solve the problem,the influence of LC filter on the motor velocity-loop control is studied.The plant model of the motor with an LC filter is established,and the transfer function of the plant is given.A new design method of LC filter is proposed,which can reduce the influence on motor velocity-loop control as well as maintaining the insertion loss.Experiments on the velocity-loop control of electro-optical pod are conducted to verify the effectiveness of the proposed method.