A robust moving horizon estimation algorithm is proposed for the state estimation of multi-agent systems under deceptive attacks and uncertain parameter constraints. Firstly,the deceptive attacks are characterized as a stochastic sequence satisfying the Bernoulli distribution,and a predictive compensation strategy is adopted. Secondly,the time-varying weight matrix in the distributed moving horizon estimation algorithm is modified to counteract the influence of uncertain parameters. Based on this,a robust moving horizon estimator is obtained by solving an min-max optimization problem. Subsequently,conclusions regarding the convergence of estimation error are given. Finally,simulation results of a multi-agent system composed of vehicles verify the feasibility of the proposed algorithm.
Considering the problem of Field-of-View (FOV) constraint in the precision guidance mission,an Impact Time Control Guidance (ITCG) law based on look-angle shaping is proposed. Firstly,a cubic polynomial reference look-angle profile with respect to the non-dimensional relative range is constructed. The impact time and maximum look-angle are expressed as equations of the profile parameter,and the reference look-angle profile satisfying the constraints is obtained by solving the parameter. Based on the dynamic model of the missile,the guidance command of tracking the reference look-angle profile is designed,the ITCG law with FOV constraint is obtained,and the achievable range of the impact time is derived. The proposed guidance law does not depend on the time-to-go estimation,linearization model and switching logic,and the terminal zero control can be achieved. Finally,the digital simulation results verify the effectiveness of the proposed guidance law.
In order to address the multi-layer hesitance phenomenon in multi-attribute decision-making problem,the type 2 hesitant dual fuzzy set is proposed. The score function of Type 2 Hesitant Dual Fuzzy Elements(T2HDFE) is defined,and the axiomatic definitions of its distance,correlation coefficient and entropy are discussed. A distance formula of type 2 hesitant dual fuzzy elements without data extension is constructed,and based on the distance formula,the formulae for the correlation coefficient and entropy of type 2 hesitant dual fuzzy elements are given. Then,their properties are proved. Finally,a multi-attribute decision-making method based on correlation coefficient and entropy of type 2 hesitant dual fuzzy elements is proposed for the multi-attribute decision-making problem under type 2 hesitant dual fuzzy environment,and a numerical analysis is discussed through the selection of strategic delivery plans.
Aiming at the problems of fast movement,occlusion,non-rigid deformation and illumination change of objects in complex scenes,an object tracking algorithm based on 3D attention and pyramid decoder is proposed. Firstly,VGG-16 neural network is introduced and its structure is optimized to improve the efficiency and quality of feature extraction. Secondly,by introducing 3D attention,the extraction ability of key features is enhanced. Then,the deep semantic fusion module is used to fuse feature information through upsampling to achieve accurate expression of features. Finally,a pyramid decoder is designed to improve the robustness of the model in complex scenes. Experimental results show that the success rate and tracking accuracy on OTB100 data set are improved by 15.8% and 16.2% respectively compared with those of the baseline algorithms.
To address the issues of insufficient path search capability and local optimum in swarm intelligence algorithms for UAV path planning in battlefield environments,a UAV low-altitude penetration path planning method based on Undirected Random exploration Dung Beetle Optimization (UR-DBO) algorithm is proposed. Firstly,terrain and threat models are established. Then,population diversity of the algorithm is enhanced by using Piecewise map in initialization. An undirected random exploration mechanism is proposed,which aims to make up for the defect of incomplete exploration of the ball dung beetles in DBO algorithm and improve the global optimization ability of the algorithm. Subsequently,the stealing dung beetle can learn from the multi-strategy optimization mechanism of Ali Baba and the Forty Thieves (AFT) algorithm,so that it can dynamically adjust its strategy according to the problem,which is beneficial to the algorithm to jump out of local optimization. Finally,six test functions and two kinds of different terrains are selected for test. The experimental results show that the UR-DBO algorithm has better convergence speed and accuracy than the contrast algorithms,and is more suitable for UAV three-dimensional path planning.
The significant size difference of targets and the complex background in remote sensing images may lead to false detection and missed detection. Meanwhile,existing algorithms have problems of large parameter number and high computational cost. Therefore,a lightweight remote sensing target detection algorithm based on YOLOv8n is proposed. Firstly,the C2f module of the neck is replaced with the CSPStage module to enhance the learning of features in different feature layers by introducing the integration mechanism of gradient change. Secondly,CG module is introduced to reconstruct a Bottleneck module in C2f,and the feature processing ability of the network is enhanced by combining context information. Then,based on PConv,lightweight detection header PDetect is designed to reduce the waste of redundant information on computing resources. Finally,the Focaler-Shape-IoU loss function is designed to make the model focus on the shape and scale factors of the border itself,make up for the influence of sample difficulty distribution on the border regression,and improve the convergence speed and generalization performance of the model. The experimental results show that the mAP value obtained by the improved network model on the open remote sensing dataset NWPU VHR-10 is 4.1 percentage points higher than that obtained by the original YOLOv8n algorithm,the parameters is reduced by 37%,and the FLOPs is reduced by 45%,which proves the effectiveness and advanced nature of the improved algorithm.
To address the issues of data theft and unauthorized access during the transmission of aerial images,a novel aerial image encryption algorithm based on five-dimensional hyperchaotic system and immune algorithm is proposed. Firstly,the design of the five-dimensional hyperchaotic system overcomes the limitations of the insufficient key space in low-dimensional chaotic systems. The chaotic properties of the system are verified through equilibrium point analysis and other methods. Subsequently,the chaotic sequence is used to perform the initial scrambling of the aerial image,followed by a second scrambling by using the immune algorithm. Finally,chaotic diffusion is applied to complete the encryption process. The algorithm has a key space of 2185, a ciphertext information entropy of 7.999 8,a correlation close to zero,a pixel change rate of 99.608 6%,and an unified average change intensity of 33.471 3%,which can effectively resist outside attacks. The experimental results show that the algorithm has excellent security and robustness.
Aiming at the problems of low object detection accuracy in remote sensing images caused by special angle of view,complex background information,dense objects,and diverse scales,an object detection algorithm for remote sensing images based on multi-scale feature fusion,EMD-YOLOv8 is proposed. Firstly,an enhanced backbone network,EnhancedDarkNet is introduced to realize shallow feature reuse and enhance the extraction capability of object details and texture features. Secondly,a Multi-scale Interactive Fusion Feature Pyramid Network (MIFFPN) is designed to reconstruct the neck structure and strengthen feature fusion in multi-scale spaces. Finally,a Dimension Interactive Polarization Attention (DIPA) mechanism is proposed to reduce background noise and interference of redundant information,enhancing the response of key object features. The proposed algorithm achieves an mAP@0.5 and mAP@0.5∶0.95 of 85.9% and 62.5% on the remote sensing dataset DIOR,which is 5.1 and 5.6 percentage points higher than that of the original YOLOv8n respectively,while the detection rate reaches 90.9 frames per second. Experimental results demonstrate that EMD-YOLOv8 can improve the accuracy of object detection in remote sensing images and meet the requirements of real-time detection performance.
Aiming at the issue of identifying aerial sensitive targets in the context of incomplete trajectories,a rapid similarity matching algorithm using the K-Dimensional Tree (KD Tree) is proposed to extract civil aviation routes based on the relatively fixed nature of civil aviation routes,and civil aviation targets are removed based on route matching feature identification method,thus a dual-module approach for screening aerial sensitive targets is constructed. The traditional density clustering method is used to preprocess partial complete routes,and a nearest neighbor point rapid search algorithm based on KD Tree is designed. Combining with the constraint of trajectory start and end points,a logical discrimination of similar trajectories is conducted,so as to achieve rapid aggregation of routes with incomplete trajectories. Based on the position and motion characteristics,the civil aviation targets are confirmed and removed,which solves the limitation that the existing identification methods rely heavily on the completeness of historical data. The experimental results indicate that compared with traditional methods,the proposed method improves trajectory similarity matching accuracy by at least 24.63 percentage points,and significantly reduces time overhead. It can rapidly and accurately extract civil aviation routes while excluding civil aviation targets,providing a new perspective for aerial intelligence analysis.
How to deploy multistatic sonar to achieve the maximum coverage of the monitoring area is an important issue in multistatic sonar application. Based on the bistatic detection model,the detection range of bistatic sonar is approximated as a Cassini oval,a multi-objective optimization model for the coverage problem of multistatic sonar is established,and the monitoring area is discretized into a deployable grid to refine the multi-objective optimization model of receiving nodes. Nondominated Sorting Genetic Algorithm II (NSGA2) is improved by removing the limitation of the number of nodes and increasing the reconfiguration operation of nodes. Simulation results show that the Improved NSGA2 (INSGA2) has faster convergence speed and better global search ability than the traditional NSGA2,and obtains a better Pareto front,which provide a reference for the optimal deployment of multistatic sonar.
In order to suppress the drift of inertial navigation system and improve the seamless navigation ability of Inertial Navigation System/Magnetism Navigation System (INS/MNS) in geomagnetic lockout environment,a hybrid seamless INS/MNS strategy combining Adaptive Cubature Kalman Filter with Deep Self-Learning (DSL-Adaptive-CKF) is proposed. It mainly includes two innovative steps. Firstly,an adaptive optimization auxiliary method based on residuals and innovation is combined to enhance the robustness of the initial error of measurement noise and process noise. The heading RMSE of the Adaptive-CKF method is 2.78°,which improves the heading accuracy by 89.51% compared with that of the traditional single INS,and greatly improves the robustness of the combined navigation system and the accuracy of the heading measurement. Secondly,by introducing the Nonlinear Autoregressive model with Exogenous inputs (NARX) neural network,the Adaptive-CKF can learn deeply,which means that it can realize continuous high-precision navigation estimation even during lockout period,and its heading RMSE reaches 3.08°,thus the heading accuray is improved by 88.38% compared with that of the single INS.
A Robust Weighted Likelihood Constant False Alarm Rate detector algorithm based on Bayesian interference control,BRWL-CFAR,is introduced,which dynamically evaluates the clutter level in Weibull background by equally dividing the clutter range profile and optimizing the decision selection. At the same time,the clutter level carries out feedback control on the former. Then,in order to realize multi-target predictive inference in Weibull background,based on Bayesian interference control theory,the idea of using constant false alarm detector in different scenarios by Bayesian classified interference control is proposed. Therefore,the anti-interference ability of the detector is improved while reducing the computational complexity. The false alarm rate and judgment expression are given,and the detection is extended when the segmentation and interference are arbitrary. The SAR image data obtained by TerraSAR-X satellite is used for simulation experiments. The results show that the proposed detector is more robust than the traditional detector algorithms of CA-CFAR,OS-CFAR,TM-CFAR and WAI-CFAR.
In response to the difficulty of obtaining paired datasets in existing underwater image enhancement methods,an unsupervised underwater image enhancement method based on data-driven optimization and atmospheric light adjustment is proposed. Firstly,the underwater image degradation module is designed,and the latent components of the original image are extracted by parallel network,and the re-degraded image is generated by using the improved Koschmieder model. Then,a data-driven optimization module is designed to compare the original image with the re-degraded image and extract the perceptual transmission of different regions. It adopts a double-branch gating unit and a channel attention mechanism to selectively adjust the input features and realize cross-channel feature fusion. Finally,an atmospheric light optimization module is used to process the global information in frequency domain to offset the influence of atmospheric light on underwater images. The experimental results show that the evaluation indexes of UCIQE,UIQM,NIQE and CIEDE of the proposed method on LSUI dataset,UIEB dataset and RUIE dataset are improved by an average of 4.4%,2.7%,1.6% and 27.5% respectively compared with the suboptimal solution,and it also has obvious advantages in visual quality.
Aiming at the problem of low detection rate of infrared small target in complex background,an target detection algorithm based on Improved High Boost Filtering (IHBF) and weighted local contrast is proposed. Firstly,the infrared image is processed by IHBF operation to suppress most background clutter and extract candidate pixels of the target. Then,the local contrast is calculated by using the gray difference ratio between the target region and the background region. At the same time,according to the dissimilarity between the target and the background,a weighted function is designed to further improve the contrast between the target and the background. Finally,the target is extracted by adaptive threshold segmentation. Experimental results show that the proposed algorithm has excellent detection performance in a variety of complex scenes.
In traditional methods,acquiring the helicopter pilot's driving status requires the use of wearable devices,which shall affect the driving. Therefore,a helicopter driving state detection method is proposed based on keypoint detection. The helicopter driving state is decomposed into human posture and the positional information of control components,and a self-built helicopter driving data set is used for training designed models. A method of determining the state of control components via coordinate regression of the components and a method of building pilot posture model through coordinate regression of the pilot posture is realized respectively. Test results indicate that the method achieves an average recognition accuracy of 0.923 1 for pilot driving actions and 0.986 1 for the position of control components,thereby validating the feasibility of establishing a pilot posture model through coordinate regression.
Communication system and audio system are important systems of aircraft. When testing the system,the voltage deviation of the input signal is carried out to meet the test requirements of special scenarios. Therefore,a signal generation circuit based on multilevel inverter is proposed by improving the class D power amplifier. The structure and operating modes of new topology are analyzed in detail,and the feasibility of DC bias,frequency and amplitude adjustment is demonstrated. According to this topology,different control modes are proposed,and the realization methods of three-level SPWM modulation and five-level SPWM modulation are explained in detail. The correctness of the control strategy and theoretical analysis is verified by PSIM simulation. Finally,the correctness of theoretical analysis and the feasibility of new topology are demonstrated by setting up an experimental platform,and the method for reducing the high-frequency oscillation at the output end of the bridge arm is given.
An adaptive neural network prescribed performance control strategy is proposed based on Hamilton-Jacobi Inequality (HJI) theory with the background of manipulator control. Firstly,the nonlinear transformation with a prescribed performance function is utilized to convert the tracking error to an unconstrained form,thereby allowing the trajectory tracking error to converge into a prespecified range at a user-specified convergence rate. Secondly,the backstepping technique is used to design the virtual control law based on the unconstrained tracking error. Further,according to the universal approximation characteristics of neural network,RBF neural network is used to approximate the model uncertainty. Finally,a novel prescribed performance control method is designed based on the HJI theory and the approximation provided by the RBF neural network. The Lyapunov function proves the stability of the trajectory-tracking closed-loop system in this paper,and the effectiveness of the proposed control method is verified in the simulation of a two-joint manipulator.
Aiming at the complexity of state detection and fault diagnosis of spacecraft power system,a non-invasive real-time monitoring method of power electronic converter is proposed. Firstly,the digital model of the physical model is built. Secondly,the output voltage collected from the physical model is input into the digital model,and the objective function is constructed according to the output voltage and ripple of the physical model and the digital model. The component parameters of the digital model are updated by using the Particle Swarm Optimization (PSO) algorithm,and the fault diagnosis of the physical model is realized. The experimental results show that the proposed method has short time for fault diagnosis and high average accuracy of parameter identification,which can realize real-time monitoring of converter.