Abstract:To address the conflict between the complexity of aerospace tasks and the design of traditional fixed-configuration satellites, aerospace institutions are focusing on the study of modular self-reconfigurable satellites with flexible configuration change capabilities. Among these efforts, configuration planning emerges as a particularly challenging area of research. Aiming at the configuration problem of modular satellites, cube-lattice satellites are taken as the research object, and based on graph theory, a configuration matrix and an extension matrix are proposed to describe the satellite topology. Through the study of the motion characteristics of the satellite module, an algorithm for solving the accessible space of the module motion is given. Considering the satellite configuration problem as a sequential decision-making problem, and based on the theory of deep reinforcement learning, the modification process is modeled as a Markov decision process. An intelligent modification planning method based on the actor-critic model is designed, incorporating a multi-layer neural network to approximate the actor and critic functions. Through training the neural network, the performance of the satellite reconfiguration strategy is progressively improved. The simulation experimental results show that the proposed configuration method yields progressively improved satellite reconfiguration strategies for the given satellite case studies. This approach exhibits generality across different satellite configurations with varying numbers of modules. Additionally, compared with the traditional configuration method based on heuristic search, it has advantages in the number of configuration steps, calculation time and configuration success rate, which validates that the proposed intelligent planning method has potential value in future modular satellite design work.

Abstract:To improve the aerodynamic prediction capability during fighter jet spin maneuvers and enhance the simulation accuracy of stable spin motion, a novel neural network model is proposed, leveraging the powerful function approximation capabilities of deep neural networks. This model enables accurate modeling of the unsteady aerodynamic forces during spin maneuvers and achieves high-precision spin attitude prediction through spin-coupled simulation. Focusing on the aerodynamic characteristics of fighter jets in post-stall spin, this study first utilizes the neural network model to achieve high-precision modeling of the unsteady aerodynamic moments observed in vertical wind tunnel tests. Secondly, based on the features of the neural network model and traditional aerodynamic database construction methods, an incremental superposition neural network model is proposed. This model incorporates control surface deflection increments from aerodynamic databases into the neural network, enabling high-precision modeling of unsteady aerodynamic moments under varying control surface configurations. Finally, the resulting model is then coupled with the spin motion equations to conduct stable spin simulations and spin characteristic predictions. The research results indicate that the proposed model effectively captures variations in spin aerodynamics under different control surface combinations. Compared to traditional aerodynamic databases, the aerodynamic moment prediction error is reduced by 77%. Using this model enables high-precision prediction of stable spin characteristics, with the relative error in stable spin period prediction reduced by 34%, demonstrating the engineering effectiveness of machine learning methods in simulating complex aircraft dynamics.

Abstract:Regarding range-extended missiles by pulse motor, in order to meet the needs of deciding optimal time for pulse engine ignition and the real-time generation of optimal overload commands, a nonlinear optimal guidance method is studied and a real-time generation method of optimal guidance commands is proposed in this research. First, a nonlinear optimal control problem model for the missile is established. Optimality conditions are then derived by fully differentiating the augmented objective function. Subsequently, a parameterized method which constructs a set of parameterized differential equations based on the optimality condition is proposed for the fast generation of pulse optimal trajectory datasets, allowing for generating datasets of optimal trajectories through numerical integration. Finally, based on the datasets containing the pulse engine optimal ignition timing and overload commands, feedforward neural networks are trained to decide the pulse engine optimal ignition time and generate the optimal overload commands in real time. Numerical simulations demonstrate that the proposed method can decide the pulse optimal ignition timing and generates the optimal overload commands within 1 ms. Moreover, the range of missile trajectory achieved is either superior to or comparable to that obtained through conventional optimization methods. Thus, this method has the capability of generating optimal guidance commands for range-extended missiles by pulse motor.

Abstract:To further improve the rate of trajectory convergence, terminal miss distance and miss distance rate are defined as performance index for guidance law derivation based on linear quadratic differential game theory. The derivation results realize the control purpose of minimizing miss distance while maximizing miss distance convergence rate. The control systems of intercept missile and its target are modeled in a general sense, applicable to forms where both exhibit higher-order dynamic characteristics. The derivated guidance law has a wide general adaptability. The proposed guidance law is extended to the typical situation of missile and its target with one-order control system dynamics. A comparative analysis of nonlinear system simulations has been performed for proportional navigation, conventional differential game guidance law, and the dual-weighted differential game guidance law proposed in this study. The simulation scenarios include three types of maneuvers: constant-maneuvering, S-type maneuvering and random starting maneuvering of target. The single shot kill probability is used for guidance performance evaluation. The simulation results show the advantages of fast trajectory convergence rate and low acceleration requirements of the proposed guidance law. The proposed guidance law minimizes the miss distance while maximizing the trajectory convergence rate of the miss distance, achieveing the control purpose of fast trajectory convergence of interception missile.

Abstract:To reveal the inherent relationship between the closed-loop stability of delayed vehicle platoons and communication topologies, a universal method of searching the most exigent eigenvalue (MEE) is proposed for second-order vehicle platoons under arbitrary communication topologies. First, for the case where all eigenvalues of the Laplacian matrix are real, by analyzing the monotonicity relationship between the maximum allowable delay and the eigenvalues, it is proven that the MEE of the second-order vehicle platoon must be the maximum eigenvalue of the Laplacian matrix. Then, for the case where some eigenvalues are complex conjugates, it is found that the maximum allowable delays corresponding to a pair of complex conjugate eigenvalues are equal in size, and an analytical expression for the maximum allowable delay is provided. Furthermore, during the analysis of the aforementioned monotonic relationship between the maximum allowable delay and the eigenvalues, the influence of the magnitude and phase of the complex conjugate eigenvalues on the MEE is revealed, and a concise MEE search rule is proposed. The results show that the proposed theoretical method is validated through simulation examples. Compared with traditional traversal methods of calculating the maximum allowable delay, the proposed method significantly reduces the computational burden. This method of searching the MEE is applicable to the closed-loop stability analysis of second-order vehicle platoons under arbitrary communication topologies and demonstrates universality.

Abstract:To improve the problems of poor real-time performance and inability to automatically adjust parameters of the existing anti-lock braking system (ABS) using proportional integral differential (PID) control method, a PID control method for anti-lock braking system based on multi-strategy aquila optimizer (AO) is proposed. Taking a single-wheel vehicle model as an example, firstly, the PID controller simulation model of the vehicle anti-lock braking system is constructed. Secondly, a differential evolution combining reverse learning combined with stagnation perturbation for aquila optimizer (DERLSP-AO) is proposed to overcome the limitations of the standard AO, particularly its tendency to converge to local optima and its limited search precision. By designing the reverse learning strategy of hunting perspective to increase the search range, the efficiency of the algorithm is improved. Additionally, a differential evolution strategy was integrated to evolve the aquila population by eliminating weaker individuals. By mixing multiple strategies, the DERLSP-AO method design is completed. Then, the optimal individual tuning PID parameters are used to obtain the optimized DERLSP-AO-PID controller. Finally, different road conditions are selected to simulate the anti-lock braking process of the vehicle. The results show that, compared to existing algorithms, the slip rate curve of the ABS output based on DERLSP-AO-PID control shows improved performance in maintaining the desired range. The vehicle exhibits reduced braking time and shorter stopping distances, which further validates the effectiveness of the improved algorithm and demonstrates enhanced braking performance.

Abstract:To enhance the reliability of multimedia information and mitigate the negative impact of image forgery events on society, there is an urgent need to develop image inpainting forensics to detect and locate tampered regions of images. This paper proposes a bridge-type attention forensics network (BAFNet) for image inpainting. The network receives tampered images directly and outputs the tampered regions end-to-end. The network adopts an encoder-decoder architecture as the basic framework. Firstly, the encoder selects two backbones, Swin Transformer and RepVGG, to extract multi-domain inpainting features. Then, a bridge-type attention module is used to connect the same-level stages of the two backbones, enhancing the encoder’s modeling capability in both local and global dimensions. Finally, a semantic alignment fusion module is built between the encoder and the decoder to eliminate semantic inconsistencies between the features extracted by the two backbones, thereby improving the forensic performance of the network. Experimental results on different inpainting forensic datasets demonstrate that the proposed model, compared with other mainstream forensic models, can more accurately locate the inpainting areas. In particular, on the challenging DeepFillV2 dataset and Diffusion dataset, the proposed BAFNet achieves IoU scores of 91.37% and 82.34%, respectively, which improves the IoU metrics by 8.77% and 10.46% compared to the mainstream forensic network MVSS-Net. In addition, combining the results of several experiments, BAFNet achieves a good balance between forensic performance and model complexity.

Abstract:To improve the electromagnetic performance of the built-in U-type permanent magnet synchronous motor and suppress the low-frequency vibration and noise of the motor, a motor topology is proposed with a U-type pole structure combined with the Halbach magnetizing method and rotor slotting. Firstly, the relevant electromagnetic performance expressions of the motor are derived, and the low-frequency vibration and noise fluctuation composition of the radial magnetic flux density and radial electromagnetic force density between the stator and rotor are analyzed. Next, using analytical and finite element methods, the spatial and temporal distribution of the radial electromagnetic force in the air gap is obtained. Parameter sensitivity analysis, parameter scanning method, and response surface method are used to multi-objective optimization of the selected rotor topology parameters to obtain the optimal parameter solutions. Finally, no-load reaction potential, radial flux density, cogging torque, output torque, torque pulsation, radial electromagnetic force density, vibration acceleration and sound pressure level of the U-type pole structure are compared and analyzed with four other U-type pole structures. The U-type motor structure is validated by combining electromagnetic, mechanical and acoustic fields analyses. The results show that by adding a type I magnetic pole to the U-type magnetic pole structure, changing the magnetizing method, and modulating the sinusoidal degree of the air gap magnetic field by dq-axis slotting, the motor’s cogging torque decreased by 91.3%, the no-load reverse potential is increased by 53 V, the output torque is increased by 39.6%, the amplitude of radial electromagnetic force waves in the air gap decreased at all harmonic orders, the vibration acceleration of the stator assembly decreased significantly at all frequencies, and the maximum sound pressure level around the motor decreased by 9 dB at the peak, with a reduction rate of 10.1%.

Abstract:To evaluate the dynamic risks of industrial cyber-physical saystems (ICPS) under cyber attacks, this study investigates a Markov-improved evolutionary game model. Based on the vulnerability nodes within the ICPS, a system attack-defense state transition diagram from the information domain to the physical domain is designed, providing a foundation for the Markov-improved evolutionary game analysis. First, in the single-stage attack-defense process, an evolutionary game model incorporating a parameter mechanism is studied to determine the payoffs of attack and defense entities with varying degrees of rationality and exploration after the game. Second, in the multi-stage attack-defense process, based on the single-stage attack-defense game model, transition probabilities and discount factors are introduced. The attack payoffs of different vulnerability nodes are calculated according to the attack-defense state transition diagram, enabling dynamic deduction of multi-stage attack-defense confrontations. Finally, the dynamic risks of ICPS are assessed based on the magnitude of attack payoffs. This study conducts numerical experiments and simulations of an industrial cyber-physical system model, using a boiling water power plant as the simulation object. The Markov-improved evolutionary game evaluation method is simulated using Matlab, and the dynamic risks of ICPS are evaluated based on the attack payoffs. The results demonstrate that the proposed model emphasizes the differences between the attack and defense sides, reasonably calculates the attacker’s payoffs in ICPS based on the varying levels of rationality and exploration of both parties. This provides a theoretical foundation for the dynamic risk assessment of ICPS under cyber attacks and offers significant reference value for enhancing the security of industrial cyber-physical systems.

Abstract:To enhance the accuracy of cross-modal fusion and interaction in visual question answering (VQA) while mitigating the loss of multimodal feature information, we propose a novel cross-modal adaptive feature fusion approach for VQA. First, the method designs a convolutional self-attention unit consisting of self-attention layers and dilated convolution layers-the former captures global feature information while the latter extracts spatial relationships between visual objects. Subsequently, an adaptive feature fusion layer effectively integrates global relationships with spatial correlations, enabling the model to simultaneously consider both global contextual information and inter-object spatial relationships during image feature processing, thereby addressing the limitation of traditional attention mechanisms in overlooking spatial relationships. Furthermore, based on the varying contributions of different modal features to answer prediction, we construct a multimodal gated fusion module that adaptively combines features according to their relative importance, effectively reducing information loss across modalities without introducing additional computational overhead. Experimental results demonstrate that our method achieves overall accuracies of 71.58%, 72.00%, and 58.14% on the VQA2.0 test-dev, test-std, and GQA datasets respectively, significantly outperforming traditional self-attention approaches without requiring additional pre-training datasets. This research provides valuable insights and serves as an important reference for cross-modal feature fusion studies.

Abstract:To explore the differences in structural and aerodynamic characteristics of the variable-sweep wings on aircraft under various swept deformation methods, as well as the underlying physical mechanisms of these differences, this paper proposes two shearing variable-sweep schemes based on parallelogram unit shear deformation topology, and a comparative study was conducted with a conventional rotary variable-sweep wing. First, the structural characteristics of the three deformation modes are investigated by four main parameters: wing area, chord ratio, root-to-apex ratio and relative thickness of the wing. Then, numerical simulations of the winding flow field in a wide range of speeds are carried out to analyze the aerodynamic characteristics and mechanisms under the three deformation modes. Finally, for the diagonal shear variable-sweep wing with optimal aerodynamic performance, the length-width ratio of the parallelogram unit is optimized with the optimization objective functions of wing area, chord ratio and root chord length at supersonic cruise state. A deformable prototype is developed for wind tunnel testing. The results demonstrate that, across a wide range of speed, diagonal shear variable-sweep can obtain a better lift-to-drag ratio, the difference mainly arises from the fact that the airfoil of the wingtip section of the diagonal shear variable-sweep is intact and the relative thickness of the wing is smaller. When the aspect ratio of the parallelogram unit is 1.75, the comprehensive aerodynamic performance of the morphing wing is the best.

Abstract:To solve the problem of feature localization with multi-feature intersection and improve the performance of machining feature recognition for complex parts, this paper proposes Brep3pNet, a machining feature recognition method under the framework of instance segmentation. Firstly, based on the boundary representation (B-rep) of 3D models, Brep3pNet extracts geometric and topological data such as face point clouds and face adjacency graphs to construct a graph representation of the 3D model. We utilize point cloud learning networks and graph neural networks to learn surface-level embedding representations of the 3D model. Secondly, a probabilistic positional embedding method is proposed, which introduces spatial position prior information to encode the face into ternary Gaussian distribution, and measures the similarity among those face embeddings by Bhattacharyya kernel for the purpose of locating machining features and generating candidate machining feature instances. Finally, a score network is designed to predict the quality of the instance generated, so as to guide the non-maximum suppression between instances to remove redundant feature instances, thereby obtaining the final machining features. Brep3pNet is evaluated on four multi-feature datasets, including MFCAD, MFCAD++, MFInstSeg and a synthetic dataset of rotary parts. The research results indicate that Brep3pNet outperforms other state-of-the-art methods on feature localization accuracy, and can achieve optimal feature recognition accuracy with lightweight model parameters, demonstrating its potential application in intersecting features recognition.

Abstract:Binderless tungsten carbide (BTC) is an ideal material for high-temperature resistant molds, but the high hardness and low toughness characteristics lead to a poor machinability, making efficient and precise machining difficult with existing cutting methods. In order to improve the machinability of binderless tungsten carbide and achieve high-efficiency and high-quality machining of binderless tungsten carbide, a laser-oxidation-assisted micro-milling process for tungsten carbide was proposed in this study. The oxidation ablation experiments were carried out on the surface of BTC using a 1 065 nm fiber continuous laser. The effects of different laser power levels, scanning speeds, and the number of scans on the morphology of the ablation grooves were studied to analyze the oxidation mechanism of BTC. Then micro milling experiments on the oxidation grooves were carried out. Meanwhile, a control group without laser-induced oxidation process was set up for comparison with the micro milling experiments. The advantages of the laser-oxidation-assisted micro-milling process in the machining of high-hard, brittle tungsten carbide were explored. The results show that the surface of BTC showed obvious oxidation and ablation traces when the laser power was greater than 7 W. Higher power levels and the slower scanning speed lead to more intense oxidation ablation. Under the high-temperature effects of the laser, thermal cracks generated at the bottom of the grooves, and the length of the thermal cracks was reduced by multiple laser scans. The tungsten carbide grains were oxidized at high temperatures, and the oxidation product was mainly the loose WO₃. The laser oxidation process can lower the tool wear and improve the machinability of BTC.

Abstract:To study the temperature characteristics of MEMS electrothermal actuators at micro-scale, taking into account the influence of air on the dynamic heat balance of the actuator, an electrical-thermal-fluid-solid coupled multi-field model based on air convective heat transfer is proposed. This model is based on the principles of energy conservation along with theories of gas convection and heat conduction. Finite element simulations are conducted to analyze the model. An experimental platform for temperature characterization of MEMS electrothermal actuator is established, and the experimental results of the temperature response of the actuator under constant voltage excitation are compared and analyzed against the simulation results from both the electric-thermal-fluid-solid coupling model and the traditional heat transfer model. The results indicate that compared with the model based on the constant convective heat transfer coefficient and empirical equations, the electric-thermal-fluid-solid coupling model achieves higher accuracy, with the steady-state temperature distribution error ranging between 0.8% and 7.6%. Additionally, the convective heat transfer coefficient varies across different surfaces of the actuator. Specifically it decreases then increases on the upper surface, increases steadily on the lower surface, and increases then decreases on the vertical walls. Despite these variations, the convective heat transfer coefficients on all three characteristic surfaces reach the steady-state almost simultaneously, coinciding with the actuator’s temperature reaching steady state. These findings, based on the uneven characteristics of air convective heat transfer, the obtained temperature characteristics of the electrothermal actuator provide a foundation for the application of MEMS electrothermal actuators in microelectromechanical systems.

Abstract:To fulfill the practical requirements of constructing large-size structures in orbit for extraterrestrial exploration and space station trusses, a direct in-orbit strip-orming method was proposed, alongside the development of a small, lightweight and low-power metal pipe fitting manufacturing equipment. Based on Kirchhoff hypothesis and the principle of equivalent strain energy, a nonlinear mechanical model of bending and winding was established to analyze the influence of different parameters on the winding torque. The ABAQUS/explicit solver was used to solve the winding torque and validate the effectiveness of the analytical model. The winding and locking performances were described and optimized by defining three performance indexes, namely stable winding moment, positive locking pressure, and maximum locking edge stress in a quantitative manner. The lower the indexes, the higher the winding locking performance. The polynomial surrogate model of the index parameters for the curved winding pipe was established using the response surface method, and the metal pipe bending was sujected to multi-objective optimization design through an improved non-dominated genetic algorithm (NSGA-II). The optimization results demonstrate a reduction of 26.23% in the winding torque, a decrease of 4.71% in positive pressure at the locking edge, a decline of 2.14% in maximum stress at the locking edge, and an enhancement in the fluctuation of the torque curve. A prototype for winding locking was developed, featuring a roller seat with a diameter of 70 mm, a roller group spacing of 100 mm, and a locking groove bending box with a corrective adjustment function, along with a core shaft diameter of 50 mm. The experimental results demonstrate the impact of various process parameters on rolling forming and elucidate the relationship between forming angle and pipe fitting diameter during spiral forming. The findings of this study offer a crucial theoretical and experimental foundation for the implementation of metal tube forming in orbit.

Abstract:Aiming at the problem of outliers, noise and irregular disturbances prevailing in the monitoring parameters of the thermal system, a noise reduction method for monitoring parameter of the thermal system based on median regression empirical mode decomposition (MREMD) and variational mode decomposition (VMD) is proposed. The purpose is to enhance the accuracy of monitoring system regulation and the level of system operation management, while minimizing noise and disturbances in the monitoring parameters, all while preserving as much of the original data’s effective information as possible. The method firstly performs MREMD of the monitoring parameters to obtain a number of intrinsic mode functions (IMF). Secondly, chaotic time series analysis is applied to filter out the IMF components containing noise using permutation entropy, reconstructing them as the noise portion of the original data. Then the noise part is decomposed by VMD, and the optimal envelope entropy of the IMF obtained by the decomposition is used as the fitness function. The northern goshawk optimization (NGO) algorithm is used to optimize the VMD decomposition parameters, yielding the IMF with the lowest envelope entropy within the optimization range, which contains the effective information of the noise portion. Finally, this part was reconstructed by summing with the low frequency IMF component and residual component obtained by MREMD decomposition which both are contained trend information, to obtain the monitoring signal after noise reduction. The results demonstrate that through case studies, the modal double-decomposition noise reduction method proposed in this paper has highest signal-to-noise ratio and lower information entropy and power spectral entropy compared to mainstream wavelet threshold denoising methods and moving average filtering techniques.