Advances in Astronautics Science and Technology最新文献

The online prediction of reentry footprints is critical for autonomous systems in scenarios like emergency landing and mission replanning, yet it remains challenging to balance computational speed with predictive accuracy. This work presents a fast and accurate online prediction method based on a Multi-Head Attention Neural Network (MHANN) to overcome the limitations of traditional numerical and analytical approaches. The proposed model is trained on high-fidelity samples generated offline using the Gauss Pseudospectral Method (GPM). To handle the periodicity of longitude, we introduce a state expansion technique that represents longitude as continuous sine and cosine features, effectively eliminating numerical discontinuities. The network employs a lightweight Multi-Head Attention mechanism to capture complex dependencies between flight states and footprint boundaries efficiently. Furthermore, a custom loss function incorporating a trigonometric identity constraint is designed to ensure the physical consistency of predictions. Simulation results demonstrate the model’s superiority, achieving a training loss at least one order of magnitude lower than control groups. With only 86,969 parameters, the MHANN accomplishes an average inference time of 0.221ms on a platform with an Intel Core i9-14900HX CPU, reducing the prediction error by up to 88.6% compared to benchmarks. The combined use of state expansion and the physics-informed loss function is shown to be crucial, contributing to a dramatic 94% error reduction in ablation studies. This study confirms that the MHANN-based method delivers millisecond-level, high-accuracy footprint prediction, fulfilling the stringent real-time requirements of autonomous onboard systems.

In response to the demand for space debris mitigation, this paper discusses an adhesion mechanism based on van der Waals forces and a Stewart platform. This mechanism has high adhesion, high transmission accuracy and high reliability characteristics. The STEWART mechanism adapts to the relative attitude deviation between the service and the target, absorbs the kinetic energy of the collision, reduces the collision force, prolongs the contact time between the adhesion structure and the target surface, improves the impact of non-ideal contact caused by relative attitude deviation on adhesion performance. The system scheme including the mechanism configuration and composition, the adhesive structure and the adhesive material is introduced. In the third section, the collision adhesion dynamic model and the adsorption-desorption behavior model is introduced. The simulation and analysis of collision adhesion performance is detailed explained in fourth section. In the fifth section, through buffer performance testing, adhesion force testing, and air flotation platform experiments, the correctness of the simulation parameter calculation has been verified. The attitude adaptability of the final adhesion based on the non-cooperative target adhesion mechanism of the STEWART platform is determined by the mechanism design, and the adhesion force only plays a role in maintaining adhesion.

Folding-wing aircraft can modify their aerodynamic configurations during various flight phases, enabling adaptation to diverse atmospheric conditions, making them a crucial area in aerospace engineering advancements. The coordinated deformation of lightweight structures is a key technology for realizing folding wings. Mechanical metamaterials, with their lightweight and customizable deformation properties, are promising for use in folded wing structures. The KinetiX-based 3D-deformed metamaterial, with inclined ledges and flexible hinges, can produce out-of-plane bending deformation under in-plane loading. This study establishes a mechanical model of the KinetiX 3D-deformed metamaterial unit-cell structure, evaluates its deformation through simulations and experiments, and analyzes its bending behavior in combined configurations. The metamaterial is applied to wing and airfoil designs, with finite-element analysis and experimental tests validating its bending performance, demonstrating its potential for folded aerodynamic structures.

This paper proposes a novel trajectory planning method for multiple UAVs (Unmanned Aerial Vehicle, UAV) collaborative standoff tracking using a consistency GVF (Guiding Vector Field, GVF ) with an adaptive gain (AG-GVF), which is modulated by a modulation matrix for collision-free navigation in unknown complex environments. Firstly, the AG-GVF is established based on the desired path, with an adaptive gain and an additional virtual coordinate introduced to elevate dimensionality. The adaptive gain aims at reducing steady-state error and eliminating oscillations. This virtual coordinate is not only used for eliminating singular points but also utilized as a state variable for consistency control, ensuring uniform phase distribution among multiple UAVs during standoff tracking. Subsequently, in environments with obstacles, a modulation matrix is proposed to adjust the original GVF motion by estimating the normals of unknown obstacles using point clouds and constructing a modulation matrix to modify the AG-GVF direction for effective obstacle avoidance. Finally, the obtained desired path is optimized to generate flight trajectories that satisfy the kinematic constraints of fixed-wing UAVs. Simulation results demonstrate that the proposed method enables multiple UAVs to achieve collaborative standoff tracking with collision-free navigation in unknown complex environments.

With the rapid expansion of satellite constellation system scales, the design of constellation configurations is increasingly confronted with critical challenges including inter-satellite collision risks and mission area coverage effectiveness. Building upon the existing calculation method for minimum inter-satellite distance (MISD), this study further proposes a geometric analysis-based approach to solve the latitude corresponding to the MISD. By investigating the correlations among the MISD, its corresponding latitude, and constellation coverage requirements, a constraint model for constellation configurations oriented to dual coverage of target areas is established. Through systematic simulation analysis, the variation patterns of the MISD and its corresponding latitude with key constellation configuration parameters are revealed, and the distribution characteristics between the constellation’s dual-coverage target areas and the latitude of MISD are clarified. Based on the aforementioned calculation model and pattern models, the constellation configuration design process is streamlined, and an integrated design method tailored to application requirements and constellation safety is presented. The proposed method enables the automatic generation of a configuration recommendation list that satisfies both dual-coverage and safety constraints, thereby providing a reliable theoretical and technical basis for constellation configuration optimization.

With the increasing interest in cislunar space exploration, a variety of cislunar missions are being proposed. To support the implementation of diverse missions, it is necessary to establish different types of cislunar constellations. Although there is abundant research on cislunar constellation design, a comprehensive review is still lacking. This review addresses this gap by establishing a theoretical framework for the design and optimization of cislunar constellations through an examination of relevant literature. Firstly, common orbital types and their applications in cislunar constellations are summarized. Subsequently, various types of existing cislunar constellations are introduced, including those in the development phase and those still in the conceptual design phase. A unified constellation performance evaluation system is then established, consisting of two types of performance metrics: metrics associated to specific missions and general metrics applicable to all types of constellations. Following this, design strategies and results for cislunar constellations are collated and summarized. An analysis based on the research overview reveals the trends in cislunar constellation design and optimization. Finally, future prospects on cislunar constellation design and optimization are discussed.

Aluminum diboride (AlB₂) has been proposed as a viable substitute for elemental boron as the fuel for boron-containing solid propellants, owing to its favorable compatibility with propellant formulations, high gravimetric heat of combustion, and the anticipated synergistic effects from boron and aluminum. Nevertheless, its practical implementation is impeded by intrinsically low energy-release rates and pronounced agglomeration. In this study, two-dimensional graphene fluoride (GF), three-dimensional polytetrafluoroethylene (PTFE), and two-dimensional graphene oxide (GO) were strategically incorporated to construct composite fuel systems capable of fully exploiting the energetic potential of AlB₂. Laser-ignition experiments demonstrated that GF, GO, and PTFE all generate abundant gaseous products while markedly intensifying the combustion of AlB₂, while GF exhibits better enhancing effects. Thermal analyses reveal that fluorinated additives markedly enhance the thermal oxidation characteristics of AlB₂, effecting a pronounced reduction in its initial decomposition temperature. The difference in the combustion mechanism of AlB₂ with GF and GO lies in the fluorine, although they both open alternative reaction pathways, GF exhibits more notable capacity to suppress agglomeration of condensed combustion products. The formulation developed and mechanisms revealed by this work could potentially advance the applications of boron-based fuels in ramjets and scramjets.

The orbital inspection game (OIG) is characterized by its strong coupling with observational conditions and higher complexity compared to pursuit-evasion games. This study employs reinforcement learning techniques to investigate the OIG problem. Building upon the Markov decision process and its hybrid variant, a two-stage decision-making model for spacecraft is proposed, decomposing the OIG into approach and inspection tasks. With a specifically designed hierarchical network architecture and gradient-driven exploration strategy, the gradient-driven exploration-based serial solution method (GDE-SSM) is developed. GDE-SSM stratifies the agent’s exploration space through model switching and task decomposition, thereby significantly improving training effectiveness. Compared to the prediction-reward-detection multi-agent deep deterministic policy gradient algorithm and the deep deterministic policy gradient algorithm, the defender trained via GDE-SSM exhibits a tendency to adopt more aggressive maneuvering strategies, resulting in average success rate improvements of 57.8% and 21.8%, respectively. Generalization and robustness analyses demonstrate that GDE-SSM achieves excellent robustness against uncertainties arising from environmental parameters, suboptimal observation conditions, and abrupt disturbances.

The concept of spacecraft modularization has gradually attracted attention. Due to the problems of low efficiency and high transfer consumption in the current large-scale reconfiguration research of space modular self-reconfigurable satellites (SMSRS) in orbit, we propose a distributed self-reconfigurable motion planning optimization method. Firstly, the real-time minimum map and the corresponding undirected connected graph are established based on the connection condition and the limited observation constraint condition. Then, using the Dijkstra search algorithm, we create the real-time minimum map and the connected graph, leading to a new local optimal reconfiguration path algorithm that combines minimum map and shortest path planning, which significantly cuts down on the total calculations and the number of steps needed for reconfiguration in large-scale planning. Simulation results show that the new method is better than the traditional centralized graph-based method in the on-orbit reconfiguration mission of random configuration with 10–500 module scale, about 11.2% transfer steps and 35.7% reconfiguration planning time were saved.

Ammonium dinitramide (ADN)-based thrusters are pivotal for future spacecraft propulsion due to their low toxicity, adjustable specific impulse, and environmental benefits. However, complex fault patterns observed during ground tests challenge traditional fault detection methods, which struggle with high-dimensional, nonlinear time-series data. This study proposes a machine learning-based approach for robust fault diagnosis in ADN-based thrusters. Using 189 real engine test time-series datasets, we performed systematic preprocessing and feature engineering to extract statistical and correlation characteristics inside experimental data, creating a standardized dataset of normal and faulty conditions. Ten algorithms—six traditional machine learning and four deep learning—were evaluated for fault identification. The multilayer perceptron achieved 98.2% accuracy and 100% recall, while random forest and XGBoost, attained accuracies of 99.1% and 98.2% respectively, with superior computational efficiency. Deep learning excels in complex scenarios but demands longer training, whereas traditional methods suit real-time applications. Feature analysis highlighted pre-injection pressure and capillary outlet temperature as key fault indicators. A Simcenter AMESim-based simulation model further augmented the dataset, supporting fault mechanism studies. This approach enhances fault diagnosis, health monitoring, and design optimization for ADN-based thrusters, offering significant engineering value.