Abstract:Space mission design software spans the entire lifecycle of space missions and is regarded as the cornerstone of the aerospace industrial software system. Leveraging nearly three decades of technical expertise in the field of manned spaceflight, the research team from the National University of Defense Technology initiated the development of the indigenous software ATK (Aerospace Tool Kit) in 2020. By the end of 2024, the ATK 3.0 version was officially released, encompassing five major categories and 21 functional modules, including standard platform, visibility and coverage analysis, mission analysis, orbital design, and secondary development. The development of ATK has been deeply and continuously supported by China’s Manned Spaceflight Program, pioneering a collaborative research and development model characterized by "engineering-driven demand + centralized academic development." This software has had a significant impact in aerospace engineering, space security, and academic education, marking a solid first step toward replacing STK (Systems Tool Kit). This paper reviews the development history of ATK, elaborates on its functional features, and focuses on analyzing its breakthroughs in core technologies such as large-scale software architecture design, computational kernel development, and orbital maneuver planning. While ATK faces challenges in enhancing its functionality and building its application ecosystem, the paper outlines its future development roadmap and strategic objectives .

Abstract:As a representative of the next generation of gyroscopes , atomic gyroscopes have become a focal point of research in the field of high-precision inertial navigation and has garnered significant attention due to its theoretically ultra-high precision, exceptional long-term stability, and immense potential for miniaturization and integration. Among them, the atom interferometric gyroscope, as a type of atomic gyroscope, had attracted widespread interest in the field of inertial navigation.. The development of atomic interferometric gyroscopes was systematically reviewed. Beginning with fundamental principles, it elaborates on critical technical components including atomic source preparation, interferometric loop construction, and phase resolution. Through rigorous analysis, the paper establishes intrinsic correlations between core performance parameters such as sensitivity and ultimate accuracy while elucidating their mutual constraint mechanisms. Furthermore, it reveals the physical origins of engineering bottlenecks including limited data update rates and narrow dynamic ranges.?Finally, the article outlines future development directions and trends for atomic interference gyroscopes, emphasizing the need for in-depth research in several areas: solving external interference issues to improve accuracy, improving chip processing technology for miniaturization and integration, enhancing combined inertial sensors, increasing data update rates, and exploring ways to expand dynamic range.

Abstract:A modified physical vapor deposition method for the controllable growth of α-In2Se3 was proposed, and the broad-spectrum response performance of three thicknesses of α-In2Se3 nanosheets in the visible to near-infrared wavelength range was systematically studied. The results indicate that the thickness of α-In2Se3 nanosheets can significantly regulate the photoelectric performance, and the photoresponsivity and specific detection rate increase with increasing thickness. In addition, it was found that the α-In2Se3 with a thickness of 32.8 nm exhibited a photocurrent anisotropy ratio (dichroic ratio) of 4 at 635 nm, indicating good polarization-sensitive detection functionality. In summary, the two-dimensional α-In2Se3 prepared by the physical vapor deposition method demonstrates a wide visible-infrared spectral response and good polarization detection ability, making it an ideal candidate material for two-dimensional multifunctional optoelectronic devices.

Abstract:Due to the low damping characteristic of the superconducting electrodynamic suspension (EDS) system which may cause suspension instability problem, a six degree-of-freedom dynamic model for the high speed maglev sled system is established, and it is found that its suspension damping would become negative when the traveling speed exceeds 23.6 m/s, resulting in suspension instability to the maglev sled. The vertical damping ratio of the EDS system is calculated using the least square fitting method, and the relationship between the vertical damping and the speed is then obtained. To stablilize the suspension system, a distributed dynamic vibration absorbers (DVA) scheme is proposed, and the effects of the DVA parameters on the suspension stability are investigated; to minimize the influence of the DVA scheme on the acceleration performance of the maglev sled, the feasibility of applying low mass (≤ 1kg) DVAs to the maglev sled is disccussed; further, the vibration suppression effect of this scheme is investigated considering the vertical misalignments of guideway girders, which shows that the proposed scheme can significantly increase the damping and stability of the suspension system, and it also well suppresses the vibration of the sled-body caused by misalignments of the ground suspension coils. This scheme provides a useful reference for the design of the suspension system of the maglev sled system, the vaccum tube high speed maglev train, etc.

Abstract:In view of the insufficient mining of implicit associative relationships and the problem of neglecting the temporal evolution factor, we propose a domain-specific social events detection method via temporal evolution feature mining. Our model divides data into slices by time, and takes duplicate event records from different sources into account which reflects the importance of events, and constructs entity interaction graph, which reduces the impact of database error. Multi-relational graph convolutional network is improved to update graph structure information of historical evolution sequence by interaction relationships. Attention mechanism is used to learn core features to obtain global embedding of sequence units. Implicit association is mined sufficiently. Based on recurrent neural network, temporal evolution features are extracted to obtain the global embedding and the temporal evolution factor is mined effectively. Experiment results show that our method can be applied to detection task, which is better than existing methods.

Abstract:The efficient online exploration of intelligent agents is important in reinforcement learning tasks, but still faces the problem of low utilization of interactive data with environment, or the need for additional tasks’ data. To solve this problem, an online exploration latent variable that obtained the characteristics of current task to assist the agents to behave was introduced. There was no need for additional multi-task data or additional environmental interaction steps in the current task. The exploring latent variable was updated in the learnable environment model, and the environment model underwent supervised updates based on the intelligent agent and real environment interaction data. The exploration in advance in the simulated environment model was assisted by the exploring latent variable, and thus the performance of agents in the real environment was improved. The performance in typical continuous control tasks was raised by about 30% in the experiments, which was of guiding significance for the single-task exploration and meta reinforcement learning research.

Abstract:To enhance the ability of large language models to generate legal documents for the power grid sector under few-shot conditions, a few-shot legal document generation method based on large language models (LLM) was proposed, integrating data augmentation and rule guidance techniques. The proposed method addressed key challenges in power grid legal document generation, such as data scarcity, high domain specificity, and the complexity of legal practice. Experimental results show that the method achieves excellent performance in generating power grid legal defense documents, significantly improving the quality and professionalism of the generated texts.

Abstract:In the complex environment of strong confrontation, the entity perception information is incomplete and the real-time response is required, which poses a challenge to the long-term and forward-looking dynamic force deployment decision. How to realize efficient exploration of strategies through explainable effective rewards and incentives is the key to drive strategic planning of dynamic force deployment by using learning methods. Aiming at the dynamic force deployment problem, this paper first proposes a multi-agent reinforcement learning strategy planning method based on SVD (Shapley value decomposition). The reward distribution among cooperative multi-agents is explained by SVD, and the reward distribution is analysed by SVD reinforcement learning method to solve Markov convex game strategy; Secondly, based on the scenario of naval and air cross-domain cooperative confrontation, this paper analyses the allocation of space domain combat resources in heterogeneous multi-entity cooperative confrontation, builds a dynamic force deployment strategy planning model, and designs the state space, action space and reward function of the problem. Finally, based on typical application scenarios, simulation experiments are organized to verify the dynamic force deployment problem with the military chess deduction system. The results show that compared with the multi-class baseline algorithm, the proposed method in this paper has excellent performance in strategic planning of dynamic force deployment, and it is theoretically interpretable. The proposed method learned the strategy of "layer upon layer interception, zone confrontation, cover core, and layered breaking". The method of project address: https://gitee.com/jrluo2049/shapleymarl.

Abstract:The bevel gear is the core transmission component in the main reducer of a helicopter that realizes the power transmission of intersecting axes, whose performance and dynamic characteristics directly affect the operational stability of the helicopter transmission system. Taking into account factors such as transmission error and backlash, a nonlinear damage dynamic model of a bevel gear system with bending-torsion-axial coupling was established using the lumped-parameter method. In order to obtain the key parameter in the damage dynamics model, a slicing method for calculating the time-varying meshing stiffness of spur bevel gear pair was proposed, which breaks through the limitations of traditional potential energy method that can only be applied to cylindrical gear pair and can achieve rapid evaluation of time-varying meshing stiffness under different states. By comparing the model simulation and experimental results, it is shown that the established model can effectively simulate the dynamic response characteristics of the bevel gear system under normal and fault conditions, reveal the system fault response mechanism and fault mechanism, and provide theoretical basis and data support for the development of a helicopter transmission system health and usage monitoring system.

Abstract:Building upon the traditional physics-informed neural network, two improved methods, EmPINN and DL-PINN, are presented by incorporating dimension expansion and diverse physics loss functions. EmPINN innovatively introduces a neural network structure with residual connections and a dimension-expanding mechanism. DL-PINN, based on EmPINN, combines the dimension-expanding mechanism with gradient enhancement and variational physical information to incorporate physical information more effectively and improve the fitting capability of the neural network. Experimental results demonstrate that the proposed methods outperform traditional PINN method, achieving up to two orders of magnitude improvement in the accuracy of solving different partial differential equations.