With the rapid advancement of artificial intelligence (AI) technology, AI is frequently talked about and debated in space systems. The strong interaction between AI and space systems is an exciting new field, as the autonomy of AI systems enables them to continuously monitor space debris and satellite orbits in real time, calculating the potential risk of collisions. By predicting dangerous approaching in advance, AI systems suggest or even execute maneuvers automatically to avoid collisions, ensuring the safety and longevity of satellites. This will help break through the current limitations of available resources and personnel costs, and promote new developments in the field of space exploration. However, every benefit brought by AI in space systems may have opposite risks manifested as potential attack, infiltration and compromise. Because AI systems have autonomous consciousness, the unpredictability of their behaviour can introduce uncontrollable situations to people, which in turn affects the completion of space missions. Therefore, the current problem caused by the combination of AI and space exploration is that people cannot be assured with the reliability and safety of space systems that utilize AI technology. In response to the above problems, this paper takes AI technology in space systems as the research object. Firstly, it investigates the application progress of AI technology in space systems. And the classification of AI technology in space systems is preliminarily determined according to current mainstream research theory or standard. Considering the high-risk environment such as radiation and temperature change, the key points of safety design for AI in space systems are analysed from the data, model and system. The rough set theory based on genetic particle swarm optimization is used to reduce attributes, and the focus of safety design for AI in space systems is further optimized. The authors hope that through the research of this paper, the level of AI technology in space systems can be clearly elucidated, providing a solution for effectively assessing the factors of safety design for AI in space systems, and strengthening the significant benefits of AI driving the development of space systems. (c) 2025 International Association for the Advancement of Space Safety. Published by Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

This study proposes a Bayesian network-based method for evaluating the security of human-machine interaction systems. It considers the effects of time pressure and scenario fusion control transitions on group member performance factors (such as cognitive load, BDI, and distributed situational awareness), and their relationship with system security. By constructing a multi-agent discrete dynamic event framework, it investigates the independent impacts of time pressure and external environmental disturbances on system security and establishes a model for system security. Additionally, it analyzes various risk scenarios in aerospace systems, including ground risks, launch and in-orbit teleoperation risks, and extravehicular activity risks, while considering risk factors in different scenarios. Ultimately, the dynamic risk level of the system is assessed through dynamic Bayesian networks, resulting in a calculated accident probability of 0.14‰, demonstrating the effectiveness of the method. This research provides a new technological approach for evaluating the security of human-machine interaction and contributes to enhancing the safety of aerospace systems.

Space station is a very complex system, and its remaining useful life will be affected by the key equipment, cosmonauts' maintenance activities as well as space environments. It is important for the operation management of a space station to predict its remaining useful life (RUL). A valid RUL prediction model is the key foundation for this issue, which motivates the research presented in this paper. Firstly, different types of space station life are defined. Secondly, the function and performance requirements as well as the operation mission program of the space station are analysed, which are further used to confirm the model development precondition. A life prediction model is then proposed by synthetically taking account of the safety, reliability and maintainability restrictions. Finally, the data requirement for supporting the RUL, prediction is determined. Based on this work, a comprehensive procedure for RUL prediction model development is constructed for the operation management engineers of the space station. If the data of the development and operation is adequate, RUL prediction of the space station can be well implemented, and can be further leveraged to support the space station operation management.

Recently, with the world-wide enthusiasm for exploring outer space and the development of astronautics technologies, Spaceflight tasks become more complicated. Human, system and environments exchanging information during the Human Involved Complex Spaceflight Tasks (HICSTs), makes comprehensive analysis of safety and reliability a critical problem to solve. We have to determine the safety and reliability of HICSTs before the tasks are implemented. This paper analyzes factors which influence the safety and reliability of HICSTs, and study the interdependent relation among human, system and environments. Then we make further study on safety and reliability modeling and assessment methodologies and provide a systematic method to assess the reliability and safety for the HICSTs. Finally, we choose a typical HICST to explain how to model and evaluate the reliability and safety of a HICST. The results show that the method proposed in this paper is effective. The further study and development trends of this are also outlined.

The generalized Man-Machine System (MMS) is defined as a complex aggregation having specific function which is made up of interactive man, machine and environment. The object of this paper is Space Man-Machine System (SMMS) which is represented by manned spacecraft, space laboratory and space station. SMMS refers to numerous man, machine and environment safety-influencing factors. There exist some disadvantages of artificial screening of safety compensation factors (CPFs) and irreplaceable factors (IRFs), independent factors (IDFs) and combination factors (CBFs), such as heavy workloads, easy to confuse and omission. This paper proposed a safety influencing factors screening method for SMMS based on Analytic Hierarchy Process (AHP), Decision-Making Trial and Evaluation Laboratory (DEMATEL) and Fault Tree Minimum Cut Set (FT-MCS), which is able to screening numerous safety influencing factors quickly and effectively, and find the key point of controlling accident to ensure safety operation of SMMS.

The attitude control system (ACS) is crucial for the satellite operation in the space, and it is a typical phased-mission system (PMS). Some features caused by the system designs bring difficulties in the ACS mission risk analysis, and the traditional risk analysis method probabilistic risk assessment (PRA) must be modified to model and analyze the redundant and dynamic features in the system. This paper proposes an improved method for mission risk analysis based on the traditional PRA framework, and combines the dynamic fault tree (DFT) and Bayesian networks (BN) methods with the traditional framework to solve the redundancy features, operation sequence dependencies, and relevant failure properties of some components in the ACS. This method can give a more accurate description of the ACS mission, and the analysis results show that this proposed method can provide more useful information for system design improvements.

In the development of complex spacecraft, new technologies or new processes are often introduced in the design. Therefore, the problems of lack of information and data are usually encountered in the decision-making of designs selection. Relative risk evaluations can be applied in the selection of designs to cope with these problems. Also, in many instances, evaluators are more comfortable making relative decisions versus assigning absolute values. In this paper, a framework based on relative risk evaluations is presented which assists in making the decision on the design or designs to select. The proposed approach starts with applying the Fussell-Vesely measure to obtain the importance rankings of different risk contributors of different end states. The analytical hierarchy process (AHP) method is then used to determine the criteria weights of different risk end states. Accordingly, the relative importance of the particular contributors to the total risk for the reference design is calculated. On the other hand, the relative difference ratios in the contribution for alternative designs compared to reference design are calculated. Combined with the relative importance and the relative difference ratios of contributors, the relative risk difference ratios between different designs are evaluated. Finally, the optimal alternative design is determined by ranking relative risk evaluations. The performances of the proposed approaches are illustrated and validated using an example of a spacecraft propellant distribution system development. The results show that the proposed approaches are capable of helping designers to systematically consider relevant information and effectively determine the optimal design alternatives in the spacecraft development.

针对现代战争信息化和网络化的发展趋势,探讨信息技术标准在战争中的重要作用,通过研究美国国防部联合技术体系结构(JTA)标准的发展历程和主要内容,分析其应用特点以及对我国作战信息技术标准建设的启示。

FMEA分析验证技术是一种基于FMEA的可靠性仿真分析技术。本文详细介绍了FMEA分析验证技术在国内航天领域的研究状况,包括基本分析流程、仿真建模技术研究;并介绍了该技术在国内航天应用中的发展前景。

全面总结了国内外光纤陀螺技术发展现状，并对其空间应用前景进行了具体分析，提出了光纤陀螺空间应用需突破的关键技术，对后续工作进行了展望。

Aiming at the localization of traditional Bayesian Networks, Monte Carlo method is used to Bayesian Networks. Applying MC-BN(Monte Carlo Bayesian Networks) to availability analysis of Satellite Ground Station, combining actual engineering data, and clearing the relationships of stations, systems, subsystems and equipments, the detail steps of simulation, availability, the most probable modes causing the system faults and sequence of importance are given in this paper. The result shows that MC-BN is effective in availability analysis of Satellite Ground Station and engineering application.

This paper firstly analyzes the concept of availability of satellite constellation. Then the method of constellation availability analysis is proposed on the basis of Markov chain model, outage analysis, as well as MTBF and MTTR of satellite constellations. Finally, an illustrative example is presented to validate the method.

简要介绍了我国航天产品环境应力筛选应用背景,分析了航天产品环境应力筛选现状,提出了提高航天产品环境应力筛选有效性的结论与建议。

备件保障是保证导弹系统战备完好性和作战能力的重要因素,谋求导弹产品维修备件需求与备件配置的平衡是一项长期未解决的难题。本文对导弹产品贮存中的备件问题作一全面分析,为导弹产品筹划备件提供思路和方法。

主要针对共因失效产生的原因及共因失效分析的开展时机、实施流程、跟踪报告表格格式等做了说明,为在航天型号内开展共因失效分析提供了指导。

简要介绍潜在电路分析(SCA)技术的原理及实施方法,以及在我国航天控制系统中的初步研究成果,以期引起重视,推动这项研究在我国的进一步开展.