摘要
航天器遥测数据是地面运管系统判断其在轨运行状态的唯一依据,其异常检测成为增强航天器在轨可靠性和安全可靠运行的重要依据。但是,遥测数据量大、维度高、参数关系复杂、专业性强的特点为高检测率、低误检率、强解释性的遥测数据异常检测实现带来严峻挑战。本文对航天器遥测数据异常检测领域进行充分的调研,阐述遥测数据异常检测内涵、方法研究及应用现状,尤其针对数据驱动的遥测数据异常检测方法体系进行了系统、详细的分析,并在此基础上归纳总体发展动态,展望未来发展方向,以此为相关研究者提供参考。
Spacecraft telemetry data is the only basis for ground operation and management system to determine its on-orbit state, and the anomaly detection of which has become the important mean to enhance the spacecraft' s on-orbit reliability and safety. However, telemetry data has the characteristics of large amount of data, high dimension, complex relation which bring severe challenges for the realization of anomaly detection with high detection rate, low false detection rate, and strong interpretation. In this work, the full analysis has been conducted on the anomaly detection area of spacecraft telemetry data. The anomaly detection definition, anomaly detection connotation, anomaly detection method research and application status have been explained in detail. Especially, the data- driven methods are emphasized system atically and effectively. Furthermore, the paper summarizes the overall development and look to the future direction to provide reference for the related researchers.
作者
彭喜元
庞景月
彭宇
刘大同
Peng Xiyuan Pang Jingyue Peng Yu Liu Datong(Department of Automatic Test and Control, Harbin Institute of Technology, Harbin 150080, China)
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2016年第9期1929-1945,共17页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(61571160
61301205)
部委重点基金课题(9140A17050114HT01054)
哈尔滨工业大学学科发展新方向(01509421)
中央高校基本科研业务费专项资金(HIT.NSRIF.2014017)项目资助
关键词
航天器
遥测数据
异常检测
数据驱动
Spacecraft
Telemetry data
Anomaly detection
Data-driven