首页 | 本学科首页   官方微博 | 高级检索  
     

自适应光学系统运行失稳检测方法
引用本文:贾启旺,李新阳,罗曦. 自适应光学系统运行失稳检测方法[J]. 红外与激光工程, 2020, 49(10): 20200299-1-20200299-10. DOI: 10.3788/IRLA20200299
作者姓名:贾启旺  李新阳  罗曦
作者单位:1.中国科学院自适应光学重点实验室,四川 成都 610209
摘    要:自适应光学(AO)系统校正像差是提高光学系统性能的有效技术手段。为了保证AO系统长时间安全、稳定地工作,需要对AO系统运行的数据进行监测,识别系统失稳状态以提供决策建议。基于以上目的,建立了一套127单元AO系统失稳数据仿真平台,通过该平台仿真得到了4种闭环失稳异常。基于变形镜控制电压RMS指标使用Kmeans聚类、K-NN分类和ARIMA预测3种机器学习方法进行识别检测。3种方法在不同类型异常数据中的检测结果有所不同,说明3种异常检测方法对系统失稳检测均有一定的效果和适用范围,在实际使用时,可以根据需要选择一种或综合多种方法进行检测。

关 键 词:自适应光学   异常检测   Kmeans聚类   K-NN分类   ARIMA预测
收稿时间:2020-04-05

Detection methods for instability of adaptive optics system
Affiliation:1.Key Laboratory on Adaptive Optics, Chinese Academy of Sciences, Chengdu 610209, China2.Institute of Optics and Electronics, Chinese Academy of Sciences, Chengdu 610209, China3.University of Chinese Academy of Sciences, Beijing 100049, China
Abstract:Applying adaptive optics (AO) system to correct aberrations is an effective technical way to improve the performance of optical systems. In order to ensure the long-term, safe and stable operation of the AO system, it is necessary to monitor the operating data of the AO system and identify the instability state of the system to provide decision making suggestions. Based on the above purpose, a set of 127 units AO system instability data simulation platform was established. The abnormal data frames were inserted into the closed-loop operation of the simulation platform, and the abnormal data sets under four kinds of closed-loop instability were obtained. Once the deformable mirror, the core component of AO system, worked abnormally, it will threaten the safety of the system. Based on the deformable mirror control voltage rms index, three machine learning methods were used: Kmeans clustering, K-NN classification and ARIMA prediction for recognition and detection. The detection results of the three methods in different types of abnormal data are different, indicating that the three anomaly detection methods have certain effects and scope of application for system instability detection. In actual application, one or a combination of multiple methods should be selected for testing.
Keywords:
本文献已被 万方数据 等数据库收录!
点击此处可从《红外与激光工程》浏览原始摘要信息
点击此处可从《红外与激光工程》下载免费的PDF全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号