Anomaly Detection with Improved Similarity Measure for Satellite Telemetry Data  
Author Datong Liu

 
Co-Author(s) Shaojun Wang; Jing Chen; Jun Zhou; Yu Peng

 
Abstract Anomaly detection with telemetry data can discover potential failure in advance and improve the operating safety for satellite. Most of the former studies used Euclidean distance to measure the similarity of telemetry data which are typical time series. However, the Euclidean distance has many limitations such as lacking of dynamic capability and stability with correlation relationship for multiple dimensions. Therefore, different improved data distance measures are studied for telemetry data mining in this paper. Then, an improved anomaly detection framework with different similarity measures is presented. We achieve the anomaly detection to satellite telemetry data based on the hierarchical clustering and improved similarity measures. Experimental results show that the proposed anomaly detection framework can obtain satisfied performance on the actual satellite telemetry data sets.

 
Keywords Satellite, Telemetry Data, Anomaly Detection, Dynamic Time Warping, Hierarchical Clustering
   
    Article #:  22192
 
Proceedings of the 22nd ISSAT International Conference on Reliability and Quality in Design
August 4-6, 2016 - Los Angeles, California, U.S.A.