Gautam Biswas, Hamed Khorasgani, Gerald Stanje, Abhishek Dubey, Somnath Deb, Sudipto Ghoshal
Prognostics and Health Management Conference 2016
Spacecraft telemetry monitoring requires detecting unanticipated failures and operational anomalies without complete prior knowledge of system behaviors. This work is innovative in combining data-driven clustering with expert knowledge to differentiate between specialized operational modes and genuine anomalies, enabling more effective real-time monitoring and failure prevention in complex space missions.
This work proposes a mixed method combining unsupervised learning and expert analysis for detecting anomalies in spacecraft telemetry data. The approach divides mission timelines into segments, applies wavelet transforms for feature extraction, and uses hierarchical clustering to group normal versus anomalous operational behaviors. Expert input validates clustering results and identifies special operating modes.
The method successfully identifies anomalies in LADEE spacecraft electrical power system data across 223 mission days with extensive telemetry. Wavelet transform feature extraction combined with hierarchical clustering effectively groups similar operational patterns. The expert-guided interpretation distinguishes nominal operations from anomalous behaviors, validating the mixed-method approach for spacecraft health monitoring.
@inproceedings{Biswas2016,
author = {Biswas, Gautam and Khorasgani, Hamed and Stanje, Gerald and Dubey, Abhishek and Deb, Somnath and Ghoshal, Sudipto},
booktitle = {Prognostics and Health Management Conference},
title = {An application of data driven anomaly identification to spacecraft telemetry data},
year = {2016},
abstract = {In this paper, we propose a mixed method for analyzing telemetry data from a robotic space mission. The idea is to first apply unsupervised learning methods to the telemetry data divided into temporal segments. The large clusters that ensue typically represent the nominal operations of the spacecraft and are not of interest from an anomaly detection viewpoint. However, the smaller clusters and outliers that result from this analysis may represent specialized modes of operation, e.g., conduct of a specialized experiment on board the spacecraft, or they may represent true anomalous or unexpected behaviors. To differentiate between specialized modes and anomalies, we employ a supervised method of consulting human mission experts in the approach presented in this paper. Our longer term goal is to develop more automated methods for detecting anomalies in time series data, and once anomalies are identified, use feature selection methods to build online detectors that can be used in future missions, thus contributing to making operations more effective and improving overall safety of the mission.},
category = {conference},
contribution = {colab},
file = {:Biswas2016-An_application_of_data_driven_anomaly_identification_to_spacecraft_telemetry_data.pdf:PDF},
keywords = {anomaly detection, spacecraft telemetry, wavelet analysis, clustering, unsupervised learning, feature extraction, health monitoring},
tag = {ai4cps}
}