Fangzhou Sun, Abhishek Dubey, C Kulkarni, Nagbhushan Mahadevan, Ali Guarneros Luna
Conference Proceedings, Annual Conference of The Prognostics And Health Management Society 2018
Small satellite missions depend on reliable battery health monitoring to ensure mission success, but traditional statistical approaches fail to capture complex battery aging patterns. This work is innovative because it combines offline deep learning with online adaptation mechanisms enabling satellites to detect imminent failures without relying on ground station resources. The approach enables extended mission duration through proactive battery management.
This paper proposes a data-driven health monitoring framework for extending small satellite missions using Deep LSTM networks to detect battery anomalies during operation. The approach combines offline pre-training with online transfer learning to adapt battery prediction models to individual satellite configurations. The system uses stacked LSTM and auto-encoder architectures to detect failures in highly imbalanced datasets while accounting for the spatial-temporal dynamics of battery degradation.
The two-layered LSTM architecture successfully detected battery anomalies and predicted remaining useful life within acceptable margins for operational decision-making. The system identified degradation patterns specific to charging cycles and environmental conditions through transfer learning. Online prediction enabled timely battery replacement decisions, extending mission duration while preventing catastrophic failures.
@inproceedings{Sun2018a,
author = {Sun, Fangzhou and Dubey, Abhishek and Kulkarni, C and Mahadevan, Nagbhushan and Luna, Ali Guarneros},
booktitle = {Conference Proceedings, Annual Conference of The Prognostics And Health Management Society},
title = {A data driven health monitoring approach to extending small sats mission},
year = {2018},
abstract = {In the next coming years, the International Space Station (ISS) plans to launch several small-sat missions powered by lithium-ion battery packs. An extended version of such mission requires dependable, energy dense, and durable power sources as well as system health monitoring. Hence a good health estimation framework to increase mission success is absolutely necessary as the devices are subjected to high demand operating conditions. This paper describes a hierarchical architecture which combines data-driven anomaly detection methods with a fine-grained model-based diagnosis and prognostics architecture. At the core of the architecture is a distributed stack of deep neural network that detects and classifies the data traces from nearby satellites based on prior observations. Any identified anomaly is transmitted to the ground, which then uses model-based diagnosis and prognosis framework to make health state estimation. In parallel, periodically the data traces from the satellites are transported to the ground and analyzed using model-based techniques. This data is then used to train the neural networks, which are run from ground systems and periodically updated. The collaborative architecture enables quick data-driven inference on the satellite and more intensive analysis on the ground where often time and power consumption are not constrained. The current work demonstrates implementation of this architecture through an initial battery data set. In the future we propose to apply this framework to other electric and electronic components on-board the small satellites.},
category = {conference},
contribution = {minor},
file = {:Sun2018a-A_data_driven_health_monitoring_approach_to_extending_small_sats_mission.pdf:PDF},
keywords = {health monitoring, deep learning, LSTM networks, satellite systems, battery diagnostics},
project = {cps-reliability},
tag = {platform}
}