Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Adaptive Active Learning for Online Reliability Prediction of Satellite Electronics

Shixiang Li, Yubin Tian, Dianpeng Wang, Piao Chen, Mengying Ren

TL;DR

Numerical experiments and a practical case study from the Tiangong space station demonstrate that the proposed method markedly improves reliability prediction accuracy while significantly reducing data requirements, offering an efficient solution for the prognostic and health management of complex satellite electronic systems.

Abstract

Accurate on-orbit reliability prediction for satellite electronics is often hindered by limited data availability, varying operational conditions, and considerable unit-to-unit variability. To overcome these obstacles, this paper proposes a novel integrated online reliability prediction framework. The main contributions are twofold. First, a Wiener process-based degradation model is developed, incorporating a generalized Arrhenius link function, individual random effects, and spatial correlations among adjacent units. A customized maximum likelihood estimation method is further devised to facilitate efficient and accurate parameter inference. Second, a two-stage active learning sampling scheme is designed to adaptively enhance prediction accuracy. This strategy initially selects representative units based on spatial configuration, and subsequently determines optimal sampling times using a comprehensive criterion that balances unit-specific information, model uncertainty, and degradation dynamics. Numerical experiments and a practical case study from the Tiangong space station demonstrate that the proposed method markedly improves reliability prediction accuracy while significantly reducing data requirements, offering an efficient solution for the prognostic and health management of complex satellite electronic systems.

Adaptive Active Learning for Online Reliability Prediction of Satellite Electronics

TL;DR

Numerical experiments and a practical case study from the Tiangong space station demonstrate that the proposed method markedly improves reliability prediction accuracy while significantly reducing data requirements, offering an efficient solution for the prognostic and health management of complex satellite electronic systems.

Abstract

Accurate on-orbit reliability prediction for satellite electronics is often hindered by limited data availability, varying operational conditions, and considerable unit-to-unit variability. To overcome these obstacles, this paper proposes a novel integrated online reliability prediction framework. The main contributions are twofold. First, a Wiener process-based degradation model is developed, incorporating a generalized Arrhenius link function, individual random effects, and spatial correlations among adjacent units. A customized maximum likelihood estimation method is further devised to facilitate efficient and accurate parameter inference. Second, a two-stage active learning sampling scheme is designed to adaptively enhance prediction accuracy. This strategy initially selects representative units based on spatial configuration, and subsequently determines optimal sampling times using a comprehensive criterion that balances unit-specific information, model uncertainty, and degradation dynamics. Numerical experiments and a practical case study from the Tiangong space station demonstrate that the proposed method markedly improves reliability prediction accuracy while significantly reducing data requirements, offering an efficient solution for the prognostic and health management of complex satellite electronic systems.
Paper Structure (13 sections, 29 equations, 3 figures, 3 tables)

This paper contains 13 sections, 29 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. A Hierarchical Spatiotemporal Degradation Model
  3. Efficient Profile Likelihood Inference
  4. Optimal Spatiotemporal Data Acquisition via Active Learning
  5. Spatial Active Learning: Uniform Coverage via Space-Filling Design
  6. Sequential Temporal Active Learning: A Balanced Information Criterion
  7. Gradient with Respect to $\boldsymbol{\gamma}=(\gamma_1,\gamma_2)^\top$
  8. Gradient with Respect to $\alpha$
  9. The Proposed Criterion
  10. Illustrative Examples
  11. Artificial Example
  12. Real Case Study
  13. Conclusion

Figures (3)

  • Figure 1: Illustration of the proposed spatial active learning via threshold accepting (left) and random swap algorithms (right) for a toy example with $L=8$, $o=10$, and $c=5$. Red points denote selected units.
  • Figure 2: Average reliability prediction results over 1000 replications when $(S_1,S_2,S_3,S_4)$ is $(1,1,1,0)$, $(0,1,1,0)$, $(1,1,0,0)$, and $(0,1,0,0)$.
  • Figure 3: Average reliability predictions over 12 devices during 10-12 years.