![[2411.12159] Sensor-fusion based Prognostics for Deep-space Habitats Exhibiting Multiple Unlabeled Failure Modes](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2411.12159] Sensor-fusion based Prognostics for Deep-space Habitats Exhibiting Multiple Unlabeled Failure Modes
Summary
This paper presents a novel unsupervised prognostics framework for deep-space habitats, addressing multiple unlabeled failure modes through sensor fusion and machine learning techniques.
Why It Matters
As deep-space missions require autonomous systems to function without ground support, understanding and predicting failures is crucial. This research enhances the reliability of such habitats by improving the accuracy of remaining useful life predictions, which is essential for mission success and safety.
Key Takeaways
- Introduces an unsupervised framework for identifying failure modes in deep-space habitats.
- Utilizes a mixture of Gaussian regressions for modeling system failures.
- Demonstrates improved accuracy in predicting remaining useful life using real-world datasets.
- Combines sensor relevance assessment with failure mode diagnosis.
- Addresses the challenges of working with unlabeled training data in complex systems.
Statistics > Machine Learning arXiv:2411.12159 (stat) [Submitted on 19 Nov 2024 (v1), last revised 14 Feb 2026 (this version, v4)] Title:Sensor-fusion based Prognostics for Deep-space Habitats Exhibiting Multiple Unlabeled Failure Modes Authors:Benjamin Peters, Ayush Mohanty, Xiaolei Fang, Nagi Gebraeel, Stephen K. Robinson View a PDF of the paper titled Sensor-fusion based Prognostics for Deep-space Habitats Exhibiting Multiple Unlabeled Failure Modes, by Benjamin Peters and 4 other authors View PDF HTML (experimental) Abstract:Deep-space habitats are complex systems that must operate autonomously over extended durations without ground-based maintenance. These systems are vulnerable to multiple, often unknown, failure modes that affect different subsystems and sensors in mode-specific ways. Developing accurate remaining useful life (RUL) prognostics is challenging, especially when failure labels are unavailable and sensor relevance varies by failure mode. In this paper, we propose an unsupervised prognostics framework that jointly identifies latent failure modes and selects informative sensors using only unlabeled training data. The methodology consists of two phases. In the offline phase, we model system failure times using a mixture of Gaussian regressions and apply an Expectation-Maximization algorithm to cluster degradation trajectories and select mode-specific sensors. In the online phase, we extract low-dimensional features from the selected sensors to diagnose the ...
