Deep Learning-Based Anomaly Detection in Spacecraft Telemetry on Edge Devices

2026-03-31

AI Summary

This study addresses the challenge of achieving both real-time performance and security in telemetry anomaly detection on resource-constrained spaceborne edge devices. For the first time, multi-objective neural architecture search is introduced to spacecraft telemetry anomaly detection, systematically evaluating three paradigms: prediction-plus-thresholding, direct classification, and image-based classification. The approach integrates time series modeling, thresholding strategies, and model compression into a unified optimization framework. The resulting lightweight model enables efficient inference on CubeSat-class hardware, occupying only 59 KB of memory (a 97.1% reduction) and reducing computational cost by 99.4%. It achieves a CEF0.5 score of 88.8% while consuming merely 0.36–6.25% of onboard memory, thereby attaining a Pareto-optimal trade-off between performance and resource overhead under extreme constraints.

Abstract

Spacecraft anomaly detection is critical for mission safety, yet deploying sophisticated models on-board presents significant challenges due to hardware constraints. This paper investigates three approaches for spacecraft telemetry anomaly detection -- forecasting & threshold, direct classification, and image classification -- and optimizes them for edge deployment using multi-objective neural architecture optimization on the European Space Agency Anomaly Dataset. Our baseline experiments demonstrate that forecasting & threshold achieves superior detection performance (92.7% Corrected Event-wise F0.5-score (CEF0.5)) [1] compared to alternatives. Through Pareto-optimal architecture optimization, we dramatically reduced computational requirements while maintaining capabilities -- the optimized forecasting & threshold model preserved 88.8% CEF0.5 while reducing RAM usage by 97.1% to just 59 KB and operations by 99.4%. Analysis of deployment viability shows our optimized models require just 0.36-6.25% of CubeSat RAM, making on-board anomaly detection practical even on highly constrained hardware. This research demonstrates that sophisticated anomaly detection capabilities can be successfully deployed within spacecraft edge computing constraints, providing near-instantaneous detection without exceeding hardware limitations or compromising mission safety.

Problem

• anomaly detection
• spacecraft telemetry
• edge devices
• hardware constraints
• on-board deployment

Innovation

• edge computing
• neural architecture optimization
• anomaly detection
• spacecraft telemetry
• on-board deployment