Transformer-Based Time-Series Forecasting for Telemetry Data in an Environmental Control and Life Support System of Spacecraft

• DOI: 10.3390/electronics14030459
• Publication Date: 2025-01-23T17:24:09Z
• Authors: Bin Song; Boyu Guo; Wei Hu; Zhen Zhang; Nan Zhang; Junpeng Bao; Jianji Wang; Jingmin Xin
• Keywords: Life support system; Environmental data

Safety and stability are critical in manned space missions, requiring an environmental control life support system (ECLSS) of spacecraft to operate reliably. This study analyzed the time-series characteristics telemetry data, including total pressure, temperature, humidity, predict ECLSS’s operational state. Existing algorithms for forecasting, ARIMA, LSTM, TCN, NBEATS, often struggle with long-sequence forecasting discrepancies data distribution, which hinder their ability deliver accurate predictions. To address these challenges, this introduces a two-stage normalization method, mean instance (MeanIN), designed adjust input distributions restore output distributions, thereby significantly enhancing predictive performance. Experimental evaluations on ECLSS demonstrate that MeanIN consistently improves model accuracy, informer achieving superior results tasks. These underscore efficacy its potential applications anomaly detection analysis data.