AI-Based Smart System Developed for Deep Space Missions
Karnataka: Professor Anil Kumar from Adichunchanagiri University has developed an advanced AI-based real-time anomaly detection and automatic decision support system for deep space missions.
He explained that one of the biggest challenges in deep space missions is the communication delay between spacecraft and Earth, which can range from a few minutes to over 20 minutes. Because of this delay, real-time human intervention during critical events is not possible. To address this problem, the smart system was designed.
According to Professor Anil Kumar, the system was designed, developed, and tested using two important NASA datasets- CMAPSS (engine degradation data) and SMAP/MSL (satellite telemetry data).
He said the system has four main parts:
- Anomaly Detection Engine – uses models like Isolation Forest, LSTM Autoencoder, and One-Class SVM
- Time-Series Forecasting Module – estimates the Remaining Useful Life (RUL) of components
- Automatic Decision System – suggests the correct actions using both rule-based logic and machine learning
- Real-Time Dashboard – allows easy visualization of all data and manual control when needed
About the results, he mentioned that the system achieved an F1-score of 0.91 on SMAP/MSL data, which is excellent. On CMAPSS data, the RUL prediction had an average error (MAE) of 12.4 cycles. Additionally, the automatic decision system provided 87% correct corrective action recommendations.
He added that these results show that in the future, AI systems like this can help monitor spacecraft health and make timely decisions, especially in situations where communication delays make real-time human control impossible.