NASA Spacecraft Reliability Engine
This section analyzes and dedect anomalies for Spacecraft working on real NASA dataset.
Reliability Simulation
Expected Failures
--
Observed Failures
--
Failure Rate
--
Simulation Accuracy
--
Anomaly Assessment
--
Failure Attribution
V2 uses Bayesian inference to estimate the possible cause of a failure.
P(C|E) = P(E|C)P(C) / P(E)
Confusion Matrix
No simulation data yet.
Attribution Performance
--
Simulation Details
--
Analysis Summary
--
Limitations
- V1 uses stimulated system failures(for actual Nasa dataset move to v3).
- V2 depends on the probability evidence used for the stimulated causes that have happened.
- The result is on the basis of probilistic and hardcoded anomalies.
NASA Telemetry Analysis
Precision
--
Recall
--
F1 Score
--
Anomalies
--
Threshold
--
Persistence
--
Telemetry Result
--
Detection Details
--
Performance
--
Analysis Summary
--
Limitations
- Different Telemetry channels have different type of anomalies.
- Model performance can change between telemetry channels its not fixed.
- Threshold and persistence are calibrated from validation experiments.